there were a lot of events recorded by gharchive.org of which 2,359,089 were push events containing 3,607,691 commit messages that amount to 276,624,311 characters filtered with words.py@e23d022007... to these 58 messages:
[MIRROR] AI actions won't unassign each other's movement targets & Mice stop being scared of people if fed cheese [MDB IGNORE] (#18268)
AI actions won't unassign each other's movement targets & Mice stop being scared of people if fed cheese (#72130)
Fixes #72116
I've had a persistent issue with basic mob actions reporting this error
and think I finally cracked it
When replanning with AI_BEHAVIOR_CAN_PLAN_DURING_EXECUTION it can run
Setup on one action leading to the plan changing, meaning that it runs
finishCommand to cancel all other existing commands
If you triggered a replan by setting up a movement action in the middle
of another movement action, cancelling the existing action would remove
the target already set by the current one.
We want actions to be able to remove their own movement target but not
if it has been changed by something else in the intervening time.
I fixed this by passing a source every time you set a movement target and adding a proc which only clears it if you are the source... but this feels kind of ugly. I couldn't think of anything but if you have a better idea let me know.
Also while I was doing this I turned it into a feature because I'm crazy. If you feed a mouse cheese by hand it will stop being scared of humans and so will any other mice it attracts from eating more cheese. This is mostly because I think industrial mouse farming to pass cargo bounties is funny. Mice controlled by a Regal Rat lose this behaviour and forget any past loyalties they may have had.
Mouse.Friend.-.Made.with.Clipchamp.mp4
Oh also I removed a block about cancelling if you have another target from the "hunt" behaviour, everywhere using this already achieves that simply by ordering the actions in expected priority order and it was messing with how I expected mice to work. Now if they happen to stop by some cheese they will correctly stop fleeing in order to eat it before continuing to run away.
Fixes a bug I kept running into. Makes it possible to set up a mouse farm without them screaming constantly. Lets people more easily domesticate mice to support Ratatouille gameplay.
🆑 add: Mice who are fed cheese by hand will accept humans as friends, at least until reminded otherwise by their rightful lord. fix: Fixed a runtime preventing mice from acting correctly when trying to flee and also eat cheese at the same time. /🆑
Co-authored-by: Jacquerel [email protected]
Removes skull balaclava and skull facepaint from loadout, places them hidden on the Almayer. (#2526)
This removes the skull balaclavas and the facepaints from the loadout menu and instead places them in 2 places hidden around the Almayer. The reason I have done this is that they are almost exclusively used by people who who are referencing a character- usually Ghost from MW2 (either version) or the characters from COD Ghosts. See below for more details.
This is an OOC meme item that doesn't fit the tone of CM, particularly because we already have an item with a skull on it in case you want to use it: Armor! They wrote things on armor in the movie, including a skull!
Instead, we have these types of people, running a skull 'clava in every round even as command or medical characters. This is a modern 'operator' look, not a Space 'Nam-esque look and not an Aliens look. If you want something that'd remind you of Space 'nam, just look at the classic 'born to kill' helmet. Now, look at these CALL OF DUTY CHARACTERS THAT THIS ITEM REFERENCES!
This is goofy as hell and kind of an outlier among the customization options since it is very clearly a reference to COD. Look at its description:
"The face of your nightmares. Heed the call of duty as the ghost in the night with this metal 'clava. Additionally protects against the cold. Now in black!"
You'd get laughed off a real marine base for wearing this, let alone wearing one on an op. We don't need more people running this every round, and this gives them something to fight over between eachother- if you want to larp as Simon 'Ghost' Riley from hit video game COD MW2 (either version) then you'll have to hunt it down.
Screenshots & Videos
Put screenshots and videos here with an empty line between the
screenshots and the <details> tags.
🆑 del: Removed skull balaclavas and skull facepaint from the loadout options maptweak: adds a single skull facepaint and balaclava, each hidden in their own locations on the Almayer. /🆑
Simplify with-google-analytics example (#43894)
- Make sure the linting passes by running
pnpm build && pnpm lint - The "examples guidelines" are followed from our contributing doc
First of all thanks for this amazing project and all the help you provide with these examples.
It seems there is code duplication in this example. After some tests
locally seem to _document.js is not necessary for gtag to work
properly.
I am aware of vercel/next.js#40645 and I would like to ask @dave-hay, @SukkaW and @ijjk to consider this is still necessary. If so why then not move all content of the scripts from _app to _document?
In any case, I am open to adding here some comments explaining what is the reason for this code duplication if necessary.
Changes that come from vercel/next.js#43897
- The
eventhashChangeComplete should be removed since/homeand/home/#sectionis not new pageview, but just reference to the same page.
If we go from /home to /home/#section (with a button click or a link for
example) this shouldn't trigger a new page visit on gtag.
For this reason, I think we should revert the changes from vercel/next.js#36079. If there is a better argument of why this should stay I am also open to creating comments to clarify this on the example since I don't think should be the default behavior and not useful in most cases.
- The
id="gtag-init"was added with no context to the example from vercel/next.js#29530
If there is a reason for this id in the script to existing I am open to adding a comment that clarifies this since in my experience is not necessary at all.
Edit: Batching with vercel/next.js#43897 as recommended by vercel/next.js#43897 (comment)
Co-authored-by: JJ Kasper [email protected]
add torch.autograd._set_view_replay_enabled, use in aot autograd (#92588)
tldr; this should fix some minor perf regressions that were caused by adding more as_strided() calls in aot autograd.
This PR adds a new context manager, torch.autograd._set_view_replay_enabled().
Context: AOT Autograd has special handling for "outputs that alias graph intermediates". E.g. given this function:
def f(x):
y = torch.mul(x, 2)
out = y.view(-1)
return out
AOT Autograd will do the following:
def fn_to_compile(x):
y = torch.mul(x, 2)
out = y.view(-1)
# return the graph intermediate
return y, out
compiled_fn = compile(fn_to_compile)
def wrapper(x):
y, out = compiled_fn(x)
# regenerate the alias of the graph intermediate
return out._view_func(y)
What's annoying is that out._view_func() will result in a .as_strided call, because out is an ordinary runtime tensor. This (likely?) caused a perf regression, because when running the backward, out as_strided_backward() is slower than our view_backward().
In this PR, I added some TLS for instructing autograd to do view replay instead of as_strided, even when given a normal tensor. I'm definitely interested in thoughts from autograd folks (cc @albanD @soulitzer). A few points that I want to bring up:
(1) One reason that this API seems generally useful to me is because of the case where you torch.compile() a function, and you pass in two inputs that alias each other, and mutate one of the inputs. Autograd is forced to add a bunch of as_strided() calls into the graph when this happens, but this would give users an escape hatch for better compiled perf in this situation
(2) To be fair, AOT Autograd probably won't need this TLS in the long term. There's a better (more complicated) solution, where AOT Autograd manually precomputes the view chain off of graph intermediates during tracing, and re-applies them at runtime. This is kind of complicated though and feels lower priority to implement immediately.
(3) Given all of that I made the API private, but lmk what you all think.
This is a followup of pytorch/pytorch#92255.
Pull Request resolved: pytorch/pytorch#92588 Approved by: https://github.com/ezyang, https://github.com/albanD
Small VoxPrimalis Fixes (#18944)
-
fuck you i don't need to give you a fucking summary
-
fixes
I said a funny joke about ripping a body part off and got banned. We made the second PID Loop, and Bortnite became the favorite mentor of the day. We are now scared of the cv->?->Output thingy but once we suceed we will rule all of FRC. If we fail then arson would be kinda funny but a bad idea.
A small point release, various smaller items, ongoing updates to matching table features, bug fixes, but nothing major.
SPECIAL THANKS:
- Thanks Umio-Yasuno in github issue #281 for actually being proactive and finding some Intel/AMD gpu device id lists. I wish more issues would be like that.
KNOWN ISSUES:
- DEBUG: --debug-arg and --debug-arg-use must use the full format: --debug-arg="-GS", or else the command line eats the args, even if in quotes. The error handlers will then complain about no data supplied, and it will exit.
BUGS:
- GRAPHICS: An accidental 'and' instead of 'or' test (see Code 1) led to systems without gpu or dri graphics drivers not showing their xorg driver even when present. This was due to a mistake, and also due to how Perl handles || and && in sequence, which made this bug not show up until I tested on a system with xorg graphics driver, but without dri or gpu drivers. Virtually no modern hardware or operating systems would trip this condition, but older hardware and operating systems, which may not have gpu or dri drivers, might. And did, in my case. This is by the way why I try to test on old hardware at least now and then.
FIXES:
- CODE: A poorly done attempt at optimization would have broken case insensitive pre-compiled regex with $pattern = qr/../ because you can't add /$pattern/i to precompiled pattern, but qr/.../i support only added perl 5.014. This should impact almost nobody, but it is/was a glitch. Basically qr/../ can only be used when no /i type modifier is required if supporting Perl less than 5.014.
See inxi-perl/docs/optimization.txt section REGEX for more on this.
Note that Perl already compares the values in the variable each iteration via a simple equality test, so the only real gain from using qr// is not having to do that equality test each iteration of a loop.
-
OUTPUT: Fixed a few small inner key name failures to use '-' instead of ' ' to separate key terms:
-
REPOS: Called urpm urpmq, which is the query tool, not the actual type.
-
GRAPHICS: Fixed some gpu_id.pl matching rules. Thanks Umio-Yasuno in github issue #281 for noticing that some of the matching rules were either wrong or not loose enough.
ENHANCEMENTS:
1a. OPTIONS: Long time oversight, no option to test or do one time change of key: value separator string ':'. This goes along with existing config option SEP2_CONSOLE. Added --separator/--sep {character}.
1b. OPTIONS: Added synonym for --output: --export, and for --output-file: --export-file.
2a. GRAPHICS: New Intel gpu data source, from intel, finally. This let us add a lot more gpu ids. Thanks Umio-Yasuno in github issue #281 for finding these.
2b. GRAPHICS: New AMD data source, from github. This let me fill in some more, albeit not as accurately as previous sources, but added more so fine. Thanks Umio-Yasuno in github issue #281 for finding these.
-
CONFIG: In a first, took a feature from acxi, --config, and imported it into inxi! This shows active current configuration, by file.
-
CPU: updated, fine tuned amd cpu microarch ids.
-
DISKS: More disk vendors added. Not as many as usual, I think the high tech sanctions against China may be slowing the rate of new Chinese SSD/USB vendors. But still some new ones, as always. Not many new IDs for existing ones though, that is noteworthy. A few new data sources to help pinpoint vendor names found too, though those won't in general impact users, but can be used to determine if a string is in fact a company name.
CHANGES:
- OUTPUT: Fix 2, -t 'started by:' key name changed to: started-by: -G 'direct render:' changed to 'direct-render:'.
DOCUMENTATION:
1a. MAN: there were a few <...> instead of [...] for required option arguments. Fixed those.
1b. MAN: also added --debug-id [string] since that is in general useful info.
1c. MAN: Added qualifiers about when xwayland: and compositor: items appear for default -Ga output.
1d. MAN: Typo in config path in man page, .conf/ should be .config/.
1e. MAN: for --output json/xml, added pointer to doc page on smxi.org, people being unable to grasp the output is getting tiresome.
1f. MAN: Added synonym for --output, --export.
2a. SMXI.ORG DOCS: added --output json/xml documentation page: https://smxi.org/docs/inxi-json-xml-output.htm - this is also linked to from the github wiki page, though of course nobody is going to read it, as well as from a few pages in smxi.org.
2b. Updated inxi-man,options,changelog.htm files.
- CHANGELOG: Changed to use same format as acxi.changelog, leading topic id's in upper case, makes it easier to scan read and organize.
4a. DOCS: docs/inxi-cpu.txt - cleaned up, re-arranged a bit, added cpuid data explanation, and updated header on inxi-perl/data/cpu/microarch to better explain the way amd does ext fam / ext model, which are not the same, bizarrrely, very confusing.
4b. DOCS: New: docs/inxi-disks.txt. Split out from inxi-resources.txt, part of the ongoing to documentation modularization, slowly splitting out sub topics from inxi-data.txt and inxi-resources.txt. Note this is in general only done when I'm working on that specific feature. But slowly, surely.
CODE:
- GRAPHICS: Test when no gpu drivers and no dri drivers but x drivers never showed x driver. Was supposed to be all || for tests:
if (@$gpu_drivers || $graphics{'dri-drivers'} && @$x_drivers){
https://perldoc.perl.org/perlop. I believe this led to test 1 being false, test 2 being false, and since that left tests 2 and 3 needing to be true for the && logical and to be true. Since only one of the two was true, the last bit was seen as false.
- GRAPHICS: Connected with 1, noticed that for some weird reason, I'd decided to assign the array ref for drivers like this:
@$x_drivers = (a, b, c); when it was supposed to be: $x_drivers = [a,b,c];
This did not cause any issues, since they mean the same thing, but it was silly to write it that way.
3a. DEBUG: Added --debug-arg-use which allows testers to run a specific argument combination that may be causing issues.
3b. DEBUG: Also added more validation, to make sure arg for --debug-arg / --debug-arg-use start with - or -- followed by a letter.
-
START: Removed this code block from set_konvi_data. I had left this in place for a release or two to make sure no need for it was found, but it will never be used since it never worked in the first place.
-
OPTIONS: in OptionsHandler::post_process(), reorganized the various run and exit triggers, help, configs, recommends, version, etc. All on top now.
