1,552,504 events, 912,161 push events, 1,267,908 commit messages, 72,603,475 characters
Fix (TCollection *) to (TSortedCollection *) cast in TSortedListBox::list()
God damn. Because TCollection and TSortedCollection are no more than forward declarations by the time TSortedListBox::list() is defined, the cast is implemented as a reinterpret_cast. As a consequence, invoking TSortedListBox::list() would provide the wrong result.
Amazingly, Borland C++ handles this fine, so I have been hugely confused while debugging this.
This is my first experience where C-style casts have silenced a compilation error.
housekeeping: yarn cache && update deps
Holy shit. 1123 things are staged. What the actual fuck is happening.
I'm scared.
Hidden gestural bar: fix visual glitches when switching states
hacky way but it works (TODO: find a proper fix in the new year, i don't have motivation to dig into navbar/keyguard code fuckery now lol)
setting the height to 1px keeps the navbar almost invisible but fixes the annoying visual glitches when going from screen off to ambient pulsing or lockscreen (more noticeable on some devices like bonito) to replicate the issue without this commit:
- screen ON, then screen off, then double tap to go to ambient, then double tap to go to lockscreen, then double tap to switch screen off, then switch screen on again
- or just switch screen off/on a few times with the power button
Also sync the hide pill code with Pulse hide pill feature
Change-Id: Ib1cc83492f8a091be5cac4563d844010cef69dbc
"11:10am. I did not go to bed too late yesterday, but I had a hard time falling asleep. Sigh.
11:15am. Let me go for dinner, and chores. Then I'll resume the docs. I am really groggy right now, but I need to get a move on.
12pm. Let me do the chores and then I will start this thing. I slacked way too much.
12:25pm. Done with chores.
12:30pm. Focus me. Let me start work on the docs. I think it is time to showcase macros.
During the night I had some awkward thoughts.
If we assume that Spiral's let statements are equivalent to F#'s, while inl is let inline, then wouldn't that mean that F# in facts supports the Spiral style inlining in full?
If somebody asked me this, I'd find it embarassing to reply. I actually do not know how F# works under the hood. Though I am a functional programming expert by now, I simply never cracked open the disassembler and looked at what comes out.
The reason for that is, even if in theory F# could hit the permance goals, it would not give me the language introp capabilities that I need. So I won't go into it in the docs.
Since I am going to be applying to chip companies, instead of pushing performance, I really do need to highlight the language interop benefits of Spiral.
Let me take a look at the tests.
12:45pm. Let me just go with macros. It seems that I want to get the differences from existing languages out of the way so let me go with theme.
1:35pm. Macros come up to two pages. This is nice and easy.
1:45pm. I am thinking how to approach the next section and even what it should be.
2:10pm. Had to take a break. Let me resume.
2:15pm. Focus me. Let me start writing this thing. I think I've covered 30% of the stuff in the top down segment. It does not have to be 10k exactly. I can do this for two weeks. I'll do this for as long as needed. Until I feel satisfied.
Let me do the layout types next.
3:10pm.
inl main () =
inl a = mut {q=1i32; w=2i32}
a <- {q=3; w=4}
type Heap0 = {l0 : int32; l1 : int32}
and Mut0 = {mutable l0 : int32; mutable l1 : int32}
let v0 : Mut0 = {l0 = 1; l1 = 2} : Mut0
v0.l0 <- 3
v0.l1 <- 4This is a compiler error. That Heap0 is not supposed to manifest. What is going on here?
Also when it comes to doing the real segment I really will require some way of taking advantage of print_static.
3:15pm. Focus me. Let me switch gears so I can debug this error. I'll have to track this down.
Let me make this one of the tests.
let rec heap = layout (fun s x ->
line s (sprintf "Heap%i = {%s}" x.tag (x.free_vars |> Array.map (fun (L(i,t)) -> sprintf "l%i : %s" i (tyv_proxy t)) |> String.concat "; "))
)
and mut = layout (fun s x ->
line s (sprintf "Mut%i = {%s}" x.tag (x.free_vars |> Array.map (fun (L(i,t)) -> sprintf "mutable l%i : %s" i (tyv_proxy t)) |> String.concat "; "))
)Hmmm, these are both layout.
3:25pm. What is going on here, I have no idea.
3:30pm. I finally got the plugin to run from the host again. I forgot how to do that recently. Does heap actually trigger at some point? What is going on here?
let rec heap = layout (fun s x ->
line s (sprintf "Heap%i = {%s}" x.tag (x.free_vars |> Array.map (fun (L(i,t)) -> sprintf "l%i : %s" i (tyv_proxy t)) |> String.concat "; "))
)I'll put a break point here and then take a look at the call stack.
| TyLayoutToHeap(a,b) -> sprintf "{%s} : Heap%i" (layout_vars a) (heap b).tag |> simple
| TyLayoutToHeapMutable(a,b) -> sprintf "{%s} : Mut%i" (layout_vars a) (heap b).tag |> simpleAh, look at that heap b on the last line. Whops.
It should be mut b.
type Mut0 = {mutable l0 : int32; mutable l1 : int32}
let v0 : Mut0 = {l0 = 1; l1 = 2} : Mut0
v0.l0 <- 3
v0.l1 <- 4Yeah, now I get correct compilation. This was a minor copy paste error still lurking inside the compiler, probably there are more of a similar kind in there. Let me publish the change as a patch.
