Make linked test cases point to the Internet Archive

[timvandermeij]

We have quite a number of linked test cases in the https://github.com/mozilla/pdf.js/tree/master/test/pdfs folder. Some of them point to sources other than the Internet Archive or Bugzilla and might become unavailable, which is to be avoided.

It is better to fetch the files from the Internet Archive to be sure that they remain available. This should always be possible if the current source exists because once you input the link on the Internet Archive website it will archive the file automatically.

Below is a list of linked test cases that point to external sources and should be updated (obtained with grep -L "bugzilla.mozilla.org\|archive.org" *.pdf.link). Make sure that the hash of the Internet Archive file is the same as the hash in the test manifest.

• 20130226130259.pdf.link
• JST2007-5.pdf.link
• P020121130574743273239.pdf.link
• SFAA_Japanese.pdf.link
• TaroUTR50SortedList112.pdf.link
• aboutstacks.pdf.link
• artofwar.pdf.link
• bpl13210.pdf.link
• bug766138.pdf.link
• bug808084.pdf.link
• bug887152.pdf.link
• cable.pdf.link
• ecma262.pdf.link
• fips197.pdf.link
• fit11-talk.pdf.link
• geothermal.pdf.link
• hmm.pdf.link
• html5checker.pdf.link
• hudsonsurvey.pdf.link
• ichiji.pdf.link
• issue1010.pdf.link
• issue1015.pdf.link
• issue1049.pdf.link
• issue1096.pdf.link
• issue1127.pdf.link
• issue1133.pdf.link
• issue1155.pdf.link
• issue1169.pdf.link
• issue1233.pdf.link
• issue1257.pdf.link
• issue1309.pdf.link
• issue1317.pdf.link
• issue1419.pdf.link
• issue1466.pdf.link
• issue1597.pdf.link
• issue1629.pdf.link
• issue1658.pdf.link
• issue1685.pdf.link
• issue1687.pdf.link
• issue1709.pdf.link
• issue1721.pdf.link
• issue1729.pdf.link
• issue1796.pdf.link
• issue1810.pdf.link
• issue1878.pdf.link
• issue1912.pdf.link
• issue1936.pdf.link
• issue1998.pdf.link
• issue2006.pdf.link
• issue2129.pdf.link
• issue2139.pdf.link
• issue2386.pdf.link
• issue2442.pdf.link
• issue2531.pdf.link
• issue2537.pdf.link
• issue2627.pdf.link
• issue2799.pdf.link
• issue2829.pdf.link
• issue2853.pdf.link
• issue2881.pdf.link
• issue3062.pdf.link
• issue3384.pdf.link
• issue3666.pdf.link
• issue3848.pdf.link
• issue3903.pdf.link
• issue3925.pdf.link
• issue3999.pdf.link
• issue4926.pdf.link
• issue5592.pdf.link
• issue5726.pdf.link
• issue818.pdf.link
• issue919.pdf.link
• jai.pdf.link
• kdchart.pdf.link
• liveprogramming.pdf.link
• mao.pdf.link
• ocs.pdf.link
• ohkubo-SS04.pdf.link
• pal-o47.pdf.link
• pdf.pdf.link
• piperine.pdf.link
• preistabelle.pdf.link
• protectip.pdf.link
• tcpdf_033.pdf.link
• tutorial.pdf.link
• txt2pdf.pdf.link
• unix01.pdf.link
• usmanm-bad.pdf.link
• vesta.pdf.link
• yo01.pdf.link

Additionally:

• Investigate hmm.pdf. The MD5 hash in the manifest matches the MD5 hash of the file from the link, but the bots complain about a mismatch.

[Snuffleupagus]

Regarding issue1155: I'm not having any luck in obtaining a working link from the Internet Archive. However, looking at the issue/PR (and quickly browsing through the structure of the PDF file), it seems that it shouldn't be too hard to hand-craft a reduced test-case, so I'll look into that later. Fixed

[Snuffleupagus]

It appears that we've now reached the point where every file, perhaps with one or two exceptions, that is possible to obtain through the Internet Archive have been fixed.

Below is a summary of the ones that are left to fix, so we probably need to start looking into the possibility of replacing them with reduced test-cases, if at all possible.

A number of these files are font related, which in theory should make it reasonable easy (albeit time-consuming) to create/verify reduced test-cases. (It might be simplest to see if Brendan manages to create the tool mentioned in #6650 (comment), before undertaking more work on that.)

test/pdfs/issue2799.pdf.link
test/pdfs/issue5726.pdf.link
test/pdfs/yo01.pdf.link