| 09:27 | <smaug> | Content-type parsing is nicely inconsistent in all the implementations. |
| 09:27 | <smaug> | And even wpt tests look buggy |
| 15:48 | <zcorpan> | Yoav Weiss: thanks for the review for sizes=auto! What do you think about the association concept? Storing a pointer on the attribute also feels weird because the attribute can be omitted. Storing it on the <source> element still seems wrong because it could apply to multiple <img>s e.g. <picture><source srcset='x 50w, y 100w' sizes=auto><img width=500><img width=1000></picture>. Maybe the parsed sizes result should be stored on the img? |
| 15:58 | <zcorpan> | I'll ponder on it and come up with something tomorrow :) |
| 19:23 | <smaug> | annevk or anyone familiar with the setup for https://searchfox.org/mozilla-central/source/testing/web-platform/tests/mimesniff/mime-types/resources/generated-mime-types.json and tests around it |
| 19:23 | <smaug> | looks like the tests are expecting that blob.type doesn't modify the value |
| 19:24 | <smaug> | but https://w3c.github.io/FileAPI/#dfn-type at least says that it should be always lower case |
| 19:24 | <smaug> | and hints also that validation is done too |
| 19:25 | <smaug> | webkit seems to have currently rather reasonable behavior |
| 19:26 | <smaug> | But I'm not sure how to modify generated-mime-types.json |
| 19:28 | <smaug> | https://wpt.fyi/results/xhr/overridemimetype-blob.html?label=master&label=experimental&aligned has results for mime-types.json and https://wpt.fyi/results/mimesniff/mime-types/parsing.any.html?label=experimental&label=master&aligned also for generated-mime-types.json |