Implement SymBool (#92149)
We have known for a while that we should in principle support SymBool as a separate concept from SymInt and SymFloat ( in particular, every distinct numeric type should get its own API). However, recent work with unbacked SymInts in, e.g., pytorch/pytorch#90985 have made this a priority to implement. The essential problem is that our logic for computing the contiguity of tensors performs branches on the passed in input sizes, and this causes us to require guards when constructing tensors from unbacked SymInts. Morally, this should not be a big deal because, we only really care about the regular (non-channels-last) contiguity of the tensor, which should be guaranteed since most people aren't calling empty_strided on the tensor, however, because we store a bool (not a SymBool, prior to this PR it doesn't exist) on TensorImpl, we are forced to immediately compute these values, even if the value ends up not being used at all. In particular, even when a user allocates a contiguous tensor, we still must compute channels-last contiguity (as some contiguous tensors are also channels-last contiguous, but others are not.)
This PR implements SymBool, and makes TensorImpl use SymBool to store the contiguity information in ExtraMeta. There are a number of knock on effects, which I now discuss below.
- I introduce a new C++ type SymBool, analogous to SymInt and SymFloat. This type supports logical and, logical or and logical negation. I support the bitwise operations on this class (but not the conventional logic operators) to make it clear that logical operations on SymBool are NOT short-circuiting. I also, for now, do NOT support implicit conversion of SymBool to bool (creating a guard in this case). This does matter too much in practice, as in this PR I did not modify the equality operations (e.g.,
==on SymInt) to return SymBool, so all preexisting implicit guards did not need to be changed. I also introduced symbolic comparison functionssym_eq, etc. on SymInt to make it possible to create SymBool. The current implementation of comparison functions makes it unfortunately easy to accidentally introduce guards when you do not mean to (as boths0 == s1ands0.sym_eq(s1)are valid spellings of equality operation); in the short term, I intend to prevent excess guarding in this situation by unit testing; in the long term making the equality operators return SymBool is probably the correct fix. I modify TensorImpl to store SymBool for theThis appears to cause performance problems, so I am leaving this to an update PR.is_contiguousfields and friends onExtraMeta. In practice, this essentially meant reverting most of the changes from pytorch/pytorch#85936 . In particular, the fields on ExtraMeta are no longer strongly typed; at the time I was particularly concerned about the giant lambda I was using as the setter getting a desynchronized argument order, but now that I have individual setters for each field the only "big list" of boolean arguments is in the constructor of ExtraMeta, which seems like an acceptable risk. The semantics of TensorImpl are now that we guard only when you actually attempt to access the contiguity of the tensor via, e.g.,is_contiguous. By in large, the contiguity calculation in the implementations now needs to be duplicated (as the boolean version can short circuit, but the SymBool version cannot); you should carefully review the duplicate new implementations. I typically use theidentitytemplate to disambiguate which version of the function I need, and rely on overloading to allow for implementation sharing. The changes to thecompute_functions are particularly interesting; for most of the functions, I preserved their original non-symbolic implementation, and then introduce a new symbolic implementation that is branch-less (making use of our new SymBool operations). However,compute_non_overlapping_and_denseis special, see next bullet.- (Update: the Python side pieces for this are still in this PR, but they are not wired up until later PRs.) While the contiguity calculations are relatively easy to write in a branch-free way,
compute_non_overlapping_and_denseis not: it involves a sort on the strides. While in principle we can still make it go through by using a data oblivious sorting network, this seems like too much complication for a field that is likely never used (because typically, it will be obvious that a tensor is non overlapping and dense, because the tensor is contiguous.) So we take a different approach: instead of trying to trace through the logic computation of non-overlapping and dense, we instead introduce a new opaque operator IsNonOverlappingAndDenseIndicator which represents all of the compute that would have been done here. This function returns an integer 0 ifis_non_overlapping_and_densewould have returnedFalse, and an integer 1 otherwise, for technical reasons (Sympy does not easily allow defining custom functions that return booleans). The function itself only knows how to evaluate itself if all of its arguments are integers; otherwise it is left unevaluated. This means we can always guard on it (assize_hintwill always be able to evaluate through it), but otherwise its insides are left a black box. We typically do NOT expect this custom function to show up in actual boolean expressions, because we will typically shortcut it due to the tensor being contiguous. It's possible we should apply this treatment to all of the othercompute_operations, more investigation necessary. As a technical note, because this operator takes a pair of a list of SymInts, we need to support convertingArrayRef<SymNode>to Python, and I also unpack the pair of lists into a single list because I don't know if Sympy operations can actually validly take lists of Sympy expressions as inputs. See for example_make_node_sizes_strides - On the Python side, we also introduce a SymBool class, and update SymNode to track bool as a valid pytype. There is some subtlety here: bool is a subclass of int, so one has to be careful about
isinstancechecks (in fact, in most cases I replacedisinstance(x, int)withtype(x) is intfor expressly this reason.) Additionally, unlike, C++, I do NOT define bitwise inverse on SymBool, because it does not do the correct thing when run on booleans, e.g.,~Trueis-2. (For that matter, they don't do the right thing in C++ either, but at least in principle the compiler can warn you about it with-Wbool-operation, and so the rule is simple in C++; only use logical operations if the types are statically known to be SymBool). Alas, logical negation is not overrideable, so we have to introducesym_notwhich must be used in place ofnotwhenever a SymBool can turn up. To avoid confusion with__not__which may imply thatoperators.__not__might be acceptable to use (it isn't), our magic method is called__sym_not__. The other bitwise operators&and|do the right thing with booleans and are acceptable to use. - There is some annoyance working with booleans in Sympy. Unlike int and float, booleans live in their own algebra and they support less operations than regular numbers. In particular,
sympy.expanddoes not work on them. To get around this, I introducesafe_expandwhich only calls expand on operations which are known to be expandable.
TODO: this PR appears to greatly regress performance of symbolic reasoning. In particular, python test/functorch/test_aotdispatch.py -k max_pool2d performs really poorly with these changes. Need to investigate.
Signed-off-by: Edward Z. Yang [email protected] Pull Request resolved: pytorch/pytorch#92149 Approved by: https://github.com/albanD, https://github.com/Skylion007
finalize build.yml v20 fuck you macos build times FUCK AMFI
Settings: Refactor and clean connectivity check preference
wow this was just horrible...
Made this preference way more maintainable and runtime effective:
- Declare the preference in xml instead of hardcoding - this makes it searchable, more maintainable and is also better runtime wise
- Use arrays instead of manually naming each URL
- Use an ArrayList to handle index <-> setting relationship more easily
- Use static imports instead of literal calls
- Actually handle cases of non availability
- There is absolutely no reason to handle OnResume here
- Get rid of useless functions
- Get rid of useless imports
- Move resources to the proper custom files so we don't confuse translators
- More, too lazy. Don't write shit code please. Thank you.
finalize build.yml v21 fuck you macos build times FUCK AMFI
fix so that it downloads build time not configure time
Update sudokumain.c
Renamed "nosolution" to "no_solution" Added several probably unnecessary checks in a failed attempt at debugging the code Added a section to reset the grid to a Zero Matrix if no solution is found, however the code still hangs before reaching that point Gave use to the pointer "*p". Guess I forgot to do that when originally tackling this code. Rebuilt the Message Loop. Not sure why, that part I did a while ago and I don't remember the purpose
Notes to self: Hardcoding the "valid" int to 0 or 1 produces the "No Solution" and "Success" results, respectively. However, setting that to the isValid function makes the program commit suicide still when there is no solution. I've provided blocks to prevent the isValid from trying to process array elements that aren't on the array, but that doesn't fix it
Rebuilds Luxury Shuttle (mostly), makes it emag-only (#72940)
Changes the really goofy old lux shuttle to a cooler layout with some
additions to make it a luxury and not just
"anti-poor-people protection + food"
Shuttle was made bigger to make it less cramped for the luxury class, pool was moved to its own room, added an arcade and a bar corner, has real lasers to shoot poors with (20c each shot), has fire extinguishers now Adds a new preopen variant of hardened shutters Adds a paywall pin subtype for the luxury shuttle, and a laser gun subtype
Made emag-only at a price of 10000 credits
The old luxury shuttle looked REALLY awful with its pool, was pretty cramped even in the luxury section and BARELY resembled a luxury.. This luxury shuttle provides luxuries such as a less poorly designed pool, more space for legs, arcade, to make it resemble a luxury unlike the old one
🆑 add: Luxury Shuttle is now bigger, and less ugly! Poor people still get it rough though... /🆑
Make the build work under more JDK versions.
(Guava is already usable under plenty of verions. This change affects only people who build it themselves.)
And run CI under JDK17. Maybe this will make CI painfully slow, but we'll see what happens. If we want to drop something, we should consider whether to revert 17 or to drop 11 instead (so as to maintain coverage at the endpoints of [8, 17]).
I expected Error Prone to work, but I saw invalid flag: -Xep:NullArgumentForNonNullParameter:OFF, even though that flag is already part of the same <arg>, which works fine for other JDK versions. So I disabled Error Prone for that version.
Then I had a Javadoc problem with the --no-module-directories configuration from cl/413934851 (the fix for google/guava#5457). After reading JDK-8215582 more carefully, I get the impression that that flag might not have been added until 11: "addressed in JDK 11, along with an option to revert to the old layout in case of need." So I disabled it for 9-10.
Then I ran into a problem similar to bazelbuild/bazel#6173 / JDK-8184940. I'm not sure exactly what tool produced a file with a month of 0, but it happened only when building guava-tests. At that point, I gave up, though I left the 2 above workarounds in place.
This fails with some kind of problem finding a Guice dependency inside Maven. I didn't investigate.
These fail with the TreeMap bug that our collection testers had detected but we never got around to reporting. Thankfully, it got reported and fixed for JDK17. We could consider suppressing the tests under that version.
These fail with SecurityManager trouble.
I set up maven-javadoc-plugin to use -source ${java.specification.version}. Otherwise, it would take the version from maven-compiler-plugin. That's typically fine: Guava's source code targets Java 8, so -source 8 "ought" to work. But it doesn't actually work because we also pass Javadoc the JDK sources (so that {@inheritDoc} works better), which naturally can target whichever version of the JDK we're building with.
While Error Prone is mostly usable on JDK11+, some of its checks have problems under some versions, at least when they're reporting warnings.
This stems from its use of part of the Checker Framework, which doesn't support JDKs in the gap between 11 and 17. And specifically, it looks like the Checker Framework is trying to look up BindingPatternTree under any JDK12+. But BindingPatternTree (besides not being present at all until JDK14) didn't declare that method until JDK16.
Anyway, the problem we saw was a NoSuchMethodException during the AbstractReferenceEquality call to NullnessAnalysis.getNullness, which uses Checker Framework dataflow.
To address that, I disabled Error Prone for the versions under which I'd expect the BindingPatternTree code to be a problem.
(I also disabled it for JDK10: As noted above, Error Prone supports JDK11+. And as noted further above, Maven doesn't get far enough with JDK10 to even start running Error Prone.)
Fixes google/guava#5801
RELNOTES=n/a PiperOrigin-RevId: 488902996
[MODULAR] Fixes All Known Modular Persistence (NIF) Saving Issues (#19096)
-
Fuck
-
Holy shit
Enables strictest reasonable TypeScript settings and fixes failures.
Specific options and their reasoning include:
- Enabled
allowUnusedLabelsas labels are generally an antipattern and usually an attempt to make an object literal. - Enabled
noImplicitReturnsas forgetting to return is a very easy mistake to make and good to correct. - Enabled
noFallthroughCaseInSwitchas switch fallthroughs are a terrible idea that should never have been allowed in the first place. - Enabled
noImplicitOverrideas overriding a method should be an explicit action and clearly not accidental. - Enabled
noPropertyAccessFromIndexSignatureas I want to be explicit about statically known properties vs dynamically retrieved properties. In kind of on the fence about this one TBH, but my experience with Closure compiler makes me comfortable with this dichotomy, so it doesn't bother me. - Enabled
noUncheckedIndexAccessto require checking forundefinedwhere necessary. I'm also not entirely sold on this one, and it can be annoying to deal with, especially with arrays. However it is generally good practice and (mostly) avoidable if you're comfortable with the language.
Make compiler report structured
Full rework of the compiler message handling pipeline. Remove old-style message generation that was based purely on strings that were formatted in-place, and instead implement structured logging where main compiler code only instantiates objects with required information.
Explanation of changes for this commit will be split into several sections, matching number of edit iterations I had to make in order for this to work properly.
- File reorganization
In addition to the architectural changes this PR requires some type definition movements.
- PNode, PSym and PType definitions (and all associated types and enums) were moved to the ast_types.nim file which is then reexported in the ast.nim
- Enum defininitions for the nilability checking were moved to nilcheck_enums.nim and then reexported
- Enum definitions for the VM were moved to to vm_enums.nim
- New files
- Main definition of the report types is provided in the reports.nim file, together with minor helper procedures and definition of the ReportList type. This file is imported by options, msgs and other parts of the compiler.
- Implementation of the default error reporting is provided in the cli_reporter.nim - all disguisting hardcoded string hacks were moved to this single file. Now you can really witness the "error messages are good enough for me" thinking that prevailed in the compiler UX since it's inception.
- Changed files
Most of the changes follow very simple pattern - old-style hardcoded hacks are replaced with structural report that is constructed in-place and then converted to string /when necessary/. I'm pretty sure this change already puts me close to getting CS PHD - it was unimaginably hard to figure out that hardcoding text formatting in place is not the best architecture. Damn, I can probably apply to the nobel prize right now after I figure that out.
** Changes in message reporting pipeline
Old error messages where reportined via random combinations of the following:
- Direct calls to
msgs.liMessageproc - it was mostly used in the helper templates likelintReport messagerawMessagefatalglobalError- template for reporting global errors. Misused to the point where name completely lost it's meaning and documentation does not make any sense whatsoever. [fn:global-err]localError- template for reporting necessary error location, wrapper aroundliMessagewarningDeprecated- used two times in the whole compiler in order to print out error message about deprecated command switches.