I'll commit here."
Reorganize files even less idiotically
-- note: haskell files will break unless copy pasted in the source folder because haskell path manipulation is bullshit
Data::Dumper (XS): use mortals to prevent leaks if magic throws
For example:
use Tie::Scalar; use Data::Dumper; sub T::TIESCALAR { bless {}, shift} sub T::FETCH { die } my $x; tie $x, "T" or die; while(1) { eval { () = Dumper( [ $x ] ) }; }
would leak various work SVs.
I start a new scope (ENTER/LEAVE) for most recursive DD_dump() calls so that the work SVs don't accumulate on the temps stack, for example if we're dumping a large array we'd end up with several SVs on the temp stack for each member of the array.
The exceptions are where I don't expect a large number of unreleased temps to accumulate, as with scalar or glob refs.
(cherry picked from commit 815b4be4ab7ae210f796fc9d29754e55fc0d1f0e)
parent 9f8dddc5e34f9af27f614482b08c13ea7e3a175d author Grey Moore [email protected] 1584497424 -0400 committer boom100100 [email protected] 1609002686 -0500
add judge_imposed_i589_deadline column to friends
allow filtering Friends by judge-imposed deadline in range
add judge-imposed deadline to Friend form
add judge-imposed deadline to accompaniment report form
reorder search filters
remove notification emails when application is submitted
Add has_a_lawyer and lawyer_name to friends table
allow setting has_a_lawyer and lawyer_name from accompaniment report form
validate lawyer_name presence if has_a_lawyer
allow admin to edit lawyer info for friend
allow admin to filter Friends list by has_a_lawyer
dont require lawyer_name for Friends who have a lawyer
update in_review -> review_requested
update changes_requested -> review_added
improve ux for regional admins
fix failing spec with regional admin dropdown options
[Remote Lawyers] Remove Team From View, Not Needed
[Remote Lawyers] 'Add Review' --> Button, Move By Approve, Add Edit
[Remote Lawyers] Move Status First, Color Differently for Lawyers, Rm Review Date, Add Reviews Col
[Remote Lawyer] Add Headers For Col Names
[Remote Lawyer] Change 'Review by'... to 'Your Review' if Name is Viewer
spacing adjustments
beginning
fix remote_review_action validations
mvp version of remote review action log
only send email after all database commits are successful
move remote review index to its own controller
additional associations and validations for remote_review_actions
add sort by last name asc and desc
simplify remote clinic lawyer friend list
allow remote clinic lawyers to upload drafts
[Regional Admins] Remote Lawyers Parity View for Drafts
[NY Regional Admins] Bring Page to Parity with Remote Clinic Lawyers (a.k.a. prettify)
upgrade rails to 5.2.4.3
make access time slot calendar weekly rather than monthly
regional admins navigate from friend record to remote clinic and can submit a draft
update puma to 3.12.6
remove activity event column
rename activity types in seeds
refactor some colors
remove clinic training zine link
add rollbar
cleanup
[Download] Our Best Guess (lol)
try
try
try
fix spacing
add view and download options for pdfs
fix specs
order applications alphabetically
ignore action controller routing errors in rollbar
Fix app init using docker compose
Replace chosen with select2
rename chosen classes to select2 and fix specs
fix duplicating select2 elements when the browser back button is used
clarify docker-specific instructions around postgres
initiate select2 with bootstrap theme on all pages except search
use standard activateSelect2() to initiate select2 on volunteer access page
kick the can down the road on select2 with attendance specs
turns out select2 does not work at all on dropdowns with thousands of options...but chosen does...so putting chosen back in for these...we probably should not have dropdowns with thousands of options, bet that impacts page load time
only display confirmed users in user access dropdowns
use select2 ajax search for admin volunteer access collection search
ajax search for family member association dropdown on Friend record
replace the rest of the chosen js dropdowns with select2 backed by ajax search
upgrade to rails 6
configuration changes for rails 6
remove webpacker gem
fix failing specs and deprecation issues from rails 6 upgrade
fix spec
cleanup dev and test gems and remove springs, model and controller specs pass
use selenium chrome driver and fix previously skipped specs
dont install phantomjs in circleci and dont seed db
cleanup
get rid of specs that are testing standard rails associations and validations, some spec fixes for rails upgrade
update bundler in circleci config
gem updates
explicitly set rubygems as source
fix ajax search for attendance dropdowns
update production gems
fix friend collection_select on activity form
upgrade to ruby 2.6.6
simplify flash messages
reorder so constant is at top of helper file
make remote clinic lawyer a full role
upgrade rails to 6.0.3.3 and update gems
update seeds for new faker version
[Friends] Fix Render Table
Added ampm to all datetime fields
fix datetime picker helper for specs
Update invitations_controller.rb
fix for existing functionality for regional admins to set 'Remote Clinic Lawyer' role when inviting users
create a new array of roles, rather than changing frozen array
fix combined array
remove old remote_clinic_lawyer boolean
Pull pg 9.6 correctly
Fix config
Remove mac-specific instructions, and added a bit about setting correct file contents
Removed irrelevant instruction
add helper notes specific to Windows OS
add a delay for select2 ajax calls to mitigate 429 responses
display social work notes on friend show page
add a notes label to social work referral notes
include a blank option on lawyer selects
change eoir case status options and status options
allow primary community admins to assign remote clinic lawyers and other small tweaks
fix scope for remote clinic lawyer dashboard
small tweaks
add ip blocking configuration for rack attack
remove all links from emails
add devise authy
conditionally display 2fa enable/disable
override devise require_token? method to check password_changed_at timestamp
more devise-authy customization
[Authy] Log Whether Admin Logged in Using 2Fac or Not
add to authy code instructions and limit friend pagination to 10 results per page
make logging style a bit more consistent and set 1.week authy rememberable configuration
remove pagination and limit results for friend and users index
restrict user editing for admins and data entry users
rough cut of adapting the eoir caller roler for accompaniment leader data entry
cleanup
grant some missing accompaniment leader functionality to eoir_caller
allow eoir callers to confirm/unconfirm activities
allow accompaniment leader data entry to see 4 weeks instead of 3
add data entry policies agreement boolean to users
Revamp blessed gain ability
One balance problem in NetHack is that blessed gain ability is not a very interesting item: all attributes are raised equally and there's no choice involved, so there's never any real reason you might want to save one for later. Also, I think it's rather too powerful: you only really need about 3-5 of them in most games to get high enough, if not maxed, scores in all six attributes, perhaps contributing to the issue where all high-level characters of a given role feel the same.
