Some unicodes are not being parse or handled by Chrome V49
Reported by
pierluc....@gmail.com,
Mar 23 2016
|
||||||
Issue descriptionUserAgent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/49.0.2623.87 Safari/537.36 Example URL: http://www.tamasoft.co.jp/en/general-info/unicode.html Steps to reproduce the problem: 1. Navigate to http://www.tamasoft.co.jp/en/general-info/unicode.html 2. Go see unicodes for FE00 row 3. In Chrome, FE0000 to FE000F are not showing up, no squares present, nothing in the source. 4. Browse same URL with other browser and the squares are being outputed. What is the expected behavior? We should at least see a square [] for those unicodes. What went wrong? Seems chrome is now handling correctly certain unicodes and not even inserting them in the source code. Does it occur on multiple sites: Yes Is it a problem with a plugin? No Did this work before? Yes The previous version 48 of Chrome. Does this work in other browsers? Yes Chrome version: 49.0.2623.87 Channel: stable OS Version: 6.1 (Windows 7, Windows Server 2008 R2) Flash Version: Shockwave Flash 21.0 r0 Here is also another list of unicodes that are in problem: 󰀁 󰀂 󰀃 󰀄 󰀅 󰀆 󰀇 󰀈 󰀉 󰀊 󰀋 󰀌 󰀍 󰀎 󰀏 󰀐 󰀑 󰀒 󰀓 󰀔 󰀕 󰀖 󰀗 󰀘 󰀙 󰀚 󰀛 󰀜 󰀝 󰀞 󰀟 󰀠 󰂀 󰂁 󰂂 󰂃 󰂄 󰂅 󰂆 󰂇 󰂈 󰂉 󰂚 󰂛 󰂜 󰂝 󰂞 We also use these unicodes for a custom WOFF file for glyphs.
,
Mar 24 2016
,
Mar 25 2016
,
Mar 25 2016
All the missing glyphs are in SUPPLEMENTARY_PRIVATE_USE_AREA_A. It's not clear what the desired behavior is in this case.
,
Mar 27 2016
> 3. In Chrome, FE0000 to FE000F are not showing up, no squares present, nothing in the source. Did you mean FE00 to FE0F? They are variation selectors. http://www.fileformat.info/info/unicode/char/fe00/index.htm and the Unicode Standard 8.0 23.4 Variation Selectors defines: > The variation selectors themselves are combining marks of combining class 0 and are > default ignorable. Thus, if the variation sequence is not supported, the variation selector > should be invisible and ignored. http://www.unicode.org/versions/Unicode8.0.0/ch23.pdf so the Chrome behavior looks correct to me. For the list of Unicode code points, the URL does not contain these code points, correct? Do you have a reproducing page with such a font as a web font?
,
Mar 28 2016
Able to reproduce the issue on windows 7, Linux Ubuntu 14.04 and Mac 10.11.3 using chrome version 49.0.2623.108 and canary 51.0.2692.0. This issue is observed from M30 builds. pierluc.jussaume @ Could you please respond to comment #6. Removing the bisect label,please feel free to add if required. Thanks,
,
Mar 28 2016
Comment 6 @ Yes sorry, I meant FE00 to FE0F. Last table in the provided link. With Chrome you can see that these cells are not being outputted as any character, nothing in the display and nothing in the source. Although, browsing the same with a different browser, you can see at least a [], in the display and source as well. These were maybe not the best example, and they are maybe unsupported. But still being managed by other browsers. - About the other list of unicodes, here is the top of the list: http://www.fileformat.info/info/unicode/char/f0001/index.htm You can see this test page for that specific code: http://www.fileformat.info/info/unicode/char/f0001/browsertest.htm Again, browsing the test page using different browsers you will see the same behavior, Chrome not outputting anything for this code. We have been using these codes mapped to glyphs through a font for our icons for a while now, and since update 49 this issue is causing missing some in our website. Thanks
,
Apr 5 2016
If you have a web font that have glyphs at U+F0001 and it's not rendered, I agree it's a bug. Can you post a reproducing HTML/font? It's a separate issue whether to render a box or not when a glyph is missing in the fonts.
,
Apr 5 2016
Here is a link listing the current unicodes we're using with our font. https://www.jeancoutu.com/Templates/GJC/Styles/Fonts/Pictos/index.html When I said a box it was more the rendered unicode itself, as visible in: http://www.fileformat.info/info/unicode/char/f0001/browsertest.htm
,
Jun 2 2016
Do we have any news around this defect? Or a debatable reason why the implementation changed, causing to loose our glyphs out of sudden? Thanks
,
Jun 6 2016
This is not reproducible for me in Chrome Linux 51.0.2704.79. This issue was fixed in https://chromium.googlesource.com/chromium/src.git/+/3f51d6256864b298b15feb93a57c6d986474245a - we had an integer truncation issue in some character test functions, which lead to characters incorrectly getting mapped to whitespace. Thank you for your report and your patience.
,
Jun 6 2016
Btw, I confirmed that the mentioned change is in fact the fix, and that the issue is reproducible when this change is reverted.
,
Jun 8 2016
Thank you for the responses, I can confirm that with the latest update "51.0.2704.84 m" that this is no longer an issue. |
||||||
►
Sign in to add a comment |
||||||
Comment 1 by tkent@chromium.org
, Mar 24 2016