374565 - weird characters showing at https://bugzilla.mozilla.org/

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Description

[-fullmetaljacket-]

Attachment: screenshot

"in-testsuite" flags on bugzilla.mozilla.org turns into some weird characters.

same weird character for "blocking-firefox3" and "blocking-thunderbird2"  flags.

Comment 1

[-fullmetaljacket-]

btw, i am using Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a3pre) Gecko/20070319 Minefield/3.0a3pre ID:2007031918 [cairo]

Comment 2

[Brian Polidoro]

Is this the same as Bug 365928?  Do those same type of characters so up in the testcase for that bug?  I see a different font but they are still the correct letters.

I don't see those characters for the in-testsuite flag.  But it is a different font that the rest of the flags.  So maybe it's falling back to a different font for each of us.

Which that bug leads to the can of worms that is bug 332649.

Comment 3

[-fullmetaljacket-]

(In reply to comment #2)
>Do those same type of characters so up in the testcase for that bug?

nope, testcase render perfectly fine (except that strange behavior when selecting the second line)

 
> I don't see those characters for the in-testsuite flag.  But it is a different
> font that the rest of the flags.  So maybe it's falling back to a different
> font for each of us.
> 
> Which that bug leads to the can of worms that is bug 332649.

i think so.

Comment 5

[Jean-Marc Desperrier]

Sorry, but I'm not convinced this is really a duplicate.

In my analysis :
- Firefox selects a font adequate to display the "dash" character inside the word
- It then uses that font to display the rest of the word (related to bug 332649, but in my opinion that's more or less correct behavior, font change in the middle of a word is ugly)
- But the font is completely inadequate to display ordinary characters

Two possibilities :
- the font that's gets selected is a "symbol font" (but not flagged as such) that has inadequate glyphs for ANSI codepoint. Then it's really a dup of 332649 and that's an important point to take into account for bug 332649, that is, to try to find a way to avoid using such fonts when switching fonts.
- We have a bug where the selected font correctly reports it can't display those characters, but that's not taken into account by firefox.

I suspect it might be the second case, because the second character displayed is an empty character box, so the font really seems not to have a character for it.

But that's no very easy to check.

-fullmetaljacket-, could you copy the "in-testsuite" text from bugzilla.mozilla.org to wordpad, select it, and try every font in your system on it until you find one that gives you the same display as in the bug ?

And attach to the bug the content of your registry key "HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Fonts" ? (another option : "cd WINDIR\font" "dir > fontlist.txt")

Comment 6

[-fullmetaljacket-]

(In reply to comment #5)
thank you for your thoughts. 

i am currently out of town this week and i will do what you said as soon as i am back to my computer that exhibit this bug.

Comment 7

[-fullmetaljacket-]

im back with my pc. but, wow! this bug nows wfm with the latest hourly Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a4pre) Gecko/20070403 Minefield/3.0a4pre ID:2007040223 [cairo]

but still broken with nightly build. so seems fixed by checkin for bug 376299 (?). 

i had tried to copy the "in-testsuite" and tried to test what font is showing but haven't pinpointed any. nothing render similar to what is showed on the screenshot i attached on this bug.

Comment 8

[Stuart Parmenter]

More likely this would have been fixed by 376300.  Can others verify?  If so I'd reopen and mark this fixed since it wouldn't be a dup.

Comment 9

[-fullmetaljacket-]

yes, it is fixed indeed by bug 376300

BROKEN -> 20070402 22:24pst
WFM    -> 20070402 23:51pst

CHECK-IN -> http://bonsai.mozilla.org/cvsquery.cgi?module=PhoenixTinderbox&date=explicit&mindate=1175577840&maxdate=1175583059

Comment 10

[-fullmetaljacket-]

fixed by checkin for bug 376300