Full pipeline of the error processing was:
- Message was generated in-place, without any colored formatting (was added in
liMessage)- There were several ways message could be generated - all of them were used interchangeably,
without any clear system.
- Some file had a collection of constant strings, such as
errCannotInferStaticParam = "cannot infer the value of the static param '$1'"that were used in thelocalReportmessage formatting immediately. Some constants were used pretty often, some were used only once. - Warning and hint messages were categorized to some extent, so and the enum was used to
provide message formatting via
array[TMsgKind, string]wherestringwas astd/strutilsformatting string. - In a lot of cases error message was generated using hardcoded (and repeatedly copy-pasted) string
- Some file had a collection of constant strings, such as
- There were several ways message could be generated - all of them were used interchangeably,
without any clear system.
- It was passed down to the
liMessage(removed now) procedure, that dispatched based on the global configuration state (whetherignoreMsgBecauseOfIdeToolsholds for example) - Then message, together with necessary bits of formatting (such as
Hintprefix with coloring) was passed down to thestyledMsgWriteln(args: varargs[typed])template, whcih did the final dispatch into - One of the two different /macros/ for writing text out -
callStyledWriteLineStderrandcallIgnoringStyle.- Dispatch could happen into stdout/stderr or message writer hook if it was non-nil
- When message was written out it went into
writeLnHookcallback (misused for{.explain.}[fn:writeln-explain]) (hacked forcompiles()[fn:writeln-compiles]) and was written out to the stdout/stderr.
It is now replaced with:
Reportobject is instantiated at some point of a compilation process (it might be an immediate report vialocalErroror delayed viankErrorand written out later)structuredReportHookconverts it to string internally. All decitions for formatting, coloring etc. happen inside of the structured report hook. Where to write message, which format and so on.writeLnHook/can/ be used by thestructuredReportHookif needed - right now it is turned into simple convenience callback. In future this could even be replaced with simple helper proc, for now I left it as it is now because I'm not 100% sure that I can safely remove this.
** Changes in the warning and hint filtering
Report filtering is done in the particular implementation of the error hook -
options.isEnabled() provides a default predicate to check whether specific report can be written
out, but it must still be called manually. This allows to still see all the reports if needed. For
example, cli_reporter.reportHook() uses additional checks to force write some reports (such as
debug trace in compiles() context).
Most of the hint and warning were already categorized, altough some reports had to be split into
several (such as --hint=Performance that actually controlled reports for CopiesToSink and
CannotMakeSink, --hint=Name that controlled three reports).
None of the errors were categorized (reports from the nkError progress was incorporated, but in
I'm mostly describing changes wrt. to old reporting system) and all were generated in-place
** Minor related changes
- List of allowed reports is now stored in the
noteSets: array[ConfNoteSet, ReportKinds]field of theConfigRef, instead of seven separate feilds. Access is now done via getter/setter procs, which allows to track changes in the current configuration state. - Renamed list of local options to
localOptions- added accessors to track changes in the state. - Move all default report filter configuration to
lineinfos.computeNotesVerbosity()- previously it was scattered in two different places:NotesVerbosityfor--verbosity:Nwas calculated inlineinfos, foreignPackageNotesDefault was constructed inoptions - Changed formatting of the
internalAssertandinternalErrormessages - Removed lots of string formatting procs from the various
sem*modules - ideally they should all be moved to thecli_reporterand related.
Additional notes
[fn:global-err] As I said earlier, globalError was misused - it is still not possible to fully get
rid of this sickening hack, since it is used for /control flow/ (you read this correct - it is an
error reporting template, and it is used for control flow. Wonderful design right?).
So, in short - globalError can raise ERecoverableError that can be caught in some other layer
(for example sem.tryConstExpr abuses that in order to bail out (pretty sure it was done as an
'optimization' of some sort, just like 'compiles') when expression fails to evaluate at
compile-time [fn:except])
[fn:except] nim-works/nimskull#94 (comment)
[fn:writeln-explain] Of course you can't have a proper error reporting in the nim compiler, so this
hook was also misused to the point of complete nonsense. Most notable clusterfuck where you could
spot this little shit is implementation of {.explain.} pragma for concepts. It was implemented via
really 'smart' (aka welcome to hell) solution in
https://github.com/nim-works/nimskull/commit/74a80988d9289e8147a791c4b0939d4287baaff3 (sigmatch
~704) and then further "improved" in
https://github.com/nim-lang/Nim/commit/fe48dd1cbec500298f7edeb75f1d6fef8490346c by slicing out parts
of the error message with let msg = s.replace("Error:", errorPrefix)
For now I had to reuse similar hack - I do override error reporting hook in order to collect all
the missing report information. In the future it would be unnecessary - when concept is matched it's
body will be evaluated to nkError with reports that are written out later.
[fn:writeln-compiles] when compiles() check is executed, all error output is temporarily disabled
(both stderr and stdout) via setting allowed output flags to {} (by default it is set to
[READY] DRAMATIC SHUTTLES!! You can now fly around the shuttle (#71906)
You can move around shuttles during transport now! Instead of them teleporting you instantly into deepspace, you can move around somewhat depending on your space-mobility and grip-strength.
Please watch the demonstration aswell, it should answer most questions: https://www.youtube.com/watch?v=Os77qDOVSXE
Interactions:
- Being within armsreach of a wall or solid object means you 'cling', where the shuttle pull is very weak and you can basically run around the shutt;e (but dont fuck up or you're gone)
- Being in range of nothing gives you a very heavy pull, you can barely resist if you have a decent jetpack
- Objects are instantly power-yeeted
- Being pulled or riding something excempts you from hyperspace pull
- Touching a space tile while being on hyperspace dumps you in deepspace, you either go back to the shuttle or enjoy deepspace
- On shuttle hyperspace tiles are a lot less dangerous, and will instead launch and freeze you instead of teleporting you into deepspace
- In-case it wasn't obvious, you can rest outside the shuttle as long as something is blocking your path. I think it's funny but I might nerf it
🆑 add: You can now fly around the shuttle during transit! Woohoo! You can either cling to the side or grab a jetpack and try and keep up with the shuttle! Carps can move around freely in hyperspace qol: Increased shuttle hyperspace size from 8 tiles to 16 /🆑
- Find a way to detect when a shuttle arrives and do something with the shit left in hyperspace
Things I will do in another PR:
- Engines spit fire and hurt (almost finished but I want to keep this small)
- Random shuttle events. You might run into dust meteors or migrating carps OR A CHANGELING INFILTRATOR
- Hyperspace turfs on the shuttle pull you under the shuttle
It's so much more immersive than being instantly teleported into deepspace. It gives you a chance to recover if you get spaced or for daredevils to look cool
It's also just very cool idk
Refactors memories to be less painful to add and apply, moves memory detail / text to memory subtypes. Adds some new memories to demonstrate. (#72110)
So, a huge issue with memories and - what I personally believe is the reason why not many have been added since their inception is - they're very annoying to add!
Normally, adding subtypes of stuff like traumas or hallucinations are as easy as doing just that, adding a subtype.
But memories used this factory argument passing method combined with holding all their strings in a JSON file which made it just frustrating to add, debug, or just mess with.
It also made it much harder to organize new memories keep it clean for stuff like downstreams.
So I refactored it. Memories are now handled on a subtype by subtype
basis, instead of all memories being a /datum/memory.
Any variety of arguments can be passed into memories like addcomponent
(KWARGS) so each subtype can have their own new parameters.
This makes it much much easier to add a new memory. All you need to do is make your subtype and add it somewhere. Don't need to mess with jsons or defines or anything.
To demonstrate this, I added a few memories. Some existing memories had their story values tweak to compensate.
Makes it way simpler to add new memories. Maybe we'll get some more fun ones now?
🆑 Melbert add: Roundstart captains will now memorize the code to the spare ID safe. add: Traitors will now memorize the location and code to their uplink. add: Heads of staff winning a revolution will now get a memory of their success. add: Heads of staff and head revolutionaries who lose their respective sides of the revolution also get a memory of their failure. add: Completing a ritual of knowledge as a heretic grants you a quality memory. add: Successfully defusing a bomb now grants you a cool memory. Failing it will also grant you a memory, though you will likely not be alive to see it. add: Planting bombs now increase their memory quality depending on how cool the bomb is. refactor: Memories have been refactored to be much easier to add. /🆑
[DNM][HACK] telephony: Force Class 0 SMS to Class 1
This kills Flash SMS messages. Fuck you airtel
Change-Id: Ifb0c9e8bae5c12868d178fbdaeceb2cc72a0ffb6
Makefile.lib: Lower kernel gzip compression to fastest
You're reading this - so you're trying to understand "JUST WHY OMG". That's already a good step.
First of all, this is a downstream kernel - always keep that in mind! Now, this kernel is targeting new very powerful Qualcomm platforms like SM8250 and the Sony Edo platform - which has a very fast UFS card.
Keep in mind that the bootloader sets the CPU at a frequency that is slightly faster than the "in the middle" ones, which is anyway not veeeery fast - but that's good, really. I agree.
So.. check this out: for Image.gz-dtb..... COMP_LEVEL SIZE 9 20116171 5 20220479 2 20940223 1 21231290
Remember again that we're loading from a UFS card and that we are loading ~1.1MB more out of a 20MB file. If you're smart enough you surely know already about RAM and CPU overhead of very high compression levels.
If you still disagree with what I just did, read this commit description another 20 times, or more, until you understand it. :)))
Signed-off-by: Tashfin Shakeer Rhythm [email protected] Signed-off-by: ImSpiDy [email protected]
Khorne Update
-
Added the stat gain on melee hit/kill 'Rampage' mechanics to more Khorne units; Juggernauts, Khorne Warptalons.
-
Standardized traits among Khorne units like Fear and Pain Immunity.
-
Added the Intimidation ability/weapon. This is primarily used for Khorne units that lack ranged weapons for reaction fire like Bloodletters, Juggernauts and Warptalons; drains TUs and Morale in a small AoE via paralytic terror.
Update books.md
There are so many books one can find on infosec, from Hacking Exposed to the Web Application Handbook. However, the books that I suggest in this pull request are of a different kind. They portray real life hacking scenarios where the reader shadows a hacker trying to break into a company. They get to experience the frustraiton, joy and excitment of a real hacking engagement. Furthermore, the techniques and tips focus on common systems and network configurations: Windows, Mainframes, Active Directory, Linux, etc. Cheers,
raphael_defconfig: Revert FBEv2 defconfig changes
To maiteiners who still use FBEv1 Fuck you in particular
https://www.youtube.com/watch?v=Ok3xVYz8Ibs
Three Idiots
Ajay, Binoy and Chandru were very close friends at school. They were very good in Mathematics and they were the pet students of Emily Mam. Their gang was known as 3-idiots. Ajay, Binoy and Chandru live in the same locality. A new student Dinesh joins their class and he wanted to be friends with them. He asked Binoy about his house address. Binoy wanted to test Dinesh's mathematical skills. Binoy told Dinesh that his house is at the midpoint of the line joining Ajay's house and Chandru's house. Dinesh was puzzled. Can you help Dinesh out? Given the coordinates of the 2 end points of a line (x1,y1) and (x2,y2), write a program to find the midpoint of the line.
power: Introduce OnePlus 3 fingerprintd thaw hack
Taken from Oneplus 3, this hack will make fingerprintd recover from suspend quickly.
Small fixes for newer kernels since we're coming from 3.10.108..
Change-Id: I0166e82d51a07439d15b41dbc03d7e751bfa783b Co-authored-by: Cyber Knight [email protected] [cyberknight777: forwardport and adapt to 4.14] Signed-off-by: Shreyansh Lodha [email protected] Signed-off-by: Pierre2324 [email protected] Signed-off-by: PainKiller3 [email protected] Signed-off-by: Dhruv [email protected] Signed-off-by: Cyber Knight [email protected] Signed-off-by: Blaster4385 [email protected]
Created Text For URL [www.iol.co.za/news/news/crime-and-courts/double-life-sentence-for-kzn-man-who-killed-his-ex-girlfriend-and-her-new-boyfriend-while-they-were-asleep-649ef8dc-e892-40d4-991a-49d3c87752d0]
Initial implementation of Divine Prayer Altar
This commit add a prayer system for Blessed Blood quirk users to request blessings.
The following changes are made:
- Added the Divine Prayer Altar
- Added CSS class blessedphrase
- Added CSS animation blessedcolor
- Added span define span_blessedphrase
- Added admin topic for blessed requests
- Removed Blessed Blood rewards for holy points of 3+
A learning experience I'm sure all programmers go through...
After a machine restore and cloning the repo back from GitHub it became apparent that all local files previously present but assigned to .gitignore had been lost - including the OriginalAppBuild folder which included the original application from which this unfinished one was built.
Well, obviously - I told it to ignore them didn't I...
The absolute kicker? I deleted the backup while cleaning up my documents folder on autopilot before restoring my machine...
A classic example of being reminded the hard way that computers do EXACTLY what you tell them to and that having backups are only good if you keep them somewhere you wont touch them by mistake.
The only upshots here are that my written work was in a totally separate location, and the actual code itself is of course exactly how I Ieft it - Praise the Omnissiah.
This commit recreates a documents folder and adds the base files for some documents I can recreate easily like the .drawio files. Added the assignment brief and some files for code snips and pseudocode writing for the assignment itself. The OriginalAppBuild is completely gone - there's no way I can reverse engineer what the code has now turned into back into what it was. Be like turning an omelette back into eggs.
