Sender | Message | Time |
---|---|---|
18 May 2024 | ||
HeveraletLaidCenx | I noticed that in MDN's font-family reference, the specification has already updated to "CSS Fonts Module Level 4https://drafts.csswg.org/css-fonts/#generic-font-families", there's something about the character coverage showed in the pic above, can I assume that is FIrefox's UA does some extra processing on generic font families(like monospace in the example here, which caused the result? It may solves the fallback character coverage problem to some extent, but sacrifices the freedom of allows users to freely configure the uncovered characters' font family... I'm not sure where to find related progress or design for Firefox or Gecko either | 05:28:03 |
HeveraletLaidCenx | @Caspy7 Thanks for your explanation, got it, but for now, it seems that as a front-end developer, I need to put more effort into compatibility testing for such font family consistency 😭 | 05:39:21 |
Caspy7 | HeveraletLaidCenx: I'd say bring this up in #webcompat:mozilla.org When someone can look at it they can give you a good idea of if this is in error or intentional or needs improved | 05:40:45 |
HeveraletLaidCenx | In reply to @caspy7:mozilla.orgThx! I'll try | 05:43:53 |
strike256 joined the room. | 05:51:14 | |
strike256 | Redacted or Malformed Event | 05:58:23 |
strike256 | Has there been a regression, or incomplete docs? Why is this here and it links to nothing(in the underlying html it seems to be a hidden section?). https://support.mozilla.org/en-US/kb/task-manager-tabs-or-extensions-are-slowing-firefox#firefox:win10:fx128 | 05:58:25 |
strike256 | Download image.png | 05:59:50 |
strike256 | * Has there been a regression, or incomplete docs? Why is this here and it links to nothing(in the underlying html it seems to be a hidden section?). https://support.mozilla.org/en-US/kb/task-manager-tabs-or-extensions-are-slowing-firefox#firefox:win10:fx126 | 06:00:35 |
strike256 | Currently, there seems to be no simple way to check the memory and CPU usage of extensions in Firefox. For the average user, without profiling extensions, this does not seem to help in regarding to finding performance problems. | 06:02:36 |
strike256 | I vaguely remember in earlier versions of firefox that was not an issue. | 06:03:26 |
strike256 | is there an preference in about:config that would allow a user to see this or is it a full regression? | 06:05:09 |
strike256 | * I vaguely remember in earlier versions of Firefox that was not an issue. | 06:05:32 |
GrayShade | In reply to @strike256:matrix.orgI don't remember that | 06:17:24 |
Caspy7 | strike256: might also inquire in #addons:mozilla.org | 06:21:05 |
GrayShade | https://superuser.com/posts/1088698/revisions shows the history of that page. It used to show the memory usage, but I don't remember seeing the CPU | 06:21:15 |
strike256 | In reply to @grayshade:dend.roI think they were individual processes instead of being lumped together, let me find a screenshot | 06:22:44 |
strike256 | Download image.png | 06:22:49 |
GrayShade | Yeah, but where's the CPU usage? | 06:23:09 |
GrayShade | The profiler icon might be more useful anyway. I tried it, but didn't get anything interesting with my extensions | 06:25:05 |
strike256 | you can also see this if you deselect the customization for the later features of Firefox. | 06:25:24 |
strike256 | https://support.mozilla.org/en-US/kb/task-manager-tabs-or-extensions-are-slowing-firefox | 06:25:25 |
strike256 | * you can also see this if you deselect the customization for the current version of Firefox. | 06:25:39 |
strike256 | In reply to @caspy7:mozilla.orgWill do! | 06:26:11 |
GrayShade |
Where does that link go? | 06:26:11 |
strike256 | In reply to @grayshade:dend.roenergy impact = Energy Impact: The CPU processing power being used by each task. Tasks with a higher energy impact will drain your battery quicker and may slow down the performance of your system. | 06:26:37 |
strike256 | In reply to @grayshade:dend.roto https://support.mozilla.org/en-US/kb/task-manager-tabs-or-extensions-are-slowing-firefox#w_task-manager but that section seems to be hidden because it's no longer up to date for the latest version of firefox. (which poses the question why that part isn't hidden) | 06:30:42 |
strike256 | Download image.png | 06:31:23 |
HeveraletLaidCenx | emilio: > <@heveraletlcm:matrix.org> I'd like to confirm if the inconsistency of the `font-family` property is a bug or designed to do so, and are there any other info about it. Are you sure that SimSun is not what you get if you just use sans-serif? emilio: Does the behaviour change with the lang attribute? You: > <@emilio:mozilla.org> Are you sure that SimSun is not what you get if you just use sans-serif? I tested and I'm sure SimSun is not the value of sans-serif, and it is a serif font family instead, the value of sans-serif is Arial and "Microsoft YaHei" when lang=en, is "Microsoft YaHei" only when lang=zh-Hans. (OS is Win 11, system language=en-US) emilio: Is SimSun a monospace font? emilio: If so it might be intentional emilio: But it seems worth filing a bug if not emilio: And maybe even if so emilio: Because might deserve some looking into from someone that knows our font fallback code better emilio: I think this explains the behavior: https://searchfox.org/mozilla-central/rev/17a31baccea968806230cc6dbe8d5700937370d5/modules/libpref/init/all.js#2227 emilio: Basically SimSun is in our monospace font list, so we never fall back to sans serif emilio: (Probably) You: in my understanding, the behaviour now seems like: Firefox or Gecko add a default implicitly fallback font-famiy for those CJK full-width character set as `SimSun` for the `monospace` generic font family, which override those font-family style set as fallback manually(sans-serif in the example above). But I'm not sure how it happens, is it related to the new CSS spec or maybe some feature or adjustment added by Firefox itself for accessibility purpose? emilio: No, generic fonts may fall back to different fonts in gecko by definition emilio: That's something that the spec allows emilio: In any case it might be worth filing a bug with some of these details in case there's something to change? You: hmm, hold on please, I'm doing additional test for the behaviour across different browsers... You: @emilio The result it weird... 🖼️ HeveraletLaidCenx sent a picture You: I tried to use another unique font-family as the fallback value, and this time it worked You: seems the problem now focused to: if using generic font family as fallback value, if it should be work You: > <@heveraletlcm:matrix.org> seems the problem now focused to: if using generic font family as fallback value, if it should be work I think it it was designed not to be like this, and apology for the mistake made above: I was misled by the default font family config of other browsers, I didn't do well in keep factors constant, which their default `monospace` font family is similar as `san-serif` should be, in fact both of them in such scene of **multiple generic font families** are the same: **all of them will only consider the first one**, but with exceptions, which in my tests, the "fangsong" is flagged as a generic font family in MDN, but it can worked when set "monospace, fangsong;" when your device has that font installed. In addition, not only the lang attribute of HTML, but also the language of the browser itself and the OS language of the device seem to have an impact on the behaviour. I think maybe some additional testing and clarification about this should be done, but it may be the work of MDN instead of Firefox itself. | 08:51:00 |
HeveraletLaidCenx | @Caspy7 I had some further discussions on the compatibility channel, seems not the problem of Firefox, my initial expectations of behavior were wrong. Maybe there should be some work on the related doc and reference I think. Anyway, this part of the behavior is a bit of complicated. | 08:54:36 |