/botio test

From: Bot.io (Linux m4)

Received

Command cmd_test from @calixteman received. Current queue size: 0

Live output at: http://54.241.84.105:8877/cc368382cf23509/output.txt

From: Bot.io (Windows)

Received

Command cmd_test from @calixteman received. Current queue size: 0

Live output at: http://54.193.163.58:8877/82c7bed4e68d2f5/output.txt

From: Bot.io (Linux m4)

Failed

Full output at http://54.241.84.105:8877/cc368382cf23509/output.txt

Total script time: 42.07 mins

• Unit tests: Passed
• Integration Tests: FAILED
• Regression tests: FAILED

  different ref/snapshot: 2

Image differences available at: http://54.241.84.105:8877/cc368382cf23509/reftest-analyzer.html#web=eq.log

Our js jbig2 decoder is buggy with this file.
That being said, the "regression" is because of const elem = srcPos < srcLength ? src[srcPos] : 255; changed in to const elem = src[srcPos];. For sure the perf win of removing the branching in this expression is probably not that much but it'd a pity to have a penalty for something useless.

Given that the original rendering using the JS decoder was also incorrect, I don't really mind that this patch changes only this particular file (and none of the others) to a "different kind of incorrect" so to say, mostly because the WASM decoder handles it well and we try to use that as the default where possible anyway.

From: Bot.io (Windows)

Failed

Full output at http://54.193.163.58:8877/82c7bed4e68d2f5/output.txt

Total script time: 83.99 mins

• Unit tests: FAILED
• Integration Tests: FAILED
• Regression tests: FAILED

  different ref/snapshot: 2

Image differences available at: http://54.193.163.58:8877/82c7bed4e68d2f5/reftest-analyzer.html#web=eq.log

Our js jbig2 decoder is buggy with this file.

At the risk of asking stupid questions: Can the PDF.js JBIG2 decoder be replaced with a fallback JS one based on the PDFium decoder?
Looking at PR #19525 that was feasible for the compiled JPEG2000 decoder, so might it be possible for JBIG2 as well to reduce the amount of code that needs to be maintained here and improve rendering consistency?

Our js jbig2 decoder is buggy with this file.

At the risk of asking stupid questions: Can the PDF.js JBIG2 decoder be replaced with a fallback JS one based on the PDFium decoder? Looking at PR #19525 that was feasible for the compiled JPEG2000 decoder, so might it be possible for JBIG2 as well to reduce the amount of code that needs to be maintained here and improve rendering consistency?

No it isn't a stupid question.
I didn't make a js version of the decoder because I thought it'd be bigger but I didn't check. I'll check that, I'll let you know and we'll see. But for sure it'll make things more consistent.

/botio makeref

From: Bot.io (Windows)

Received

Command cmd_makeref from @calixteman received. Current queue size: 2

Live output at: http://54.193.163.58:8877/7cdcce40dcee3e8/output.txt

From: Bot.io (Linux m4)

Received

Command cmd_makeref from @calixteman received. Current queue size: 0

Live output at: http://54.241.84.105:8877/36fade69c44c3c8/output.txt

From: Bot.io (Linux m4)

Success

Full output at http://54.241.84.105:8877/36fade69c44c3c8/output.txt

Total script time: 18.61 mins

• Make references: Passed
• Check references: Passed