This makes it so that blessed gain ability raises one attribute by d2 points, with the player able to choose which or let it choose randomly. With this change, using it becomes more of a decision into how you want to build your character (there still isn't much reason to save it for later unless you're not sure of how your build will go, though). For instance, it allows a Knight who wants to cast spells to target his low starting intelligence for boosts.
Also in this commit is a refactor of how gain ability randomly picks an attribute to raise, for uncursed potions and blessed potions if the player chooses the random option. Previously it made 6 tries to raise a random attribute, and if all 6 attempts picked a maxed one you were out of luck. Now it shuffles an array of the six attributes and iterates through them until a raisable one is found, so it's still random but doesn't have a chance of failing to do anything.
add judge_imposed_i589_deadline column to friends
allow filtering Friends by judge-imposed deadline in range
add judge-imposed deadline to Friend form
add judge-imposed deadline to accompaniment report form
reorder search filters
remove notification emails when application is submitted
Add has_a_lawyer and lawyer_name to friends table
allow setting has_a_lawyer and lawyer_name from accompaniment report form
validate lawyer_name presence if has_a_lawyer
allow admin to edit lawyer info for friend
allow admin to filter Friends list by has_a_lawyer
dont require lawyer_name for Friends who have a lawyer
update in_review -> review_requested
update changes_requested -> review_added
improve ux for regional admins
fix failing spec with regional admin dropdown options
[Remote Lawyers] Remove Team From View, Not Needed
[Remote Lawyers] 'Add Review' --> Button, Move By Approve, Add Edit
[Remote Lawyers] Move Status First, Color Differently for Lawyers, Rm Review Date, Add Reviews Col
[Remote Lawyer] Add Headers For Col Names
[Remote Lawyer] Change 'Review by'... to 'Your Review' if Name is Viewer
spacing adjustments
beginning
fix remote_review_action validations
mvp version of remote review action log
only send email after all database commits are successful
move remote review index to its own controller
additional associations and validations for remote_review_actions
add sort by last name asc and desc
simplify remote clinic lawyer friend list
allow remote clinic lawyers to upload drafts
[Remote Clinic] Abstract View Into Partial
[Regional Admins] Remote Lawyers Parity View for Drafts
[NY Regional Admins] Bring Page to Parity with Remote Clinic Lawyers (a.k.a. prettify)
upgrade rails to 5.2.4.3
make access time slot calendar weekly rather than monthly
regional admins navigate from friend record to remote clinic and can submit a draft
update puma to 3.12.6
remove activity event column
rename activity types in seeds
refactor some colors
remove clinic training zine link
add rollbar
cleanup
[Download] Our Best Guess (lol)
try
try
try
fix spacing
add view and download options for pdfs
fix specs
order applications alphabetically
ignore action controller routing errors in rollbar
Fix app init using docker compose
Replace chosen with select2
rename chosen classes to select2 and fix specs
fix duplicating select2 elements when the browser back button is used
clarify docker-specific instructions around postgres
initiate select2 with bootstrap theme on all pages except search
use standard activateSelect2() to initiate select2 on volunteer access page
kick the can down the road on select2 with attendance specs
turns out select2 does not work at all on dropdowns with thousands of options...but chosen does...so putting chosen back in for these...we probably should not have dropdowns with thousands of options, bet that impacts page load time
only display confirmed users in user access dropdowns
use select2 ajax search for admin volunteer access collection search
ajax search for family member association dropdown on Friend record
replace the rest of the chosen js dropdowns with select2 backed by ajax search
upgrade to rails 6
configuration changes for rails 6
remove webpacker gem
fix failing specs and deprecation issues from rails 6 upgrade
fix spec
cleanup dev and test gems and remove springs, model and controller specs pass
use selenium chrome driver and fix previously skipped specs
dont install phantomjs in circleci and dont seed db
cleanup
get rid of specs that are testing standard rails associations and validations, some spec fixes for rails upgrade
update bundler in circleci config
gem updates
explicitly set rubygems as source
fix ajax search for attendance dropdowns
update production gems
fix friend collection_select on activity form
upgrade to ruby 2.6.6
simplify flash messages
reorder so constant is at top of helper file
make remote clinic lawyer a full role
upgrade rails to 6.0.3.3 and update gems
update seeds for new faker version
[Friends] Fix Render Table
Added ampm to all datetime fields
fix datetime picker helper for specs
Update invitations_controller.rb
fix for existing functionality for regional admins to set 'Remote Clinic Lawyer' role when inviting users
create a new array of roles, rather than changing frozen array
fix combined array
remove old remote_clinic_lawyer boolean
Pull pg 9.6 correctly
Fix config
