Fix generation of large PDFs

[clementmas]

Browsershot can not currently save PDFs that are over 512 MB.

Browsershot::url('https://page-containing-large-images.com')->save('large-file.pdf');

The issue is described quite well here: microsoft/playwright#14675

The error Cannot create a string longer than 0x1fffffe8 characters occurs at output.toString('base64') which can not handle very large output (see Stack Overflow thread).

However, we don't need to convert the output to base64 when saving PDF files to disk. If we did, we should serialize the output to buffer but I don't think this is necessary.

Fixes #642 and #745.

[freekmurze]

Thanks! It seems like the tests are failing, could you take a look at that?

[clementmas]

The test "can evaluate page" was failing because it still needs to receive the encoded the output.

So I moved the encoding condition inside the getOutput function. All the tests are now passing. I should have checked earlier.

I also added node_modules inside .gitignore to make it easier to run npm i puppeteer locally before submitting PRs.

[clementmas]

I adapted my PR to support the new getOutput() method from #780

[rw152]

Will this be merged? Seems like a helpful addition.

[freekmurze]

Could you rebase this with main?

[clementmas]

I rebased with main (v4). The tests are failing though but it's unrelated to this update.

[freekmurze]

There are still some conflicts that prevent me from merging. Could you rebase again with main?

[clementmas]

I rebased with main again. If this doesn't work, I'll open a new fresh PR.

[freekmurze]

Best to open a new PR. Unfortunately, GitHub still detects conflicts, and won't let me merge this.