🪁🥎 ↝ #8 Add file upload feature & auth handler
- Completed authentication header for web client. To the end user it is 100% offchain, with user profiles being stored on a postgresql database. However, I've taken a dive into the Magic sdk to create wallet addresses for each user, as well as a Flask-based authentication handler for future metamask/wallet interaction. I've made this decision for a few reasons (like simplicity), but the main reason is for the client to seem like a regular journal platform and not be confusing, as well as follow the 'web3-agnostic' design language I favour for projects like this due to confusion and/or distrust of web3 products/teams. However, since each user will have a wallet address, they'll be able to interact with smart contracts and IPFS just fine. Further discussion will need to take place to discuss long-term suitability of this model, including things like gas fees (currently everything regarding transactions is occurring on Goerli [testnet] and gas fees will be processed by a "superuser" so that there's no restrictions or huge expenses) and how we go about getting users to trust the web3 nature (which I've got a lot of experience with). However, I don't fully know the exact demographics we'll be targeting & also I understand that that's quite a while away, so I'll raise it now but won't spend any time thinking about it until the time comes 2. I've continued with the contracts for proposals/publications & updating the metadata standards. I favour a lazy minting approach with data processing being handled by a Flask blueprint (which is a formula my team have developed on signal-k/sytizen). Right now I'm using Thirdweb & Moralis for the contract interactions and I have also, with much difficulty, succeeded in getting Moralis to self-host on my Postgres server. Finally, I've begun the process of optimising the base layer contracts so that the gas fees (which are already reduced post-merge) are essentially negligible at this time. 3. File upload for posts/articles feature on the web client is complete, and the smart contracts now receive all file upload metadata from this. 4. Begun a new flask blueprint (forked from point #2) to generate dataset previews based on which modules (e.g. lightkurve) are used and to add interactive nature to the 'sandbox' feature discussed earlier 5. Reluctantly continued some documentation
(above message from the Desci discord, see Signal-K/sytizen#16 for more info)
more span macro changes in brain traumas and disease files (#73273)
i was fucking around with brain traumas on a downstream and noticed they had similar issues to quirks so i decided to continue work from #73116
(search in VSC for span class = 'notice')
its going to be a bit of a thing to get all of these but this is a
decent chunk i think
there was only one annoying/tough file. imaginary_friend.dm had class = 'game say' and class = 'emote' both of which after some testing did not seem like they did anything. ill try to keep that in mind in other files if i continue to do this but i either omitted them because they didnt have any formatting or, in the case of emote, turned it into name, which i think is what you'd want those messages to look like.
there were also a few small spelling errors that i fixed
more consistent and stops people from copying brain trauma formatting wrong
they should all work the same
Co-authored-by: san7890 [email protected]
[FIX] website: redirect to case insensitive URL if not exact match
Before this commit, if a link to a page was not correct because of a case mismatch, it would simply land on a 404 page. While it's correct, as URL are case sensitive, it leads to a few bad UX flow at the admin/editor level:
- Create a link in your page (on a text or a button eg), type an URL which does not exists (to create it after) like /Page
- Click on the link/button you just made, you are redirected to /Page which display a 404 with the "Create page" option (correct)
- When you click on that button, it will actually create a page with /page URL, leading to a mismatch between the URL you created and the page URL. Your link/button will still lead to a 404 URL as it points to /Page.
Since it's just a fallback when an exact URL match is not found, it should not break anything and should not have bad impact at any level (seo/speed etc). Indeed:
- It's done through a 302 redirect
_serve_page()is already a fallback case, so it will only make thewebsite.redirectand 404 cases a bit slower due to the extra search query.
The only possible scenario seems to be if the user (mind the uppercase):
- Created a /Page page
- Created a redirect from /page to /another-page
In this case, /page won't land on /another-page but on /Page. This flow seems unlikely and is not actually wrong either way. At least, it certainly is less important than ensuring a case insensitive fallback.
Finally, note that another solution would have been to either:
- Force page URL to lower case.
-> This is not stable friendly, people might be relying on this to
create pages with different casing:
/Batman-VII-The-Dark-Knight-Whatevers, while not recommended, doesn't sounds idiot. On top of not being stable friendly, we probably want to keep offering this possibility - Redirect all URLs to lowercase endpoints. -> This is obviously not stable and not Odoo's jobs. It should be something decided by the sysadmin and done at nginx (etc) level.
task-3110294 opw-3104030
closes odoo/odoo#111400
X-original-commit: 639cfc76ba259eea8f38284192017024809173b3 Signed-off-by: Quentin Smetz (qsm) [email protected] Signed-off-by: Romain Derie (rde) [email protected]
Flash
A young man named Diffny leaves home to travel to California, to join the Team Flash. Although Diffny is not able to join this elite team immediately, he befriends the three most formidable members of the age: Joe, Ramsey and vixon and gets involved in affairs of the state and court.At that time, the Villan was planning to dethrone the king and to take the kingdom and to remove the Team Flash of the guard. Since the Villan has spies mixed with the local public , Diffny decides to send a message of his whereabouts to the team Flash in unique way.He gave a note to a boy which has the following message. I am at the midpoint of the line joining the farmhouse next to the palace and the light house. The Team Flash were puzzled. Can you help them find out the location of Diffny?Given the coordinates of the 2 places (x1,y1) and (x2,y2), write a program to find the location of Diffny.
Input & Output Format:
Input consists of 4 integers.
First value consists of x1.
Second value consists of y1.
Third value consists of x2.
Fourth value consists of y2.
Output consists of two float values.
Sample Input 2
4
10
15
Sample Output
6.0 9.5
[MIRROR] Basic mobs don't become dense upon death [MDB IGNORE] (#18679)
- Basic mobs don't become dense upon death (#72554)
In #72260 what was previously a var became a flag, which was a sensible change, however this inverted the default behaviour. In virtually all cases we want dead mobs to stop being dense, this added a requirement for the flag to be present for that to happen and then didn't add the flag to any mobs.
Rather than add this to every mob I inverted the function of the flag.
My reasoning here is that simple mobs seemingly never required this
behaviour, basic mobs are probably going to need it rarely if ever, and
including it in basic_mob_flags by default seems messy and easy to
leave off when setting other flags (plus #72524 implies to me we want to
avoid adding more default values).
Setting this manually on each mob seems kind of silly as a requirement going forward and I can't think of a way we'd unit test for people forgetting.
For the same reason I did the same thing with the
STOP_ACTING_WHILE_DEAD flag I added to the AI controller in a recent
PR, the flag should denote unusual behaviour not the default.
It looks really odd when you're constantly shuffling places with dead
mobs, they're not supposed to do that.
It's tedious to add STOP_ACTING_WHILE_DEAD to every AI controller when
that should be an obvious default assumption.
🆑 fix: Dead basic mobs are no longer "dense" objects and can be stepped on. /🆑
-
Basic mobs don't become dense upon death
-
Removes a flag we didn't need anymore.
-
Forgot to remove this one
Co-authored-by: Jacquerel [email protected] Co-authored-by: GoldenAlpharex [email protected] Co-authored-by: GoldenAlpharex [email protected]
Merge pull request #34 from RizeKishimaro/login-rs
I hate myself, Fuck
[FIX] stock,purchase,mrp: accumulative security days
Usecase to reproduce:
- Set the warehouse as 3 steps receipt
- Put a security delay of 3 days for purchase
- Set a product with a vendor and 1 days as LT
- Replenish with the orderpoint
You expect to have a schedule date for tomorrow that contains all the product needed in the incoming 4 days.
Currenly the internal transfer from QC -> Stock is for tomorrow (ok). The transfer from Inpur -> QC is plan for 2 days in the past. (not ok) The PO date is plan for 5 days in the past. (not ok)
It happens because the system check at each stock.rule application if
purchase is part of the route. If it's then it applies the security lead
time. It's a mistake because we should apply it only the first time.
To fix it we directly set it when the orderpoint run and not during
stock.move creation.
However for MTO it's not that easy. We don't want to deliver too
early the customer. So we keep applying the delay during the
stock.move creation but only when it goes under the warehouse stock
location.
Expose filtering by file type in glob (#7834)
Add flags for filtering the output of glob by file type. I find myself
occasionally wanting to do this, and getting a file's
file_type
is presumably fast to do as it doesn't have to go through the fallible
metadata method.
The design of the signature does concern me; it's not as readable as a
filter or "include" type list would be. They have to be filtered one by
one, which can be annoying if you only want files -D -S, or only want
folders -F -S, or only want symlinks --butwhy?. I considered
SyntaxShape::Keyword for this but I'll just defer to comments on this PR
if they pop up.
I'd also like to bring up performance since including these flags
technically incurs a .filter penalty on all glob calls, which could be
optimized out if we added a branch for the no-filters case. But in
reality I'd expect the file system to be the bottleneck and the flags to
be pretty branch predictor friendly, so eh
Three new flags when using glob and a slightly more cluttered help
page. No breaking changes, I hope.
Don't forget to add tests that cover your changes.
Make sure you've run and fixed any issues with these commands:
cargo fmt --all -- --checkto check standard code formatting (cargo fmt --allapplies these changes)cargo clippy --workspace -- -D warnings -D clippy::unwrap_used -A clippy::needless_collectto check that you're using the standard code stylecargo test --workspaceto check that all tests pass
If your PR had any user-facing changes, update the documentation after the PR is merged, if necessary. This will help us keep the docs up to date.
includes today's "speed: run and removes an interval pack run
Today went poorly. In retrospect, I should not have worn my ruck for the nature hike I did with Marcia and Jeff on the 2nd. I also shouldn't have done the second AA loop I did yesterday. I probably should have also done my interval training without my ruck on Monday. Oh, and I shouldn't have worn my ski boots for a few hours the other day, either.
At this point it looks like I actually have a blister where my hot spot is. That may be a good sign, because a blister is much more likely to resolve relatively quickly than a morton's neuroma. OTOH, it was quite uncomfortable today and that was with 400mg ibuprofen and 500mg acetaminophen, although the discomfort lessened toward the end.
In addition to being slowed down by my right foot, I also got off to a late start which caused me to have to come to a complete stop for a while twice at the intersection for the main entrance to Albuquerque Academy. It's hard for me to tell just looking at my Strava information, and I have too many other things to do than to write some software to tease out the data, but it looks like I lost about thirty seconds at each stop, but by giving my foot a rest I may have allowed myself to run faster later, so I can't just take 6 seconds off my average pace to make up for the lost time.
In light of my foot discomfort, I'm going to make some changes. I'm going to run less running (perhaps none) tomorrow and Friday. I'm not going to do pack intervals Monday. Instead, I'll probably drive to the library, do unweighted intervals and drive back. Sadly, Tough Ruck is looking unlikely this year, and I probably won't decide on Cocodona 250 until after Bataan Memorial.
hash-object: use fsck for object checks
Since c879daa237 (Make hash-object more robust against malformed objects, 2011-02-05), we've done some rudimentary checks against objects we're about to write by running them through our usual parsers for trees, commits, and tags.
These parsers catch some problems, but they are not nearly as careful as the fsck functions (which make sense; the parsers are designed to be fast and forgiving, bailing only when the input is unintelligible). We are better off doing the more thorough fsck checks when writing objects. Doing so at write time is much better than writing garbage only to find out later (after building more history atop it!) that fsck complains about it, or hosts with transfer.fsckObjects reject it.
This is obviously going to be a user-visible behavior change, and the test changes earlier in this series show the scope of the impact. But I'd argue that this is OK:
-
the documentation for hash-object is already vague about which checks we might do, saying that --literally will allow "any garbage[...] which might not otherwise pass standard object parsing or git-fsck checks". So we are already covered under the documented behavior.
-
users don't generally run hash-object anyway. There are a lot of spots in the tests that needed to be updated because creating garbage objects is something that Git's tests disproportionately do.
-
it's hard to imagine anyone thinking the new behavior is worse. Any object we reject would be a potential problem down the road for the user. And if they really want to create garbage, --literally is already the escape hatch they need.
Note that the change here is actually in index_mem(), which handles the HASH_FORMAT_CHECK flag passed by hash-object. That flag is also used by "git-replace --edit" to sanity-check the result. Covering that with more thorough checks likewise seems like a good thing.
Besides being more thorough, there are a few other bonuses:
-
we get rid of some questionable stack allocations of object structs. These don't seem to currently cause any problems in practice, but they subtly violate some of the assumptions made by the rest of the code (e.g., the "struct commit" we put on the stack and zero-initialize will not have a proper index from alloc_comit_index().
-
likewise, those parsed object structs are the source of some small memory leaks
-
the resulting messages are much better. For example:
[before] $ echo 'tree 123' | git hash-object -t commit --stdin error: bogus commit object 0000000000000000000000000000000000000000 fatal: corrupt commit
[after] $ echo 'tree 123' | git.compile hash-object -t commit --stdin error: object fails fsck: badTreeSha1: invalid 'tree' line format - bad sha1 fatal: refusing to create malformed object
Signed-off-by: Jeff King [email protected] Signed-off-by: Junio C Hamano [email protected]
[MIRROR] Abductor scientist self-retrieve failure/runtime fix [MDB IGNORE] (#19179)
- Abductor scientist self-retrieve failure/runtime fix (#73172)
Since the abductor outfit/implant would load before the abductor ship (and it's teleport pad) when first generating a team, a runtime would occur when trying to link the pad to the implant. Another would occur every time you attempted to retrieve yourself (as the linked pad would be null), preventing recall and completely neutering an abductor team's most important maneuver.
Now, using the implant will perform the linking process again if no linked pad is found, and provides the owner with a warning if (by some great calamity) they genuinely have no pad to teleport back to. This solves the issue of the implant sometimes not linking to a pad properly on initialize, and makes them way less prone to breaking.
Apparently this has been broken for a while, presumably since the abductor ship was made into a lazyloading template.
The funny silly grey men get to torture the poor hapless crew once again.
🆑 fix: abductor scientist's retrieval implants will now properly recall the owner, and inform them upon recall failure. /🆑
- Abductor scientist self-retrieve failure/runtime fix
Co-authored-by: Rhials [email protected]
huge update, many fixes
-
refactor the LHS tokenizer to handle possibilty of both Targets, Assignment, or plain Expression. The LHS tokenizer will now store the whitespace as individual Literals. When the statement has been tokenized far enough to determine Rule, Assignment, or Expression, then we'll deal with the whitespace.