Remove mac-specific instructions, and added a bit about setting correct file contents
Removed irrelevant instruction
add helper notes specific to Windows OS
add a delay for select2 ajax calls to mitigate 429 responses
display social work notes on friend show page
add a notes label to social work referral notes
include a blank option on lawyer selects
change eoir case status options and status options
allow primary community admins to assign remote clinic lawyers and other small tweaks
fix scope for remote clinic lawyer dashboard
small tweaks
add ip blocking configuration for rack attack
remove all links from emails
add devise authy
conditionally display 2fa enable/disable
override devise require_token? method to check password_changed_at timestamp
more devise-authy customization
[Authy] Log Whether Admin Logged in Using 2Fac or Not
add to authy code instructions and limit friend pagination to 10 results per page
make logging style a bit more consistent and set 1.week authy rememberable configuration
remove pagination and limit results for friend and users index
restrict user editing for admins and data entry users
rough cut of adapting the eoir caller roler for accompaniment leader data entry
cleanup
grant some missing accompaniment leader functionality to eoir_caller
allow eoir callers to confirm/unconfirm activities
allow accompaniment leader data entry to see 4 weeks instead of 3
add data entry policies agreement boolean to users
rpm: Autospec creation for update from version 4.10.2 to version 1
Florian Festi (1): Fix whitespace in Python doc string
Jakub Jelinek (1): Add DWARF-4 support to debugedit (RhBug:707677)
Jindrich Novy (3): Avoid freeing an unallocated variable (RhBug:688091) Understand 'PK00' zip archives (RhBug:699529) Fix find-lang so that it finds @.qm QT i18n files (RhBug:699945)
Mark Wielaard (1): Add -r flag to find-debuginfo.sh to invoke eu-strip --reloc-debug-sections.
Michael Schroeder (9): Allow uncompressed payload in packages Fix corner-case behavior on dependency matching when release not present Fix "method not permitted before handle's open method" on --verifydb Fix segfault on build with empty %prep section Do not abort if chown/chmod fails but the file is already correct Always copy macro source when expanding it Support "magic_and_path" flag in fileattrs Do not die on empty changelog section Restore %defattr() behavior on special %doc
Panu Matilainen (151): A bit of sanity checking in rpmtsRebuildDB() Kick out some long since invalid comments Add beginnings of python bindings test-suite Remove broken versioned provides rpmlib() tracking Support excluding by path or magic in file classification Rename pattern -> path for, duh, path patterns Refactor the helper execution out of rpmfcHelper() Some further preliminaries for next bits Implement filtering of autogenerated dependencies Implement transaction ordering hinting Replace verified header tracking bitfield with a hash table Cut down the initial db checked hash allocation somewhat Remove hardcoded lock and mail group id's Split user+group caching to separate source (again), rename Unify the user+group caching between librpm and librpmbuild Remove the now unused user/group caching code Eliminate ancient dependency loop whiteout mechanism Support GetEntry() for hashes without datatype too Add a "unique string cache" to rpmug Unbreak file user/groupname handling in build Permit comparison operator strings (<, >= etc) in python ds constructor Move python test-suite macros to local.at, rename Add a pile of dependency matching tests for "obvious" cases Add pile of further dependency match testcases Yet another pile of depmatch tests + added commentary Rip the stillborn, broken apply/commit transaction goo Rip the remains of --aid and --nosuggest, except for the callback Mark two more unused rpmts flags as such Free up a bunch of bits from rpmtransFlags_e Move rpmtsSELabelFoo() functions to an internal-only header Update translations to kick out removed messages etc Make rpmplugins.h private for now Only sepolicy-plugin needs linking to libsemanage Require %files section for package generation again Bump up default BDB cache- and allowed mmap size considerably A largish man-page update Tweak up BDB flags to avoid breakage from --verifydb Hide --verifydb switch again Honor --root in rpmkeys too Fix rpmsign --key-id popt alias typo Add switch to allow printing only soname dependencies to elfdeps helper Preparing for 4.9.0-beta1 Permit queries from rpmdb on read-only media (RhBug:671200) Restore default SIGPIPE handling for build scriptlets (RhBug:651463) Fix python documentation wrt dbIndex() Avoid automatic index generation on db rebuild Fix rpmdb index match iteration termination with NULL keyp (#671149) Plug potential division by zero in the rpmdb size calculation (RhBug:671056) Don't try to remove existing environment when using private environment Teach rpm about post-transaction dependencies Add an error message + comments on open(".") behavior (RhBug:672576) Eww, mono rules both buggy AND missing from tarballs, doh. Argh, yet more mono dependency generation braindamage Mark the identical signature warning as translatable Adjust OCaml detection rule for libmagic 5.04 -> 5.05 string change Fix segfault when building more than one rpm (RhBug:675565) Add + use a db error callback function Callback argument