LibGfx/JBIG2+Tests/LibGfx: Add JBIG2 tests for text refinement with custom tables #26394
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
github-actions
bot
added
the
👀 pr-needs-review
nico
force-pushed
the
jbig2-stuff
branch
2 times, most recently
from
d9e5770 to
e9b41f5
Compare
The custom tables in this test are meant to be the default tables with some rows swapped, but I had a typo in the tables that are meant to be B.1 with rows swapped: B.1 has a lowest value of 0, not 1. No real behavior change (the test checked what it was supposed to check with the "wrong" value too), just pedantry.
* The huffman "refine one" test was added in SerenityOS#26388 * We now have tests for everything tested by Annex H except for a symbol refinement segment referring to itself, so use a more precise bullet for just that
This makes no difference when the default table is used, but it does make a difference if a custom table is present.
There can be up to four custom huffman tables for refinement in text regions. Text regions also must refer to at least one symbol dictionary, so customizing all refinement tables means that a text region refers to five other segments. That causes issues, especially in other decoders, see SerenityOS#26393. Also, Acrobat Reader apparently can't handle if the table that stores refinement data sizes is customized. So instead of adding a single test that customizes all four tables, add a bunch of tests, with the goal of most of them decoding in several other JBIG2 decoders. 1. Add a file with custom tables for refinement position. Decodes fine in PDFium, Acrobat Reader. Decodes mostly fine in Preview.app. 2. Add a file with custom tables for refinement dimensions. Decodes fine in PDFium, Acrobat Reader. Decodes mostly fine in Preview.app. 3. Add a file with custom tables for refinement data size. Decodes fine in PDFium. Decodes mostly fine in Preview.app. 4. Add a file with custom tables for refinement position and dimensions. This has 5 referred-to segments. Decodes fine in Acrobat Reader (and HEAD PDFium). Decodes mostly fine in Preview.app. 5. Finally, add one with all customized. PDFium can decode this as of a few hours ago, but nothing else can. (Preview.app comes fairly close.) Given that it's just a combination of tests 1-4 and those all decode fine somewhere, there's some hope to believe that the test itself is correct. For the contents of the custom huffman tables, I used the same contents as in the default huffman tables (B.14 for the first four, B.1 for the last), but changed them a bit. For B.14, I put the 1-length symbol in each of the four slots that it isn't in the normal B.14 table, and for B.1 I swapped the first two entries.
github-actions
bot
removed
the
👀 pr-needs-review
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There can be up to four custom huffman tables for refinement in
text regions. Text regions also must refer to at least one symbol
dictinary, so customizing all refinement tables means that a text
region refers to five other segments.
That causes issues, especially in other decoders, see #26393.
Also, Acrobat Reader apparently can't handle if the table that stores
refinement data sizes is customized.
So instead of adding a single test that customizes all four tables,
add a bunch of tests, with the goal of most of them decoding in
several other JBIG2 decoders.
Add a file with custom tables for refinement position.
Decodes fine in PDFium, Acrobat Reader.
Decodes mostly fine in Preview.app.
Add a file with custom tables for refinement dimensions.
Decodes fine in PDFium, Acrobat Reader.
Decodes mostly fine in Preview.app.
Add a file with custom tables for refinement data size.
Decodes fine in PDFium.
Decodes mostly fine in Preview.app.
Add a file with custom tables for refinement position and
dimensions. This has 5 referred-to segments.
Decodes fine in Acrobat Reader (and HEAD PDFium).
Decodes mostly fine in Preview.app.
Finally, add one with all customized.
PDFium can decode this as of a few hours ago, but nothing else can.
(Preview.app comes fairly close.) Given that it's just a combination
of tests 1-4 and those all decode fine somewhere, there's some hope
to believe that the test itself is correct.
For the contents of the custom huffman tables, I used the same
contents as in the default huffman tables (B.14 for the first four,
B.1 for the last), but changed them a bit. For B.14, I put the 1-length
symbol in each of the four slots that it isn't in the normal B.14 table,
and for B.1 I swapped the first two entries.
Also, fix a bug in our implementation of writing custom tables that these tests found :^)