- whitespace is discarded in Targets
- whitespace is preserved until the parse phase
- if the statement is a directive, the first whitespace after the directive is discarded otherwise, all whitespace is preserved.
-
the LHS tokenize now returns a raw token_list for the LHS. The calling tokenizer/parser will create the Target or Assignment or Expression
-
I successfully removed two of the most ugly methods from vline: rstrip and truncate. I originally used those methods to truncate a Rule line with a dangling Recipe. Now, with a more updated Recipe tokenizer, I can pass the remaining vchars to the recipe tokenizer directly, no more need to do horrible ugly hacks directly into the VCharString.
-
added a temporary _view() fn for interactive debugging. At the pdb prompt, use print(_view(array-of-vchar)) to see the array-of-vchar as a string.
windows: ignore empty PATH elements
When looking up an executable via the _which function, Git GUI
imitates the execlp() strategy where the environment variable PATH
is interpreted as a list of paths in which to search.
For historical reasons, stemming from the olden times when it was uncommon to download a lot of files from the internet into the current directory, empty elements in this list are treated as if the current directory had been specified.
Nowadays, of course, this treatment is highly dangerous as the current
directory often contains files that have just been downloaded and not
yet been inspected by the user. Unix/Linux users are essentially
expected to be very, very careful to simply not add empty PATH
elements, i.e. not to make use of that feature.
On Windows, however, it is quite common for PATH to contain empty
elements by mistake, e.g. as an unintended left-over entry when an
application was installed from the Windows Store and then uninstalled
manually.
While it would probably make most sense to safe-guard not only Windows
users, it seems to be common practice to ignore these empty PATH
elements only on Windows, but not on other platforms.
Sadly, this practice is followed inconsistently between different software projects, where projects with few, if any, Windows-based contributors tend to be less consistent or even "blissful" about it. Here is a non-exhaustive list:
Cygwin:
It specifically "eats" empty paths when converting path lists to
POSIX: https://github.com/cygwin/cygwin/commit/753702223c7d
I.e. it follows the common practice.
PowerShell:
It specifically ignores empty paths when searching the `PATH`.
The reason for this is apparently so self-evident that it is not
even mentioned here:
https://learn.microsoft.com/en-us/powershell/module/microsoft.powershell.core/about/about_environment_variables#path-information
I.e. it follows the common practice.
CMD:
Oh my, CMD. Let's just forget about it, nobody in their right
(security) mind takes CMD as inspiration. It is so unsafe by
default that we even planned on dropping `Git CMD` from Git for
Windows altogether, and only walked back on that plan when we
found a super ugly hack, just to keep Git's users secure by
default:
https://github.com/git-for-windows/MINGW-packages/commit/82172388bb51
So CMD chooses to hide behind the battle cry "Works as
Designed!" that all too often leaves users vulnerable. CMD is
probably the most prominent project whose lead you want to avoid
following in matters of security.
Win32 API (CreateProcess())
Just like CMD, `CreateProcess()` adheres to the original design
of the path lookup in the name of backward compatibility (see
https://learn.microsoft.com/en-us/windows/win32/api/processthreadsapi/nf-processthreadsapi-createprocessw
for details):
If the file name does not contain a directory path, the
system searches for the executable file in the following
sequence:
1. The directory from which the application loaded.
2. The current directory for the parent process.
[...]
I.e. the Win32 API itself chooses backwards compatibility over
users' safety.
Git LFS:
There have been not one, not two, but three security advisories
about Git LFS executing executables from the current directory by
mistake. As part of one of them, a change was introduced to stop
treating empty `PATH` elements as equivalent to `.`:
https://github.com/git-lfs/git-lfs/commit/7cd7bb0a1f0d
I.e. it follows the common practice.
Go:
Go does not follow the common practice, and you can think about
that what you want:
https://github.com/golang/go/blob/go1.19.3/src/os/exec/lp_windows.go#L114-L135
https://github.com/golang/go/blob/go1.19.3/src/path/filepath/path_windows.go#L108-L137
Git Credential Manager:
It tries to imitate Git LFS, but unfortunately misses the empty
`PATH` element handling. As of time of writing, this is in the
process of being fixed:
https://github.com/GitCredentialManager/git-credential-manager/pull/968
So now that we have established that it is a common practice to ignore
empty PATH elements on Windows, let's assess this commit's change
using Schneier's Five-Step Process
(https://www.schneier.com/crypto-gram/archives/2002/0415.html#1):
Step 1: What problem does it solve?
It prevents an entire class of Remote Code Execution exploits via
Git GUI's `Clone` functionality.
Step 2: How well does it solve that problem?
Very well. It prevents the attack vector of luring an unsuspecting
victim into cloning an executable into the worktree root directory
that Git GUI immediately executes.
Step 3: What other security problems does it cause?
Maybe non-security problems: If a project (ab-)uses the unsafe
`PATH` lookup. That would not only be unsafe, though, but
fragile in the first place because it would break when running
in a subdirectory. Therefore I would consider this a scenario
not worth keeping working.
Step 4: What are the costs of this measure?
Almost nil, except for the time writing up this commit message
;-)
Step 5: Given the answers to steps two through four, is the security measure worth the costs?
Yes. Keeping Git's users Secure By Default is worth it. It's a
tiny price to pay compared to the damages even a single
successful exploit can cost.
So let's follow that common practice in Git GUI, too.
Signed-off-by: Johannes Schindelin [email protected]
Auto merge of #101986 - WaffleLapkin:move_lint_note_to_the_bottom, r=estebank
Move lint level source explanation to the bottom
So, uhhhhh
r? @estebank
"note: #[warn(...)] on by default" and such are moved to the bottom of the diagnostic:
- = note: `#[warn(unsupported_calling_conventions)]` on by default
= warning: this was previously accepted by the compiler but is being phased out; it will become a hard error in a future release!
= note: for more information, see issue #87678 <https://github.com/rust-lang/rust/issues/87678>
+ = note: `#[warn(unsupported_calling_conventions)]` on by defaultWhy warning is enabled is the least important thing, so it shouldn't be the first note the user reads, IMO.
struct_span_lint and similar methods have a different signature.
Before: ..., impl for<'a> FnOnce(LintDiagnosticBuilder<'a, ()>)
After: ..., impl Into<DiagnosticMessage>, impl for<'a, 'b> FnOnce(&'b mut DiagnosticBuilder<'a, ()>) -> &'b mut DiagnosticBuilder<'a, ()>
The reason for this is that struct_span_lint needs to edit the diagnostic after decorate closure is called. This also makes lint code a little bit nicer in my opinion.
Another option is to use impl for<'a> FnOnce(LintDiagnosticBuilder<'a, ()>) -> DiagnosticBuilder<'a, ()> altough I don't really see reasons to do let lint = lint.build(message) everywhere.
By moving the message outside of the closure (that may not be called if the lint is disabled) format!(...) is executed earlier, possibly formatting Ty which may call a query that trims paths that crashes the compiler if there were no warnings...
I don't think it's that big of a deal, considering that we move from format!(...) to fluent (which is lazy by-default) anyway, however this required adding a workaround which is unfortunate.
I'm sorry, I do not how to make this PR smaller/easier to review. Changes to the lint API affect SO MUCH 😢
Fix operation synchronization
PBENCH-993
This is fairly large, but yet much smaller than it started out. This solves two problems in Pbench Server task scheduling:
- The default SQLAlchemy transaction model is to start a transaction on any SELECT and end it on any UPDATE or INSERT; this means there's no potential for atomic update. This impacts the management of all internal context, but serious problems have been observed with the "operation" and "state" of the datasets.
- I began the dataset with the concept of a "state", as the dataset progresses through upload, backup, unpack, index, and delete. I quickly discovered that this wasn't ideal, but had trouble backing off completely from the concept. When I added the DB-based "operation" to replace the old filesystem links, the relationship between "operation" and "state" became even messier.
The primary change here is to divorce the Sync class entirely from general
metadata. (I originally set out to make Metadata management atomic, and the
fan-out was enormous: I'll tackle that again later, but while important in the
long run, getting Sync working is immediately critical.)
There is a new Operation DB object associated with a Dataset, and this is
entirely managed within the Sync class. For visibility, Dataset.as_json
will collect associated rows just as it does for dataset.metalog, but this
doesn't require any special transactional management. (It's a snapshot.)
I've completely removed the Dataset.state column (and its associated "last
transition" timestamp). "State" is tracked and observed by the various Operation
rows created and managed by Sync. This corresponds to the previous
dataset.statussub-object managed by the oldSync: each named operation (OperationNameenum) that's been associated to a dataset will have a row withstateandmessagecolumns. Thestatecan be advanced throughREADY, WORKING, OK, and FAILED, and a message can be associated with each row (on error via Sync.erroror as part of transition viaSync.update`).
While I was modifying the Dataset schema, I also removed the created column;
it's really redundant with dataset.metalog.pbench.date, and we'll need to
understand that for "non-Pbench-standard" tarballs we might not be able to get
this anyway. This wasn't quite as "easy" as I'd thought, because it meant that
I had to refactor date-range operations to work on uploaded (perhaps they
should have been that way originally).
Sync.__init__: Construct a sync object for a particular named operation.
Sync.next: Return a list of datasets which have Operation rows for the
sync component that are in READY state.
Sync.update: Change the state of the sync component's Operation,
optionally add a message to that Operation, and set a list of named
operations for that dataset to READY.
Sync.error: Change the state of the sync component's Operation to FAILED
and record an explanatory message for the failure.
To avoid rippling massive SQLAlchemy transaction model changes across all our
code, I've added a second session factory in Database with the most strict
integrity level, SERIALIZABLE. (In fact, I think that "REPEATABLE READ"
would be enough, and slightly more efficient, but sqlite3 doesn't support that
and I don't think I want to trust the weaker model it does support.)
Only the Sync class in sync.py currently uses that alternate session
factory. To avoid conflicts and confusions with autoflush and autocommit and
other SQLAlchemy "help", I'm creating new sessions on-the-fly for each call
and retiring them after commit/rollback. Note that the idiom
with Database.maker.begin() as session: constructs a new session with fresh
state, allows a sequence of session operations, and then implicitly tears down
the session after it commits on success or rolls back on an exception.
To avoid multiple SELECT statements within our transaction, Sync.update
fetches all relevant rows in a single SELECT, and then organizes them for
selective updates. This ensures we have no SELECT following the update of any
proxy object, which confuses SQLAlchemy in normal configuration.
Sync.update and Sync.error will loop up to 10 times on commit failure to
re-try with fresh data. Note that I've observed the retry logic in dozens of
functional test runs; and while I wonder vaguely whether I should be concerned
with the constant 10, I rarely see more than one or two retries since I added
a small delay to minimize thrashing.
shit got fucked up i literally hate this thing so much
"1:35pm. I do not know whether it is the experience in concurrent programming, or the focus of having the goal of learning webdev directly, or if the docs are simply better now, but I am making huge progress right now. I am soaking up the knowledge like a spoonge.
Breakfast time.
2:35pm. Done with breakfast. I want to chill a bit longer and then I will resume.
3:05pm. Let me resume.
Now that I think about it, I have no idea how how the data transfer between the server and client works. By that I mean, the number of requests and such. Now that I understand middleware, I understand that the app just replies to requests. But surely it does not transfer the web page in its entirety all at once. They should be back and forth talk as html, images and JS are transfered separately, right?
I am going to have to come to an understanding of those basics at some point.
3:10pm. Focus me, let turn off Youtube music.
Focus on reading this article and then the other. After that I might want to watch sometihng about the basics of how server client communication works.
For more information, see Write custom ASP.NET Core middleware.
I need to remember this link.
public class Operation : IOperationTransient, IOperationScoped, IOperationSingleton
{
public Operation()
{
OperationId = Guid.NewGuid().ToString()[^4..];
}
public string OperationId { get; }
}I did not know that it was possible to get the last 4 elems of an array in C# like this.
3:20pm. Focus me, stop reading /g/. Get into it.
public class IndexModel : PageModel
{
private readonly ILogger _logger;
private readonly IOperationTransient _transientOperation;
private readonly IOperationSingleton _singletonOperation;
private readonly IOperationScoped _scopedOperation;
public IndexModel(ILogger<IndexModel> logger,
IOperationTransient transientOperation,
IOperationScoped scopedOperation,
IOperationSingleton singletonOperation)
{
_logger = logger;
_transientOperation = transientOperation;
_scopedOperation = scopedOperation;
_singletonOperation = singletonOperation;
}
public void OnGet()
{
_logger.LogInformation("Transient: " + _transientOperation.OperationId);
_logger.LogInformation("Scoped: " + _scopedOperation.OperationId);
_logger.LogInformation("Singleton: " + _singletonOperation.OperationId);
}
}I have to admit this is really informative, because I'd never think of a shitty design pattern like this one myself.
3:35pm. I need to get used to DI. It is creating objects based on their runtime types.
Sigh, spare me. It really has a ton of these. But I get it. I really get it now.
I think last time I didn't really for some reason. But DI is a pervasive pattern in OO programming.
Now comes middleware.
The request handling pipeline is composed as a series of middleware components. Each component performs operations on an HttpContext and either invokes the next middleware in the pipeline or terminates the request.
///
The WebApplicationBuilder.Build method configures a host with a set of default options, such as:
Use Kestrel as the web server and enable IIS integration. Load configuration from appsettings.json, environment variables, command line arguments, and other configuration sources. Send logging output to the console and debug providers.
///
So it is loading from appsettings.json in this step? I've been wondering about this for a long time.
It keeps talking about IIS.
Kestrel is often run in a reverse proxy configuration using IIS.