mismatch from previous commit, meh Import the C-level ts python object as TransactionSetCore Allow installation of self-conflicting packages (ticket #806, RhBug:651951) Preparing for 4.9.0-rc1 Allow both string + unicode in python addInstall() / addErase() Fix braindamage in the depgen helper collector loop (RhBug:675002) Fix db cursor double-open, causing yum to hang on reinstall (RhBug:678644) Silence error callback during from BDB during environment open Preparing for 4.9.0 Throw an exception from Fseek() errors in python rpmfd.seek() Make peace with autoconf-2.68 Preferred color pkgs should be erased last Improve the dependency loop debug message a bit Take file state into account for file dependencies Verify some properties of replaced and wrong-colored files (RhBug:528383) Python 3 fixups Fix the PyBytes vs PyString compatibility defines Update librpm doxygen module list Missing va_end() call Plug memory leaks on macro definition error cases Fix a small memleak in rpmsign tool Fix classification of ELF binaries with sticky bit (RhBug:689182) Use pkg-config to find Lua + determine flags (ticket #88) Add a more useful example to rpm2cpio manpage Dont reference transaction set from transaction elements Fix dangling databases from iterators (ticket #820) Remember to free db index iterators too on forced termination Reflect file classifier errors in rpmdeps exit code Actually handle headerGet() / pgpPrtPkts() failure on signature verify Catch write errors when generating scriptlet temporary files Give at least some indication of error from fchdir() failures Remove redundant indentation block from rpmSign() headerPut() and headerDel() returns aren't interesting here Handle errors from moving target file into place in rpmSign() Empty transaction is not an error (RhBug:699929) -D is for --define, not --predefine (RhBug:706161) Unify fileattr include- and exclude-rule handling Whoops, flags needs to be sorted for argvSearch() to work correctly Handle HEADERFLAG_SORTED bit correctly in headerUnsort() Fix %prep parse error to abort build Bail out of debuginfo if stabs format encountered (RhBug:453506) Plug a memory leak on Lua rpm.expand() Add support for nested Lua macro expansion (RhBug:490740) Issue an error on failure to replace original package on signing Don't run collections on script stages like %pre/posttrans, ugh. Permit %verifyscript from non-installed packages Disable all scriptlets and collections centrally on --test and --justdb Both files must be colored for multilib conflict resolution (RhBug:705115) Colored conflict resolution part II Handle readlink() failure in genCpioListAndHeader() correctly Clean up + fix memleaks in readIcon() Clean up + plug memleak in parseDescription() Fix a logic error leading to unlink(NULL) call, oops. Reset cli configured flag on rpmcliFini() (RhBug:709421) Abort depgen output reading on EOF, not child exiting Handle EINTR on the spot instead of restarting the entire loop Eww, python ds.Instance() doesn't take any arguments Adjust script detection rules to work with file >= 5.07 too (RhBug:712251) Don't remove buildroot docdir on %doc usage (ticket #836) Export rpmteFailed() to python bindings Error on unclosed macros & trailing line continuations in spec (RhBug:681567) Support retrieving the spec contents in parsed format Add --parse option to rpmspec tool to dump parsed spec contents Avoid extra newlines in parsed spec output outside preamble Honor trailing slash in rpmGlob() Pay attention to dir vs file when building (RhBug:505995) Enable GLOB_ONLYDIR of the bundled glob() on platforms that support it Fix explicit directory %attr() when %defattr() is active (RhBug:481875) Fix the totally broken rpm.fd() read method Fix the broken python header getattr() behavior, take 13 (or so) zlib is mandatory, fail at configure if missing + remove conditionals Fix the sanity check on number of query/verify sources (RhBug:691930) Only increment number of query/verify sources when we encounter new types Oops, rpmPubkeyDig() should return NULL if pgpPrtPkts() fails Fix crash on PGP packets/armors with more than one key (RhBug:667582) Fix memleak on keys with more than one user id Sanity check signatures even if we dont have a key Shut up unused-but-set warnings from gcc (if supported) Preparing for 4.9.1 Remove ugly isDir recurse prevention hack on build pgpVerifySig() check of NULL hash is the wrong way around Strip trailing slash from paths at addFile() Preparing for 4.9.1.1 Only muck with signals on first and last db open/close Sanity check region offset in regionSwab() Sanity check region offset range on headerLoad() Preparing for 4.9.1.2 Sanity check region length on header load Verify the entire region trailer, not just its offset, is within data area Fix dribble length calculation on headerLoad() Specifically validate region tag on header import Differentiate between non-existent and invalid region tag Validate negated offsets too in headerVerifyInfo() Preparing for 4.9.1.3
Tero Aho (2): Fix uninitialized variable in fsm Plug a minor memleak on writeRPM() error paths
Ville Skyttä (8): Comment spelling fix. Avoid emitting empty perl() module deps. Documentation spelling fixes. Man page syntax fixes. Mention %bcond_with* in conditional build doc. Honor $TMPDIR in various scripts. Add lzip support. Add lrzip support.
