Open Bug 181722 Opened 22 years ago Updated 2 years ago

minimum font size pref only works for some encodings by default because there's a separate pref for each language group

Categories

(Core :: CSS Parsing and Computation, defect, P4)

defect

Tracking

()

People

(Reporter: tommi.komulainen, Unassigned)

References

()

Details

User-Agent: Mozilla/5.0 Galeon/1.3 (Turing 42;) Gecko Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3a) Gecko/20021003 Setting the minimum font size has no effect on the page in question. It remains as unreadable even if minimum font size is set to 24. Reproducible: Always Steps to Reproduce: 1. Load the mentioned URL 2. See how the text on the right hand side is annoyingly small 3. Set minimum font size to 24 Actual Results: No change in font size. Expected Results: Increased the font size to the defined minimum. GTK2 build with Xft enabled, running under LC_CTYPE=fi_FI@euro locale.
The minimum font size pref is per-encoding, and it works if you set it in the "Central European" encoding. (See the dropdown at the top of the fonts pref panel.) I think that's bad design, though (only partly because encoding is an incorrect way to determine content language). This is either style system or internationalization (or GFX).
Assignee: font → dbaron
Status: UNCONFIRMED → NEW
Component: Layout: Fonts and Text → Style System
Ever confirmed: true
Summary: Setting minimum font size has no effect → minimum font size pref only works for some encodings by default
On a clean install- western has a default size set, Hebrew does not. Per ecoding pref does make sense (the relative font sizes in diffrent ecodings is diffrent- for example, a 12pt Hebrew font is close in size on screen to a 11pt english font). However, we should either have no min size for any encoding set by default, or have all set by default (which is better imo). The way we have it now, with some off and some on, is terribly comfusing
The current setup is intentional because the so-called "glyph black-box" (the area attained by the 'ink') is generally widely different between certain languages/fonts in unpredictable ways. C.f. also bug 158868 comment 9. (BTW, there is a similar issue with |font-size-adjust| which is even more harsh since it brings the distinction down to individual fonts.) Unfortunately, the current behavior can lead to cases where the user is totally mystified about what is happening, such as bug 158256 comment 3. It is conceivable that the FE could group things such as "Western" and "Central European" linked together (in the background), so that the FE only presents "Western/Central European/etc", and when the user sets the minimum size, the JS propagates it to the corresponding linked list. However, a side-effect would be incompatibilities in the UI. "Western" and "Central European" would be linked here, but separated in another context, and this might cause confusions. Cc:ing cbiesinger@web.de who fixed bug 110342 in case he has any further thoughts... Re: comment 2 > On a clean install- western has a default size set, Hebrew does not. No idea why some have and some do not. Need to track people who set the defaults for the rationale behind this.
I confirm it does not work in Traditional Chinese either. I am sure that I have set the minimum font size to 14. However, it does not work on some pages such as: http://netcity5.web.hinet.net/UserData/hoow/html/topics.htm and http://98.to/ocphk/ Please try them, developers. In fact, changing enocding does not help!
I tried the first URL in comment 4, and then chose: Edit -> Preferences -> Appearance -> Fonts -> Traditional Chinese (Taiwan) and set the "Minimum font size" pref to 18. That increased the font size on the page. What am I missing? (Autodetection decided the page was in Big5 encoding.)
Sorry, I haven't give out the details enough. I have tried the links on Mozilla 1.5, Mozilla FireBird 0.7 and Camino 0.7 Nightly Build 2003100911 before posting comment 4. All ran on Mac OS X 10.2.8. They were not working at the time I wrote the Comment 4. I had been suffering this problem for about 2 weeks with Camino as my primary browser. However, After I have download an Unofficial Camino Build 2003101804 from Dave Haas ( http://www.cae.wisc.edu/~haasd/ ). The Mozilla 1.5 and Mozilla Firebird 0.7 work now. The Camino 0.7 Nightly Build 2003101612 and the Unofficial Build 2003101804 work correctly on the second link ( http://98.to/ocphk/ ) but the Minimum Font size does not work on the first link ( http://netcity5.web.hinet.net/UserData/hoow/html/topics.htm ). Thanks for your reply.
I finally found the culprit! It works if I set the encoding to "Traditional Chinese (Big5)" but it does not work with "Tradition Chinese (Big5 HKSCS)". I think developers missed to put "Big5 HKSCS" under the "Minimum Font Size" control of Traditional Chinese.
In Mozilla there are separate font prefs for "Traditional Chinese (Taiwan)" and "Traditional Chinese (Hong Kong)". If Camino doesn't have those, then that's a camino bug.
Thank for David clarifying the problem. However, I have tried on Mozilla 1.5 and Mozilla Firebird 0.7 on Mac OS X. In both Mozilla->Preferences->Appearance->Fonts and Mozilla Firebird->Preference->Web Features->Fonts & Colors, there is just a single entry for "Traditional Chinese". I cannot find the "Tradtional Chinese (Taiwan)" and "Traditional Chinese (Hong Kong)" setting. Nonetheless, I discovered the Minimum Font Size setting works with both "Traditional Chinese (Big5)" and "Traditional Chinese (Big5 HKSCS)" under both Mozilla and Mozilla Firebird. Therefore I have setted a separate bug report for the Camino alone, bug 222919.
It's possible that that split was after 1.5 was released.
Yeah, my request for a1.5 was turned down (see bug 152264).
Thanks for Jungshik Shin for more information. I have added additonal suggestion for bug 15226. :)
I think the problem is the split personality on the Preference setting. It is very confusing that the Fonts set for Traditional Chinese are used for Big5-HKSCS encoded page but the Minimum Font Size isn't. Please make it consistent that setting of Minimum Fonts Size applies to all the Traditional enocded pages (both Big5 and Big5-HKSCS).
It's about trunk build (not 1.5) on Mac OS X, right? Then, let's deal with it in bug 222919. I know why that's happening.
Sorry about that, Jungshik. I was mixed up a bit.
I think this entier thing is bogus... What I mean to say is: end-users don't give a damn about the fact that different font sizes and encodings require different minimum sizes. We humans don't work in font sizes in "points". We work in "visual size". I think we should find a way to allow the user to set the minimum visual size (not in point units) that a font can have, regardless of the underlying encoding and font and allow Thunderbird to map that back to point units. If you think I missed something, feel free to let me know. In Java, you have something called FontMetrics which lets you find out the width/height of any text in a certain font. What I am simply proposing is that you try rendering the page in whatever the normal size is, then use something equivilent to FontMetrics to calculate the minimum font size across the entire letter, then increase the font across the entire letter such that the mininum matches that specified by the user. I think the minimum size should be "minimum font height in pixels".
Assignee: dbaron → nobody
QA Contact: ian → style-system
Since this can be so baffling to the user, couldn't a change be made to group non-CJK and CJK fonts under the covers, so that: 1-any change to the minimum for any non-CJK changes it for all non-CJK encodings 2-one of the following for CJK encodings: a-handle all CJK encodings as now, or b-any change to the minimum for any CJK changes it for all CJK encodings
OS: Linux → All
Priority: -- → P4
Hardware: x86 → All
Summary: minimum font size pref only works for some encodings by default → minimum font size pref only works for some encodings by default because there's a separate pref for each language group
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.