What is a reverse proxy configuration? Does this mean I do not have to use it with IIS?
A reverse proxy server is a type of proxy server that typically sits behind the firewall in a private network and directs client requests to the appropriate backend server. A reverse proxy provides an additional level of abstraction and control to ensure the smooth flow of network traffic between clients and servers.
During development, the content root defaults to the project's root directory. This directory is also the base path for both the app's content files and the Web root. Specify a different content root by setting its path when building the host. For more information, see Content root.
It really answers all my answers here.
https://learn.microsoft.com/en-us/aspnet/core/tutorials/min-web-api?source=recommendations&view=aspnetcore-7.0&tabs=visual-studio Tutorial: Create a minimal API with ASP.NET Core
I'll go over this, but first let me just take a peek at the 3.1 documentation.
https://learn.microsoft.com/en-us/aspnet/core/fundamentals/middleware/?view=aspnetcore-3.1
Yeah, it is a bit worse, but all the aspects are still there. I was simply retarded. And in this tutorial, remember the JS no-batteries servers that I looked into. Yeah, it is the same thing.
4:10pm. https://github.com/DapperLib/Dapper
I guess Dapper and EF are the most popular database libraries in the .NET world. This one has nearly 16k stars. 3.6k forks, now that is something.
Yeah, I've been foolish. I should started my ASP.NET journey here. And then either looked into Blazor or functional frameworks, skipping over the controllers. Although to be honest, I am embarassed that I can't understand MVC properly.
https://learn.microsoft.com/en-us/aspnet/core/web-api/?view=aspnetcore-7.0
Wait, this isn't MVC.
4:20pm. Forget this. I hate programming with attributes. DI is just barely palatable to me, but this I don't like. I think functional programming will have a better response than attributes.
...Now...
Let me think.
4:25pm. I'll figure out how to run F# projects in different configurations.
I forgot how running stuff in VS Code works. Forget that for now.
You really add a lot of complications when you get away from the IDE.
4:30pm. I think I have the basics of ASP.NET down now. Let me just take a look at the Middleware page for 7.0.
https://learn.microsoft.com/en-us/aspnet/core/fundamentals/middleware/?view=aspnetcore-7.0
Ok, I am going to make an assumpting. For a particular page, just the HTML is sent back. And then the browser has the option of getting the multimedia files separately.
https://learn.microsoft.com/en-us/aspnet/core/fundamentals/middleware/?view=aspnetcore-7.0
The Routing middleware in the preceding diagram is shown following Static Files. This is the order that the project templates implement by explicitly calling app.UseRouting. If you don't call app.UseRouting, the Routing middleware runs at the beginning of the pipeline by default. For more information, see Routing.
4:40pm. I a bit confused as I read this:
app.UseResponseCaching();
app.UseResponseCompression();
app.UseStaticFiles();I think this would in fact cause the response to be written to multiple times. How does that work?
I do not get most of these middleware.
4:55pm. But that doesn't matter. Let me take a break here. I get the gist of what the framework is about now. It is pretty modular with DI and middleware. But middleware itself is nothing special. This is how all the JS frameworks that I've seen work. I forgot what the name of those servers were, but they had custom routing. In the minimal example, it was shown how to do exactly that.
5pm. Now that I have all this fundamental knowledge, it isn't such a mystery as to how ASP.NET would interact with...or rather it isn't a mystery of how functional and other kinds of frameworks could be built on top of it.
5:05pm. I can build a server with this no problem.
I just need some kind of remoting solution for the client and I am set. And I've already seen an example of that in the SAFE Dojo.
I'll also be using that in the future when I make something in Python and want to communicate to the .NET land.
Now...I wanted to say, that I need more knowledge of the existing DI services that ASP.NET library offers. I also need more knowledge of existing middleware. Right now I barely know anything about them. I certainly can't claim that I know how to build a significant website - something on the level of Mangadex yet.
But that is just a matter of learning as I go along.
It would be pointless to just dive in and read up each and every service as if I were memorizing an encyclopedia.
5:05pm. Right now I am in great shape.
What I should do here is go back to studying the SAFE stack. Apart from more of the samples, the Elmish library is the next in line. I'll just study the examples for a few days until my curiosity is stated and then start work on my own project.
5:10pm. Let me take a break here. I do not feel like starting right away.
https://manonepunch.com/manga/one-punch-man-chapter-173/
I am a bit behind on OPM so let read this for a bit. Then I'll get started on the Elmish samples.
5:25pm. https://chapmanganato.com/manga-wd951838/chapter-174
Just this chapter and I'll leave the rest for later.
5:35pm. Enough of that. Let me check out the Elmist examples.
https://elmish.github.io/elmish/
https://elmish.github.io/elmish/docs/basics.html
#r "Fable.Elmish.dll"
Ugh, what do I do about this?
Let me sidetrack a bit. What are the new features of F#. The ones I've been using are quite old.
https://devblogs.microsoft.com/dotnet/whats-new-in-fsharp-6/
let readFilesTask (path1, path2) =
task {
let! bytes1 = File.ReadAllBytesAsync(path1)
let! bytes2 = File.ReadAllBytesAsync(path2)
return Array.append bytes1 bytes2
}Interesting.
Up to and including F# 5.0, F# has used expr.[idx] as indexing syntax. This syntax was based on a similar notation used in OCaml for string indexed lookup. Allowing the use of expr[idx] is based on repeated feedback from those learning F# or seeing F# for the first time that the use of dot-notation indexing comes across as an unnecessary divergence from standard industry practice. There is no need for F# to diverge here.
Ok.
> [<return: Struct>]
let (|Int|_|) str =
match System.Int32.TryParse(str) with
| true, int -> ValueSome(int)
| _ -> ValueNoneStruct active patterns. This is great.
type Pair = Pair of int * int
let analyzeObject (input: obj) =
match input with
| :? (int * int) as (x, y) -> printfn $"A tuple: {x}, {y}"
| :? Pair as Pair (x, y) -> printfn $"A DU: {x}, {y}"
| _ -> printfn "Nope"
let input = box (1, 2)In F# 6, we’ve added a new declarative feature that allows code to optionally indicate that lambda arguments should be inlined at callsites.
This is one of the tricks from Spiral.
let inline iterateTwice ([<InlineIfLambda>] action) (array: 'T[]) =
for j = 0 to array.Length-1 do
action array[j]
for j = 0 to array.Length-1 do
action array[j]Unlike previous versions of F#, this optimization applies regardless of the size of the lambda expression involved. This feature can also be used to implement loop unrolling and similar transformations reliably.
Why did I even make Spiral?
In F# 6, the compiled form of generative list and array expressions is now up to 4x faster (see the relevant pull request). Generative list and array expressions also have greatly improved debugging. For example:
NOTE: High-performance modern tensor implementations are available through TorchSharp and TensorFlow.NET.
Has somebody done PyTorch wrappers for .NET?
One of the most anticipated features in the F# 6 toolchain is the addition of “Pipeline debugging”. Now, the F# compiler emits debug stepping points for each position in an F# pipeline involving |>, ||> and |||> operators. At each step, the input or intermediate stage of the pipeline can be inspected in a typical debugger.
Interesting.
5:50pm.
PS C:\Users\Marko> dotnet fsi
Microsoft (R) F# Interactive version 12.4.0.0 for F# 7.0
Copyright (c) Microsoft Corporation. All Rights Reserved.
For help type #help;;
>
Oh, I can enter the F# FSI like this. Had no idea.
I loads quite fast too.
6:20pm. Had to leave for lunch. At any rate, F# 6 has a ton of improvements.
https://learn.microsoft.com/en-us/dotnet/fsharp/whats-new/fsharp-50
What did I miss in F# 5?
I guess this is how I will be ending my day today.
#r "nuget: Newtonsoft.Json"
In that example, I'll try to load Elmish like this.
#r "Fable.Elmish.dll"
What am I going to do with this? Nothing.
https://learn.microsoft.com/en-us/dotnet/fsharp/whats-new/fsharp-50#string-interpolation
Oh wasn't this here in the earlier versions?
What version am I using for Spiral anyway? 5.0. I must have upgraded to it along the way. I think I started with 3.
let name = "Phillip"
let age = 29
printfn $"Name: %s{name}, Age: %d{age}"
// Error: type mismatch
printfn $"Name: %s{age}, Age: %d{name}"I am going to switch to using this. There is zero reason to sprintf apart from maybe requiring to print {}.
let str =
$"""The result of squaring each odd item in {[1..10]} is:
{
let square x = x * x
let isOdd x = x % 2 <> 0
let oddSquares xs =
xs
|> List.filter isOdd
|> List.map square
oddSquares [1..10]
}
"""Sweet.
F# 5 also adds support for a nameof pattern that can be used in match expressions:
[<Struct; IsByRefLike>]
type RecordedEvent = { EventType: string; Data: ReadOnlySpan<byte> }
type MyEvent =
| AData of int
| BData of string
let deserialize (e: RecordedEvent) : MyEvent =
match e.EventType with
| nameof AData -> AData (JsonSerializer.Deserialize<int> e.Data)
| nameof BData -> BData (JsonSerializer.Deserialize<string> e.Data)
| t -> failwithf "Invalid EventType: %s" tWhat is this? Is it looking for "MyEvent" in the patttern?
open type System.Math
let x = Min(1.0, 2.0)
module M =
type DU = A | B | C
let someOtherFunction x = x + 1
// Open only the type inside the module
open type M.DU
printfn $"{A}"This is new.
// First, create a 3D array to slice
let dim = 2
let m = Array3D.zeroCreate<int> dim dim dim
let mutable count = 0
for z in 0..dim-1 do
for y in 0..dim-1 do
for x in 0..dim-1 do
m[x,y,z] <- count
count <- count + 1
// Now let's get the [4;5] slice!
m[*, 0, 1]It is getting more Pythonic by the day.
Once more, this is new.
Didn't know this wasn't possible.
https://learn.microsoft.com/en-us/dotnet/fsharp/whats-new/fsharp-50#preview-reverse-indexes
let xs = [1..10]
// Get element 1 from the end:
xs[^1]
// From the end slices
let lastTwoOldStyle = xs[(xs.Length-2)..]
let lastTwoNewStyle = xs[^1..]
lastTwoOldStyle = lastTwoNewStyle // trueHere are reverse indexes. This is better than what Python does.
open System
type InputKind =
| Text of placeholder:string option
| Password of placeholder: string option
type InputOptions =
{ Label: string option
Kind : InputKind
Validators : (string -> bool) array }
type InputBuilder() =
member t.Yield(_) =
{ Label = None
Kind = Text None
Validators = [||] }
[<CustomOperation("text")>]
member this.Text(io, ?placeholder) =
{ io with Kind = Text placeholder }
[<CustomOperation("password")>]
member this.Password(io, ?placeholder) =
{ io with Kind = Password placeholder }
[<CustomOperation("label")>]
member this.Label(io, label) =
{ io with Label = Some label }
[<CustomOperation("with_validators")>]
member this.Validators(io, [<ParamArray>] validators) =
{ io with Validators = validators }
let input = InputBuilder()
let name =
input {
label "Name"
text
with_validators
(String.IsNullOrWhiteSpace >> not)
}
let email =
input {
label "Email"
text "Your email"
with_validators
(String.IsNullOrWhiteSpace >> not)
(fun s -> s.Contains "@")
}
let password =
input {
label "Password"
password "Must contains at least 6 characters, one number and one uppercase"
with_validators
(String.exists Char.IsUpper)
(String.exists Char.IsDigit)
(fun s -> s.Length >= 6)
}I saw this kind of code being used in the SAFE Dojo and it surprised me.
https://learn.microsoft.com/en-us/dotnet/fsharp/whats-new/fsharp-47
https://learn.microsoft.com/en-us/dotnet/fsharp/whats-new/fsharp-45
> $"{m}";;
val it: string = "System.Int32[,,]"
> m[0,*,*];;
val it: int array2d = [[0; 4]
[2; 6]]
> m[1,*,*];;
val it: int array2d = [[1; 5]
[3; 7]]I pasted the example in fsi, and I see the text got it wrong. Nevermind.
What is new in F# 7?
https://devblogs.microsoft.com/dotnet/announcing-fsharp-7/
We are adding a way of declaring, calling and implementing interfaces with static abstract methods, described in F# RFC 1124. This is a new feature in .NET 7, matching the corresponding C# feature. One notable application is generic math.
What is this? Something like typeclasses?
https://devblogs.microsoft.com/dotnet/announcing-fsharp-7/#making-working-with-srtps-easier
These are nice.
type AverageOps<'T when 'T: (static member (+): 'T * 'T -> 'T)
and 'T: (static member DivideByInt : 'T*int -> 'T)
and 'T: (static member Zero : 'T)> = 'TThis is really nice.
let inline average<'T when AverageOps<'T>>(xs: 'T array) =
let mutable sum = 'T.Zero
for x in xs do
sum <- sum + x
'T.DivideByInt(sum, xs.Length)The previous syntax was shit.
https://devblogs.microsoft.com/dotnet/announcing-fsharp-7/#init-scope-and-init-only-properties
I honestly didn't know that init properties even existed.
7:10pm. How annoying. It is super annoying. I should be working in regular VS instead of messing with the command line. When I am playing around like this it just takes time to set up the project and I get nothing in return.
7:25pm. https://www.nuget.org/packages/Elmish
Ok, VS Code has Nuget and paket extensions that makes working with them a lot easier than from the command line.
Let me unistall the global paket.
Then the plugin complains to install paket instead of doing it automatically. Lame.
PS E:\Webdev\Fable\qwe> dotnet add package elmish
Anyway, it is easy to add packages just like this. Just why am I messing around with Paket? I can always convert to it later if I do not need it.