Completed project 014 Animated Navigation
Completed the animated navigation project. Nothing special here. Very simply project. Honestly not my favourite one. The effect looks kinda gimmicky to be honest.
rpm: Autospec creation for update from version 4.16.1.2 to version 4.14.2.1
Alan Jenkins (2): ndb glue: closeEnv() should always clear rdb->db_dbenv rpmidxHandleObsolete(): Fix fd leak in error path
Andreas Schwab (1): debugedit: handle RISC-V relocation
Athos Ribeiro (1): Add missing option to rpmspec manpage
Bernhard M. Wiedemann (1): find-debuginfo.sh: sort output of find
Bernhard Rosenkränzer (1): Fix division by zero in prelink detection, issue 420
Colin Walters (1): COPYING: Minor grammar fixes
Filipe Brandenburger (1): Add shortcut to --nodebuginfo
Florian Festi (8): Fix formatting in man page Rename RPMVSF* constants to RPMVSF_MASK_* Use new RPMVSF_MASK_* constants Expose new RPMVSF_MASK_* constants in Python binding Use new RPMVSF_MASK_* constants in comment of the macros file Add --allow-empty to the initial commit of __scm_setup_git Add switch to disable systemd-inhibit plugin Call checkOwners() first to get useful error message on broken sources
Igor Gnatenko (3): metainfo.prov: scan /usr/share/metainfo and /usr/share/appdata for both types tests: Add tests for boolean dependencies brp-compress: compress fish manpages
Jan Pazdziora (2): Make python examples run with python 3, the print commands. rpm.spec's .prep seems to be an attribute.
Jeff Johnson (1): Add LMDB backend to RPM
Jun Aruga (1): Fix Python bindings library path for custom prefix.
Kir Kolyshkin (3): Factor out and unify setting CLOEXEC Optimize rpmSetCloseOnExec rpmSetCloseOnExec: use getrlimit()
Mark Wielaard (11): Rename addPackageRequires to addPackageDeps in build/files. Add debugsource recommends to debuginfo packages. Avoid warning about mbAppendStr if lui support isn't enabled. Fix rpmlog printing of off_t in fdConsume. Fix rpmlog warning in lzopen_internal from too big length specifier. debugedit: edit_dwarf2 check lndx is in range before checking r_offset. Support --jobs as an alias to -j in find-debuginfo.sh (RhBug:1518120) Fix type mismatch calculating new line program offset in debugedit.c. debugedit: Check .debug_str index is valid before use. debugedit: Only try to collect comp_dir in phase zero. find-debuginfo.sh: Handle application/x-pie-executable (file 5.33 output).
Michael Schroeder (7): Fix classification of ELF binaries with both setuid/setgid set Add support for 'unless' rich dependencies Allow rpmrichParse to be called with no callback Forbid 'if' richops in 'or' context and 'unless' richops in 'and' context Fix sigheader generation for big archives Fix inode handling for zero-sized files Improve hardlink handling in the disk space calculation
Michal Čihař (1): Follow man page conventions in rpm-misc.8
Mike Crowe (1): Make configure cope with multiple users/groups with ID 0
Miro Hrončok (3): brp-python-bytecompile: Run a pre-flight find before invoking $default_python Provide a way to opt out from Python bytecompilation out side of lib dir Python generators: Use nonstandardlib for purelib definition
Neal Gompa (2): ci: Add lmdb-devel to CI Dockerfile configure: Fix typo for libcrypto usability check for OpenSSL check
Nikola Forró (1): Only destroy lua tables if there are no BASpecs left
Panu Matilainen (191): Remember to add macro-scope.spec to tarballs Revert "Only build bundled fts if system has a bad version that doesn't handle LFS" Disable two of the sigpipe tests, they're too unstable Make the rpm 4.4.x compat layer removal explicit via an error message Revert "Support quoting of macro arguments (#222)" Add missing "il" to header sanity check error message Make _dbibyteswapped static within lmdb.c _DBSWAP() macro is identical between lmdb.c and db3.c, consolidate to dbi.h Add a flag to allow quiet test for package existence with lookupPackage() Use silent lookup for debuginfo packages (#1482144) Remove redundant max_macro_depth initialization Bump macro max recursion limit up Eliminate unnecessary "delete" argument to freeArgs() Allow running rpmbuild with debug verbosity Leave build scripts around in debug mode (#1442015) Export rpmsqBlock() to python bindings as blockSignals() Limit automatic fallback to DB_PRIVATE to read-only operations Fallback to DB_PRIVATE on readonly DB_VERSION_MISMATCH too (RhBug:1465809) Only fallback to MDB_RDONLY if readonly database is requested Eliminate a bunch of idiotic assert()'s in rpmtd getters Restore 4.13 behavior with single value tags in query format arrays Add more queryformat array testcases Make queryformat array behavior consistent for all types Add a few more parametric macro argument tests Support parametric macro argument quoting, take three (#222) Avoid macro expansion on every rpmlog() call + fix a related memleak Plug what's probably an ancient memleak in build code Soften the 4.4.x COMPAT error into a warning for 4.14.x Add documentation for all/most built-in macros Don't assume %{quote:...