Ensure that abbreviated dictionary keys always take precedence in inline images (issue 14256) #14257
Conversation
|
Alternatively, should we reverse the get's? e.g. |
Initially I thought about doing that instead, but it'd have required edits all over the code-base and seems much more error-prone with any future changes since you'd need to ensure that the correct ordering is maintained everywhere. Edit: Although looking at this again, I suppose that the patch as-is could miss some cases... |
Snuffleupagus
force-pushed
the
issue-14256
branch
from
7414a0e to
8c2139e
Compare
|
/botio test |
From: Bot.io (Windows)ReceivedCommand cmd_test from @Snuffleupagus received. Current queue size: 0 Live output at: http://54.193.163.58:8877/9c4c6a83d1a9001/output.txt |
From: Bot.io (Linux m4)ReceivedCommand cmd_test from @Snuffleupagus received. Current queue size: 0 Live output at: http://54.241.84.105:8877/448ae22931c3b71/output.txt |
|
Based of this comment it makes me think we should always do short name first even for other dictionaries. We could enforce this when using .get() with multiple arguments. |
From: Bot.io (Linux m4)FailedFull output at http://54.241.84.105:8877/448ae22931c3b71/output.txt Total script time: 23.46 mins
Image differences available at: http://54.241.84.105:8877/448ae22931c3b71/reftest-analyzer.html#web=eq.log |
|
There's also |
From: Bot.io (Windows)FailedFull output at http://54.193.163.58:8877/9c4c6a83d1a9001/output.txt Total script time: 42.28 mins
Image differences available at: http://54.193.163.58:8877/9c4c6a83d1a9001/reftest-analyzer.html#web=eq.log |
@brendandahl So I'm guessing that you'd prefer something along these lines instead: bfae1a3?w=1 |
…onary lookup (issue 14256) Note that issue 14256 was specifically about *inline* images, please refer to: - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#G7.1852045 - https://www.pdfa.org/safedocs-unearths-pdf-inline-image-issue/ - https://pdf-issues.pdfa.org/32000-2-2020/clause08.html#H8.9.7 However, during review of the initial PR in mozilla#14257 (comment), it was suggested that we instead do this *unconditionally for all* dictionary lookups. In addition to re-ordering the existing call-sites in the `src/core`-code, and adding non-PRODUCTION/TESTING asserts to catch future errors, for consistency a number of existing `if`/`switch`-blocks were re-factored to also check the abbreviated keys first.
…onary lookup (issue 14256) Note that issue 14256 was specifically about *inline* images, please refer to: - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#G7.1852045 - https://www.pdfa.org/safedocs-unearths-pdf-inline-image-issue/ - https://pdf-issues.pdfa.org/32000-2-2020/clause08.html#H8.9.7 However, during review of the initial PR in mozilla#14257 (comment), it was suggested that we instead do this *unconditionally for all* dictionary lookups. In addition to re-ordering the existing call-sites in the `src/core`-code, and adding non-PRODUCTION/TESTING asserts to catch future errors, for consistency a number of existing `if`/`switch`-blocks were re-factored to also check the abbreviated keys first.
…nary lookups (issue 14256) Note that issue 14256 was specifically about *inline* images, please refer to: - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#G7.1852045 - https://www.pdfa.org/safedocs-unearths-pdf-inline-image-issue/ - https://pdf-issues.pdfa.org/32000-2-2020/clause08.html#H8.9.7 However, during review of the initial PR in mozilla#14257 (comment), it was suggested that we instead do this *unconditionally for all* dictionary lookups. In addition to re-ordering the existing call-sites in the `src/core`-code, and adding non-PRODUCTION/TESTING asserts to catch future errors, for consistency a number of existing `if`/`switch`-blocks were re-factored to also check the abbreviated keys first.
…nary lookups (issue 14256) Note that issue 14256 was specifically about *inline* images, please refer to: - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#G7.1852045 - https://www.pdfa.org/safedocs-unearths-pdf-inline-image-issue/ - https://pdf-issues.pdfa.org/32000-2-2020/clause08.html#H8.9.7 However, during review of the initial PR in mozilla#14257 (comment), it was suggested that we instead do this *unconditionally for all* dictionary lookups. In addition to re-ordering the existing call-sites in the `src/core`-code, and adding non-PRODUCTION/TESTING asserts to catch future errors, for consistency a number of existing `if`/`switch`-blocks were re-factored to also check the abbreviated keys first.
…nary lookups (issue 14256) Note that issue 14256 was specifically about *inline* images, please refer to: - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#G7.1852045 - https://www.pdfa.org/safedocs-unearths-pdf-inline-image-issue/ - https://pdf-issues.pdfa.org/32000-2-2020/clause08.html#H8.9.7 However, during review of the initial PR in mozilla#14257 (comment), it was suggested that we instead do this *unconditionally for all* dictionary lookups. In addition to re-ordering the existing call-sites in the `src/core`-code, and adding non-PRODUCTION/TESTING asserts to catch future errors, for consistency a number of existing `if`/`switch`-blocks were re-factored to also check the abbreviated keys first.
…nary lookups (issue 14256) Note that issue 14256 was specifically about *inline* images, please refer to: - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#G7.1852045 - https://www.pdfa.org/safedocs-unearths-pdf-inline-image-issue/ - https://pdf-issues.pdfa.org/32000-2-2020/clause08.html#H8.9.7 However, during review of the initial PR in mozilla#14257 (comment), it was suggested that we instead do this *unconditionally for all* dictionary lookups. In addition to re-ordering the existing call-sites in the `src/core`-code, and adding non-PRODUCTION/TESTING asserts to catch future errors, for consistency a number of existing `if`/`switch`-blocks were re-factored to also check the abbreviated keys first.
…nary lookups (issue 14256) Note that issue 14256 was specifically about *inline* images, please refer to: - https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#G7.1852045 - https://www.pdfa.org/safedocs-unearths-pdf-inline-image-issue/ - https://pdf-issues.pdfa.org/32000-2-2020/clause08.html#H8.9.7 However, during review of the initial PR in mozilla#14257 (comment), it was suggested that we instead do this *unconditionally for all* dictionary lookups. In addition to re-ordering the existing call-sites in the `src/core`-code, and adding non-PRODUCTION/TESTING asserts to catch future errors, for consistency a number of existing `if`/`switch`-blocks were re-factored to also check the abbreviated keys first.
Please refer to:
Fixes #14256