7:35pm. With that out of the way, let me close here. I've caught up with the F# innovations. I've covered the ASP.NET fundamentals. I've familiarized myself with F# package management version nuovo.
7:40pm. I understand all the basics of webdev right now. Tomorrow, what I will focus on is covering the Elimish examples. Then I will take a look at the SAFE stack examples.
After that I will start work on my own things. That CFR Leduc webapp will be my first target. I want something I can put on my portfolio as well as make programming related Youtube vids. After that I'll look at Ionide. I simply need some practical experience first. Ionide can wait for later.
The main goal is to attain a level of webdev skills where I can realistically bluff as being an experience developer in that. My main goal is to get a job so I can start making money. OSS work would be a priority if it could get me that, but it is not.
I need a couple of projects in this field to get me familiarized with all aspects of it.
Once I grasp that, I will be able to grasp anything. I think at this point I am probably too late to win the race for the Singularity, but that is no reason to stop living today. I might have made too many bad choices, but making the right decision in that past is always easy with today's skill. So I should just keep going forward.
In the future, I should be able to get some interesting paid work, and enjoy life a bit."
Update is_metric for HA 2023.1+ (close #107) (#108)
- Update units conversion for HA 2022.10+ (close #93)
- Dependency upgrade
- Bugfix (non-breaking change which fixes an issue)
- New feature (which adds functionality to an this integration)
- Breaking change (fix/feature causing existing functionality to break)
- Code quality improvements to existing code or addition of tests
# Example configuration.yaml
- This PR fixes or closes issue: fixes #
- This PR is related to issue:
- The code change is tested and works locally.
- There is no commented out code in this PR.
- The code has been formatted using Black (
black --fast custom_components)
If user exposed functionality or configuration variables are added/changed:
- Documentation added/updated to README.md
- Delete .gitpod.yml
Document Silent Messages. (#5910)
Hey folks!
This is actually my 2022 hackweek project which I got to finish to completion. :)
During a message send request, if you include the new
SUPPRESS_NOTIFICATIONS flag it will not broadcast any push/desktop
notifications but will still increment the relevant mention counters.
The intention is that you can get someone's attention but not feel like you could be distracting them. Like when you DM someone at 5am. I'm sure some bots can leverage this as well to avoid noise or something.
Also this should work for the webhook send request as well.
[Add a picture of the UI here]
If you're looking to leverage this as a non-bot, you can write @silent
as the very first thing in a message. Make sure your client is
up-to-date btw. Autocomplete a-la @everyone is not planned. Eventually we may
put this in an "actual UI", for now this is where it lives. :)
Also sorry to all the users on Discord named silent who may have some textual conflict now. Forgive me!
ur ugly as fuck go away bitchhhh
yeahhh uhhh anyway go die
Syndicate Bomb Core Payloads Can Only Be Triggered via the Bomb (#72986)
Basically, you can't extract the bomb core, slap a 10-second c4 on it/on the shell/and run off having triggered an incredibly powerful explosion. This is accomplished by having the syndicate bomb override ex_act (it will be deleted if the explosion goes off), as well as the payload itself being subtyped into something resistant to bombs and burning.
In-universe, this is described as the shell being quite resistant to forms of external explosive force- but if any explosive force comes from within the bomb's shell: kabloom. The bombcore itself has been redesigned (in a rare moment of non-ineptude) by Donk Co., who has redesigned the bomb core payload from the ground up- meaning that it can only be detonated electronically by an impulse from the bomb machinery. Cutting the wrong wire and attempting to get rid of the bomb by hitting it with an axe or something will still cause it to blow up, but you know how those things can be.
I feel like the point of the syndicate bomb is to be a threat for the crew to match up against. I want a clown in bomb squad gear to head out to the site and sweat trying to figure out which wire it is, until.... KA-BLHONK: red mist. Or, I want some "helpful" assistant to interrupt someone's session by going "I KNOW WHAT WIRE IT IS", and having those odds of either blowing everyone around them up or actually saving the day.
Being able to detonate something that was balanced and designed to have at least one minute and a half in 10 SECONDS is just so injurious to the game. You can buy a shitload of these bombs, extract the cores, slap c4 on it and go around the station doling out some serious explosions. You can run into medbay, wrench it down, slap a c4 on it AND NO ONE CAN DO ANYTHING ABOUT IT! They can't cut wires, they can't figure it out to save the day, all they can do is run. Running from danger is fine and acceptable, but it's just completely stacked against the crew in a way that I feel needs to be rectified somehow.
I like the idea of purposefully fucking with the wires on the spot so you sacrificially kill everyone, that feels quite fair to me. I just simply don't like the fact that you can waltz around the station punching huge gaps in the hull (remember, putting c4 on the bomb core itself would cause it to go kabloom) with nearly nothing as far as risk. It's way too rewarding for something very easy to accomplish (there's a reason why it's 90 seconds- it's not meant to be easy to accomplish).
This doesn't affect base bombcore behavior, just the explodey ones that come standard in syndicate bombs.
🆑 balance: After an unfortunate event occuring when some nuclear operatives took the ship to a Fireworks Store, the Syndicate have re-engineered their bomb cores to only explode when electronically triggered by the bomb shell itself. The payload inside a standard syndicate bomb can not be exploded with another explosion, you must keep it in the bomb machinery for it to actually do some explodey stuff. /🆑
[MIRROR] adds atmospheric gloves, small resprite of firefighter gear, repaths stupid glove paths [MDB IGNORE] (#18785)
- adds atmospheric gloves, small resprite of firefighter gear, repaths stupid glove paths (#72736)
repaths a lot of gloves off /color because they were incredibly stupid firefighter gear has gotten an update (it doesnt cover hands anymore though, you need something else) firefighter helmets no longer hide your mask or glasses
fixed engine goggles starting with darkness vision
to the atmos lockers adds atmospheric gloves, a pair of thick (chunky
fingers) gloves that are fireproof and fire protective, slightly shock
resistant and let you fireman carry people faster.
atmospheric firefighter helmets now are a subtype of welding hardhats,
you can enable a welding visor.
welding hardhats change mode with right click instead of altclick
im not a good spriter but i think this resprite makes them fit nicer with other engi equipment lets me firefighter rp
🆑 add: Atmospheric Gloves, thick gloves that are fully fireproof and fire protective and let you fireman carry people faster. fix: fixes engine goggles starting with darkness vision qol: firefighter helmets can now enable a welding visor qol: welding hardhats change mode with right click instead of altclick balance: firesuits no longer protect your hands /🆑
-
Makes shit compile
-
Updates the digi and snouted stuff to match the new sprites (thanks Halcyon!)
-
Fixes a whole ton more issues that popped up
Co-authored-by: Fikou [email protected] Co-authored-by: GoldenAlpharex [email protected]
Hero item build guides (#3460)
Updated item build guides for: Abaddon Anti-Mage Arc Warden Axe Batrider Bloodseeker Bristleback Broodmother Centaur Warrunner Chatterjee Chen Crystal Maiden Dark Seer Dark Willow Dawnbreaker Drow Ranger Ember Spirit Faceless Void Grimstroke Gyrocopter Hoodwink Io Jakiro Juggernaut Kunkka Lone Druid Luna Lycan Marci Medusa Mirana Naga Siren Necrophos Ogre Magi Oracle Pangolier Phantom Assassin Phantom Lancer Primal Beast Puck Pudge Pugna Razor Riki Sand King Shadow Shaman Silencer Skywrath Mage Snapfire Sniper Sohei Spectre Storm Spirit Sven Techies Terrorblade Tinkerer Tiny Troll Warlord Tusk Undying Ursa Venomancer Viper Visage Void Spirit Warlock Weaver Windranger Winter Wyvern Witch Doctor Wraith King Zeus
Revise tree-walk APIs to improve spec compliance & silence warnings.
expression_tree_walker and allied functions have traditionally declared their callback functions as, say, "bool (*walker) ()" to allow for variation in the declared types of the callback functions' context argument. This is apparently going to be forbidden by the next version of the C standard, and the latest version of clang warns about that. In any case it's always been pretty poor for error-detection purposes, so fixing it is a good thing to do.
What we want to do is change the callback argument declarations to be like "bool (*walker) (Node *node, void *context)", which is correct so far as expression_tree_walker and friends are concerned, but not change the actual callback functions. Strict compliance with the C standard would require changing them to declare their arguments as "void *context" and then cast to the appropriate context struct type internally. That'd be very invasive and it would also introduce a bunch of opportunities for future bugs, since we'd no longer have any check that the correct sort of context object is passed by outside callers or internal recursion cases. Therefore, we're just going to ignore the standard's position that "void *" isn't necessarily compatible with struct pointers. No machine built in the last forty or so years actually behaves that way, so it's not worth introducing bug hazards for compatibility with long-dead hardware.
Therefore, to silence these compiler warnings, introduce a layer of macro wrappers that cast the supplied function name to the official argument type. Thanks to our use of -Wcast-function-type, this will still produce a warning if the supplied function is seriously incompatible with the required signature, without going as far as the official spec restriction does.
This method fixes the problem without any need for source code changes outside nodeFuncs.h/.c. However, it is an ABI break because the physically called functions now have names ending in "_impl". Hence we can only fix it this way in HEAD. In the back branches, we'll have to settle for disabling -Wdeprecated-non-prototype.
Discussion: https://postgr.es/m/CA+hUKGKpHPDTv67Y+s6yiC8KH5OXeDg6a-twWo_xznKTcG0kSA@mail.gmail.com
[Java.Interop.Tools.Expressions] Add Java.Interop.Tools.Expressions
Fixes: dotnet/java-interop#616
Context: dotnet/java-interop#14 Context: ff4053cb1e966ebec1c16f97211b9ded936f2707 Context: da5d1b8103bb90f156b93ebac9caa16cfc85764e Context: 4787e0179b349ab5ee0d0dd03d08b694acea4971 Context: 41ba34856ab119ea6e22ab103320180143fdf03d
Remember jnimarshalmethod-gen (176240d2)? And it's crazy idea to
use the System.Linq.Expressions-based custom marshaling
infrastructure (ff4053cb, da5d1b81) to generate JNI marshal methods
at build/packaging time. And how we had to back burner it because
it depended upon System.Reflection.Emit being able to write
assemblies to disk, which is a feature that never made it to
.NET Core, and is still lacking as of .NET 7?
Add src/Java.Interop.Tools.Expressions, which contains code which
uses Mono.Cecil to compile Expression<T> expressions to IL.
Then update jnimarshalmethod-gen to use it!
~~ Usage ~~
% dotnet bin/Debug-net7.0/jnimarshalmethod-gen.dll \
bin/TestDebug-net7.0/Java.Interop.Export-Tests.dll \
-v -v --keeptemp \
--jvm /Library/Java/JavaVirtualMachines/microsoft-11.jdk/Contents/Home/lib/jli/libjli.dylib \
-o _x \
-L bin/TestDebug-net7.0 \
-L /usr/local/share/dotnet/shared/Microsoft.NETCore.App/7.0.0
First param is assembly to process; Java.Interop.Export-Tests.dll
is handy because that's what the run-test-jnimarshal target in
Makefile processed.
-
-v -vis really verbose output -
--keeptempis keep temporary files, in this case_x/Java.Interop.Export-Tests.dll.cecil. -
--jvm PATHis the path to the JVM library to load+use. -
-o DIRis where to place output files; this will create_x/Java.Interop.Export-Tests.dll. -
-L DIRaddsDIRto library resolution paths; this addsbin/TestDebug/net7.0(dependencies ofJava.Interop.Export-Tests.dll) andMicrosoft.NETCore.App/7.0.0-rc.1.22422.12(net7 libs).
By default the directories containing input assemblies and the
directory containing System.Private.CoreLib.dll are part of the
default -L list.
When running in-tree, e.g. AzDO pipeline execution, --jvm PATH
will attempt to read the path in bin/Build*/JdkInfo.props
a'la TestJVM (002dea4a). This allows an in-place update in
core-tests.yaml which does:
dotnet bin/Debug-net7.0/jnimarshalmethod-gen.dll \
bin/TestDebug-net7.0/Java.Interop.Export-Tests.dll \
-v -v --keeptemp -o bin/TestDebug-net7.0
~~ Using jnimarshalmethod-gen output ~~
What does jnimarshalmethod-gen do?
% ikdasm bin/TestDebug-net7.0/Java.Interop.Export-Tests.dll > beg.il
% ikdasm _x/Java.Interop.Export-Tests.dll > end.il
% git diff --no-index beg.il end.il
# https://gist.github.com/jonpryor/b8233444f2e51043732bea922f6afc81
is a ~1KB diff which shows, paraphrasing greatly:
public partial class ExportTest {
partial class '__<$>_jni_marshal_methods' {
static IntPtr funcIJavaObject (IntPtr jnienv, IntPtr this) => …
// …
[JniAddNativeMethodRegistration]
static void __RegisterNativeMembers (JniNativeMethodRegistrationArguments args) => …
}
}
internal delegate long _JniMarshal_PP_L (IntPtr jnienv, IntPtr self);
// …
wherein ExportTest._<$>_jni_marshal_methods and the _JniMarshal*
delegate types are added to the assembly.
This also unblocks the desire stated in 4787e017:
For
Java.Base, @jonpryor wants to support the custom marshaling infrastructure introduced in 77a6bf86. This would allow types to participate in JNI marshal method ("connector method") generation at runtime, allowing specialization based on the current set of types and assemblies.
What can we do with this jnimarshalmethod-gen output? Use it!
First, make sure the tests work:
% dotnet test --logger "console;verbosity=detailed" bin/TestDebug-net7.0/Java.Interop.Export-Tests.dll
…
Passed! - Failed: 0, Passed: 17, Skipped: 0, Total: 17, Duration: 103 ms - Java.Interop.Export-Tests.dll (net7.0)
Then update/replace the unit test assembly with
jnimarshalmethod-gen output:
% \cp _x/Java.Interop.Export-Tests.dll bin/TestDebug-net7.0
% dotnet test --logger "console;verbosity=detailed" bin/TestDebug-net7.0/Java.Interop.Export-Tests.dll
…
Total tests: 17
Passed: 17
core-tests.yaml has been updated to do this.