} is surrounded by whitespace Remove leftover parentheses, no functional change. Mark ndb and lmdb experimental in configure too Limit " has N files" debug diagnostics to install/erase goals Fix %transfiletriggerpostun diagnostic showing "unknown" Always execute scriptlet callbacks with owning header (RhBug:1485389) Use pkg-config for figuring python cflags and libs Eliminate the rest of idiotic assert()'s in rpmtd.c Sync disks at the end of transactions (RhBug:1461765) Revert "Sync disks at the end of transactions (RhBug:1461765)" Fix false negatives on signature/digest tests inside "make distcheck" Enable python build during dist-check Fix PYTHONPATH in test-suite when prefix differs from system python (#265) Preparing for rpm 4.14.0-rc1 Fix file triggers failing to match on some packages (MgBug:18797) Avoid multiple strlen() calls on the constant prefix string when searching Fix Ftell() past 2GB on 32bit architectures (RhBug:1492587) Add testcases for unpackaged files and directories detection Don't follow symlinks on file creation (CVE-2017-7501) Fix excessive dependencies for elfdeps and sepdebugcrcfix Add configuration option for controlling file IO flushing behavior Create $(prefix)/rpm/macros.d directory on make install Add + use a separate helper function for debuginfo test-cases Restrict following symlinks to directories by ownership (CVE-2017-7500) Make verification match the new restricted directory symlink behavior Unbreak make distcheck again And another test-suite related distcheck regression fix... Preparing for rpm 4.14.0-rc2 Always initialize hdrp retval of rpmReadPackageFile() Ensure all indexes get created on --rebuilddb Remove excessive linking from our cli tools Don't spew out usage message just because rpm is in verbose mode Fix "substatus" (eg NOKEY) regression on package read (#330) Add tests for signed package read for both NOKEY and OK cases Avoid local path in testsuite expected results, oops Update CREDITS for new contributors since 4.13.0 Preparing for rpm 4.14.0 Don't assume per-user groups in test-suite Switch from Transifex to Zanata Test for lsetxattr() presence, don't try building IMA plugin without it Remove find-lang.sh embedded changelog, merge authors to CREDITS Add actual command + arguments to dependency generator debug output (#338) Error out on --enable-ndb if system doesn't support mremap() (#334) Generalize argvAddDir() into a helper supporting arbitrary virtual attributes Add support for new virtual file attribute "%artifact" Check for getline() as a required function in configure Add tests for file attribute filtering in query and verification Only fallback to read-only locking on EACCES (RhBug:1502134 revisited) Mark build-id and dwz entries in packages as %artifact files Fix not all %transfiletriggerpostun triggers executing (RhBug:1514085) Don't require signature header to be in single contiguous region (#270) Don't segfault on NULL fi in generated rpmfi iterators, doh Fixups to file trigger internal api docs + comments Fix file lists getting fed to file triggers multiple times (#370) Add a testcase for issue #370 Fixup transfiletrigger comments Move %transfiletriggerun after problem checking and fingerprinting Explicitly skip non-installed files on erasure Don't fire file triggers on non-installed files Fix %ghost/%missingok files causing bogus verify failures (RhBug:1533092) Preparing for rpm 4.14.1 Add rpmcliVSFlags which is populated with the actual vsflags from popt Grab number of elements in transaction into a local helper variable Allow verification callbacks to signal stopping, document the callback Typecheck header tags match our definitions prior to import (#242, #414) Handle missing keyring same as key not found in rpmKeyringVerifySig() Further relax type checking among different string tag types (#414) Streamline digest/signature info initialization Always honor vsflags when appending new items to verify set Use the same verify callback for verbose and non-verbose operation Store digest/signature verify results in the sinfo struct directly Split verify set creation to separate allocation + init steps Store digest bundle handle in verify set as part of initialization Store keyring handle in verify set as part of creation Remove now unnecessary flags test on payload digest append Add getter for verify set vsflags, use it for the one external access Pass verify sets to rpmpkgRead() instead of flags and keyrings Remember unfinalized digest context in rpmsinfo struct Separate signature/digest verification from the digest finalization Ignore non-signature tags on rpmvsInit() PGP 5 is rather ancient history, don't bother Drop unused leadtype argument to applyRetrofits() Plug the last data "leak" from rpmlead Add another corrupted header signature/digest testcase Test for rpmkeys return codes too in [rpmkeys -Kv 1] Add a separate vsflags macro for package (signature/digest) verification Add internal API for allocating and freeing hdrblob objects Use allocated hdr blobs in rpmpkgRead() instead of local structs Lift logging out of rpmpkgRead() to callers "sigset" is a terrible variable name since it's ambigious, rename Separate verification & other extra logic from package reading Make retrofit etc helpers static again now that we can Attempt to (re)verify anything that has a digest context Identify non-present signatures too (based on their assumed algo) Preliminaries for reporting non-present verification items Sort verify set results by type as well, signatures before digests Only init/fini verify ranges we're using Implement configurable, mandatory signature/digest verify level for rpm -K Handle system verify level centrally and internally in rpmvs Make verify set sorting more precise (and hopefully, more useful) Add package verification step to transactions Switch verify level to digest by default Plug a memory leak from commit 170bc61a337f74a3619be0357bb3da96306c2769 Create transaction set only after initializing rpm itself, oops Eliminate weirdo "usesys" argument from rpmvsVerify() Move verify level configuration to transaction set public API Make --nosignature/--nodigest affect verification level too Add Python bindings + simple testcase for the verification level API Oops, add verification callback symbols to python bindings Rename vslevel back to vfylevel everywhere, sigh Add a public API for controlling package verification flags Eliminate rpmcliVSFlags from core transaction code Add rudimentary documentation to the vfylevel and -flags apis Macroize build root policies for consistent disable/override ability Don't bother falling back to gpg from gpg2 in test-suite Fix misleading error message when missing TTY (#355) Avoid hard dependency on python via the python macro helper script (#387) Fix memleak on gpg-pubkey header generation Only use offset optimization for string types in HEADERIMPORT_FAST (#398) Drop a redundant binary entry size assignment Clarify Fread() and Fwrite() semantics Make rpmErase() exit code resemble rpmInstall() exit code Use rpm's non-failing allocators for ndb too (#357) Improve rpm2cpio examples a bit (RhBug:1553898) Make python addSign() and delSign() actually work (RhBug:1558126) Remember to include macros.testfile in dist tarballs Fix a leak and a race in rpmdb open/close handling Use just numeric uid/gid for user/group verification Expect failure for invalid directory symlink fails when run as root Really fix CVE-2017-7500. Ugh. Add support for --whatconflicts for good measure Log a more specific error on scriptlet exec() failure Make test-suite work with coreutils multicall binary too Adjust file verify to really follow CVE-2017-7500 rules, doh Adjust testsuite for CVE-2017-7500 verify tweak, deep sigh Fix out-of-tree build regression Preparing for rpm 4.14.2-rc1 Fix a memleak in file verification Fix ancient FILE leak in file manifest reading Fix leak on dlopen()'ed handle in case our _hooks symbol is not found Fix leaking opendir() handle on database rebuild Fix an ancient leak and missing error logging on src.rpm create failure Fix ancient memleaks on buildRoot re-expansion sanity checks Copy DISTTAG to source rpms too if present Fix an obvious typo in dwz filename generation Remove redundant rpmGlob() retval condition Fix memleak in changelog full timestamp parsing Fix ancient memleak in spec preamble parse Fix rpmlog error handling regressions from colorization (RhBug:1597274) Clean up and improve rpmlog error handling + reporting Preparing for rpm 4.14.2-rc2 Fix ancient memleaks on error paths of expression parsing Include expected output in more install test-cases Document tag=pattern query/verify selector in the manual (#451) Actually remember weak dependencies in transaction elements (RhBug:1593185) Increase order test coverage Preparing for rpm 4.14.2 Fix a blindingly obvious memleak in package verify step Fix ancient python GIL locking bug on callback (RhBug:1632488) Resurrect long since broken Lua library path Fix nasty --setperms/--setugids regression in 4.14.2 (RhBug: 1640470) Preparing for rpm 4.14.2.1
Pavlina Moravcova Varekova (17): Unified messages printed by %{echo:...}, %{warn:...}, and %{error:...} Make macro %{error:...} cause an actual failure %trace mb states at the beginning/end of expandMacro must be equal Bring manual for %{echo:...}, %{warn:...} , %{error:...} up to date When printMacro() is used, print newline after "(empty)" macro Add documentation for %load macro Move %define and %undefine tests into rpmmacro.at and simplify them Don't expand %{verbose:...} argument on false condition Fix progress bar printing for packages with wrong archive size (RhBug:1478051) Add --queryformats for displaying numbers human readable (#375) Fix a macro end detection if the first char of a macro line is '}' (#401) Update Query formats documentation Correct Query formats documentation Add popt-based options --setcaps and --restore (RhBug:1550745) Remove misleading --setperms and --setugids warnings (RhBug: 1538610) Fix creating bogus requirements by perl.req (RhBug: 1539344) Add missing documentation for --jobs in find-debuginfo.sh
Per Øyvind Karlsen (1): Add support for passing multiple names to find-lang.sh
Peter Kjellerstedt (1): Add a new option --alldeps to rpmdeps
Piotr Drąg (1): Update Polish plural forms
Richard Purdie (1): Remove dubious condition on scriptlet exec()
Ruediger Oertel (1): Define DistTag as optional tag with macro just like DistURL
Stefan Berger (4): Split off function wfd_close() to close a file Split off function wfd_open() to open a file Create first hard link file and keep open, write it at end Remove now redundant code and parameters from expandRegular()
Thierry Vignaud (1): Add support for --whatobsoletes
Ville Skyttä (4): Use separate script instead of python -c to retrieve %python_* values Get %python_version from platform.python_version_tuple python-macro-helper: Drop -tt from shebang Invoke python-macro-helper with -Es python args