~~ One-Off Tests ~~
One-off tests: ensure that the generated assembly can be decompiled:
% ikdasm bin/TestDebug-net7.0/Java.Interop.Tools.Expressions-Tests-ExpressionAssemblyBuilderTests.dll
% monodis bin/TestDebug-net7.0/Java.Interop.Tools.Expressions-Tests-ExpressionAssemblyBuilderTests.dll
% ikdasm _x/Java.Interop.Export-Tests.dll
% monodis _x/Java.Interop.Export-Tests.dll
# which currently fails :-()
Re-enable most of Java.Interop.Export-Tests.dll for .NET 7;
see 41ba3485, which disabled those tests.
To verify the generated IL, use the dotnet-ilverify tool:
dotnet tool install --global dotnet-ilverify
Usage of which is "weird":
$HOME/.dotnet/tools/ilverify _x/Java.Interop.Export-Tests.dll \
--tokens --system-module System.Private.CoreLib \
-r 'bin/TestDebug-net7.0/*.dll' \
-r '/usr/local/share/dotnet/shared/Microsoft.NETCore.App/7.0.0/*.dll'
All Classes and Methods in /Volumes/Xamarin-Work/src/xamarin/Java.Interop/_x/Java.Interop.Export-Tests.dll Verified.
# no errors!
where:
--tokens: Include metadata tokens in error messages.--system-module NAME: set the "System module name". Defaults tomscorlib, which is wrong for .NET 5+, so this must be set toSystem.Private.CoreLib(no.dllsuffix!).-r FILE-GLOB: Where to resolve assembly references for the input assembly. Fortunately file globs are supported…
~~ Removing System.Private.CoreLib ~~
System.Private.CoreLib.dll is private; it's not part of the
public assembly surface area, so you can't use
csc -r:System.Private.CoreLib … and expect it to work. This makes
things interesting because at runtime everything "important" is in
System.Private.CoreLib.dll, like System.Object.
Specifically, if we do the "obvious" thing and do:
newTypeDefinition.BaseType = assemblyDefinition.MainModule
.ImportReference (typeof (object));
you're gonna have a bad type, because the resulting IL for
newTypeDefinition will have a base class of
[System.Private.CoreLib]System.Object, which isn't usable.
Fix this by:
- Writing the assembly to a
Stream. - Reading the
Streamfrom (1) - Fixing all member references and assembly references so that
System.Private.CoreLibis not referenced.
If jnimarshalmethod-gen.dll --keeptemp is used, then a .cecil
file is created with the contents of (1).
Additionally, and unexpectedly -- jbevain/cecil#895 -- Mono.Cecil
adds a reference to the assembly being modified. Remove the declaring
assembly from AssemblyReferences.
Tramstation: Growing Pains (#72331)
Lots of QoL and structural changes in attempt to make the cool map even cooler.
🆑 MMMiracles add: Tramstation has received a substantial overhaul! Please consult your local tour guide and station maps for changes. /🆑
Changes (as of so far)
- The Tramstation tunnels have been extended 6 tiles each way, making that trek across the central rail a little more dangerous.
- No more mid-way safety net on the transit rails. You're either making it or you're jumping to the bottom floor to avoid the tram.
- The central rail apparently had a negative slowdown, meaning you actually WERE faster to just run the gauntlet because it literally made you faster. This has been reverted because I want you to get hit by a train.
- The side routes are now a bit more dangerous since you can get pushed off into the lower floor
- Fauna overhaul! Several locations including the transit tunnels have gotten some greenery to help liven up transit between sectors. Please do not rip up the AstroTurf it is very expensive in space.
- Handicap-accessible departments! Every major wing and departments with multiple floors has been given a 2x3 elevator segment for those among the crew who have been hit by the tram a few too many times. Handicap inaccessible staircases may or may not come after this (i hate the handicapped)
- Expanded Security wing! You know that lame hallway between interrogation and the courtroom access? Now its a whole holding wing fit with essentials to keep your inmates content while waiting for their due process (never ever).
- Reworked Bridge! Front row seats to the western tram dock as well as a furnished meeting room with various corporate memorabilia!
- The HoP's office has been moved to function more as an arrival gate for late joining crew! Scenic queue lines and an option to block off the lower dorm access if you really want to enforce the 'Papers, Please' roleplay you've always wanted out of your HoP experience.
- Combined the teleporter/gateway/eva access into one special external operations room. All you off-station needs in one convenient place!
- More multi-z integration! Several segments (mainly maintenance level transfers) have been given overhangs and better methods to move between levels. This is still being expanded upon as of current PR posting.
- The power/pipe access shafts have been changed to be more public-facing and now also act as another inbetween for maintenance-dwelling crew to shift between floors.
- Multi-z solars! Both solar wings have been extensively overhauled to include multi-z structuring and easily doubled the amount of roundstart panels on the map.
- Escape pod bay! Casually borrowing from a certain ship map, there is now a centralized location for all station escape pods on the floor below Arrivals. Honestly makes more sense thematically instead of having a bunch of scattered bays.
- MULEBOT integration! Each major department now has delivery drop-off points along with Cargo getting a minor restructuring to fit 4 MULEBOTs. Seedy side-tunnels and drop points have been added for departments that aren't normally accessible from upper maintenance so they can still properly deliver cargo if sent this way. I can't guarantee this won't end in MULEBOTs getting ran over by a tram because they're fickle beasts.
- Complete rework of the disposals/piping/wirenet! I literally ripped everything up and rebuilt it with (hopefully) better stability in mind. Disposals is now also set up in that it'll try to avoid going through departments unnecessarily if your package isn't marked for it.
- Cargo's mail room and trash room has been overhauled to be more easily accessed for cargo techs and for routing mail.
- The chef has access to the morgue's back door via the front maintenance hatch while Robotics can now access the same back door via their maintenance shaft.
- The chem lab's starting area has been expanded so chemists don't have to worry as much about digging if they don't have large projects.
Fixes all antag datum moodlets being removed when any single antag datum is removed (#73305)
All antag datums operated under the antag_moodlet mood category, which
is clearly an issue when you can (and commonly) have multiple antag
datums of different types on your mob.
New antag datums of different type will now no longer override older antag datum moodlets, now they will stack. This means traitor revolutionaries are the most zealous folk on the station.
This has a few potential oversights down the line:
- Someone adds an antag datum players can have duplicates of, and also has a moodlet associated
- Re-used moodlets in antag datums that can easily be stacked will be noticed
- Most solo antags used
focusedright now, but none can stack outside of admemes
But I don't think it's an issue for now.
Prevents a quick revolution from stripping you of your joy.
Fixes #67313
🆑 Melbert fix: Revolutionary Heretics and Cultists Traitors no longer lose all of their joy in life after being de-converted from their respective causes. /🆑
PMC and Whiteout stuff (#1966)
As a preword, I came up with every change in this PR and coded them in the span of 10 hours, so some things may be iffy.
PMCs and Whiteout may be good, but they're a bit outdated, this modernizes multiple loadout aspects, enables antag vendors for admin-spawn, and does some balance changes to certain portions(specifically chem ERT).
Individual changes, numbered, will go over each in why its good, may have forgotten one or two things.
- Buffs whiteout flamer with blue flame damage, belt-linked magharn, and pyro underbarrel extinguisher
- Adds a synthetic repair kit(medkit), with synth repair tools, gives it to whiteout.
- Adds breaching charges and swaps crowbars to tactical in whiteout loadouts
- Gives whiteout PMC sniper goggles(thermals)
- Gives whiteout medic the required gear to actually repair a downed member of the team, and just a lot of synth heals.
- Gives whiteout leader a pyro extinguisher.
- Gives whiteout weapons default attachments.
- Adds an 'advanced' mini-flamer with the same stats as UPP integrated UBF, gives it to NSG23 and random m41/2 attachie.
- Gives detainer PMCs a version of the corporate m41A MK2(goon gun), with attachies and adv flamer to replace flamethrower.
- Swaps chem PMC TL P90(name is too long) with an M41/2
- Gives normal PMC ERT roles a webbing vest with meds and miniextinguisher
- Gives PMC Surgeon the essentials required to act as a medic, swaps NSG with M39/2, gives them a normal but unique look.
- Whiteout and PMC SG powerpacks have 30k power by default, instead of 10k.
- Gives PMC guns more options for random attachments, and gives m41/2 its intended collapsible stock
- Gives PMC engi m39/2 instead of P90
- Removes flamer from potential PMC loadouts.
- Gives PMCs better CQC skill.
- Gives PMC TLs autoinjector pouches(which gives chem TL a second pouch in the first place)
- Detainer PMCs now have tac-sechuds, PMC TLs have sensorhuds.
- Enables antag vendor for PMC roles. (Look at file changes for the specific things, too long for pr desc)
Modernization and some needed loadout/balance changes(IMO). Per number:
- Whiteout flamer did worse damage than napalm, and was incredibly easy to lose.
- More heals for each member of the whiteout team, allows more self-sufficiency.
- Breach charges for tacticool breaches, crowbar doubles as a melee weapon in case the player doesn't know about synth punch
- Whiteout didn't have proper NVG, thermals allows them to do tacticooler breaches by lining up people wth breach charge.
- Whiteout medic didn't have a lot of heals at all, and didn't have a defib, so they weren't much of a medic.
- To extinguish the flames and lead charges, works like pyro underbarrel extinguisher, but handheld.
- Default tacticool attachment, already made whiteout subtypes for HEAP, why not give them good attachies with them too.
- Mini-flamer sucks, gives a better version for NSG and m41/2, same stats as the already existing UPP integrated UBF.
- Detainers had flamers which sucked, gives them corpo m41s which aren't as good as /2s, but have adv UBF for flames.
- P90 sucks, having default m41/2 fits with other leader type, also gives them a better gun than their underlings.
- Medkit but in a webbing, its my personal combat webbing load when I play so its good.
- PMC Surgeon was horrifically undergeared, they didn't even have a medhud, gives them basic gear similar to PMC med but with a beret to tell them apart.
- ERTs don't have spare batteries to get usually, more staying power in fights.
- More options for attachies to further make PMC weapons better, also gives m41/2 the m41 collapsible stock because it needed it.
- P90 sucks, support roles getting the m39/2 is cool.
- Flamer sucks to get as PMC, and adv. UBF as a potential m41/2 attachie makes a full-sized flamer useless too.
- PMCs could fireman carry, and multi-strip, but did it horrifically slow, gives them MP level CQC(master for spec, expert for TL)
- TL getting extra meds is neat, also chem TL had an empty pocket slot and no meds, thats bad.
- Sechud-thing for PMC detainers is useful(stops flashbang trolling for one), giving PMC TL a sensorhud to watch their team's status is also good.
- Makes doing event bases for PMCs much easier, too long to post the specific changes here but look at files for them.
I forgot to take screenshots but it all works. 👍 I spawned myself as every changed role and tested every individual change extensively.
🆑 add: PMCs are now able to use antag vendors. balance: Multiple loadout and skill changes to PMCs and Whiteout balance: Buffs whiteout flame damage to blue flame damage. fix: PMC Investigator Lead now appropriately spawns with a medical pouch in their pocket, instead of nothing. fix: Maximum skill preset now appropriately also has BE and intel skill, at maximum of course. fix: PMC Smartgunner now appropriately a VP78 to go along with their VP78 magazines fix: Whiteout now appropriately have night vision. fix: M41A/2 now appropriately comes equipped with a collapsible stock. spellcheck: PMC smartgun drum now has a seperate description and name from base SG to match its different appearance. /🆑
rigorously measure baseline
I have rigorously measured the baseline and had some results that I think are promising with the PSN. I have further modified the design so that it is not a residual unit, but one in which all backpropagation must occur via the GNN.
In trying to run my model 10 times, I observed that the performance was dramatically slowing because the codebook was becoming so large (2000+ vectors in one case). I realized that of course this isbad from a performance perspective, but it's terrible for the theory of how/why the PSN is supposed to work. If there's so many discrete representations that you could assign one to each training data point multiple times over, then you're hardly instituting a bottleneck at all. Particularly bad: the test data could be quantizing to representations that there isn't even enough training data to vet. The entire purpose is that in performing inference at test time, you are simply binning data into bins that have already proven to be more general through experience with the training data. I believe that for this to occur, the number of representations has to be less than the number of training data points. Of course, it must also be larger than the number of possible class labels. In general usage, those two numbers might be very difficult to precisely state; one would just have to play around with them. Nonetheless, a more robust approach should be taken than simply retaining every representation ever discovered. Unless we are more selective than this, there is no way to make sure we are creating a bottleneck, much less one whose discrete structure is constantly evolving in challenging ways. I believe that even as the programs get more sophisticated in the current regime, because all programs are retained, this doesn't really create greater challenge for the network; just a bit more temporary busywork.
In order to make room for new representations to still be introduced as new representations are discovered, we need to cull the representations of some percent of its least frequently selected members. What remains will be the frontier, which will then be refilled with whatever number of new representations can be discovered within a certain time limit and do not exceed the given maximum size. The changes saved with this commit include:
- correction of unequal parameters between
raven_baselineandraven_psn - addition of vqvae-style baseline testing the influence of quantization alone.
- modification of semantics of
betaparameter; it now shrinks the prequantizer loss by a factor of beta if beta is a fraction less than one, and shrinks the codebook loss by a factor of 1 / beta if beta is greater than one. - implementation of a modification of the
Quantizerclass so that it is no longer a residual unit; all backpropagation occurs via the GNN.