206060 - ASSERTION in x11shared/nsFT2FontNode.cpp w/ big5 page but the problem is really in nsFontMetricsGTK.cpp

Status: RESOLVED WONTFIX

(Core Graveyard :: GFX: Gtk, defect)

Description

[tenthumbs]

I was feeding mozilla a bunch of web pages from the squid distribution (error
templates in different languages) and hit an assertion with a big5 page. A
little work with NS_FONT_DEBUG and you get:

      hpbig5- ffre = *-*-hpbig5-, nsFontMetricsGTK.cpp 6223
        TryNodes aFFREName = *-*-hpbig5-, nsFontMetricsGTK.cpp 5636
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-hpbig5-
looking for FreeType font matching -*-*-*-*-*-*-*-*-*-*-*-*-hpbig5-,
nsFT2FontNode.cpp 122
###!!! ASSERTION: garbled pattern: unexpected end of pattern: '*p', file
nsFT2FontNode.cpp, line 432
Break: at file nsFT2FontNode.cpp, line 432

and the xlfd name is invalid. The actual problem is in nsFontMetricsGTK.cpp:

     491   { "hp-tchinese15",      &FLG_NONE,    &Unknown       },
     492   { "hp-tchinesebig5",    &FLG_ZHTW,    &Big5          },
     493   { "hp-wa",              &FLG_NONE,    &Unknown       },
     494   { "hpbig5-",            &FLG_ZHTW,    &Big5          },
     495   { "hproc16-",           &FLG_NONE,    &Unknown       },

The strings in lines 494 and 495 don't seem to have the correct form.

This also affects some other X fotn metric files.

Comment 1

[timeless]

Attachment: guess

Comment 2

[tenthumbs]

I thought of something like this but it could just as easily be 
"hp-big5" which is similiar to the preceding entries. Presumably this is 
a AIX thing so we need an expert.

Comment 3

[Jim Dunn]

Well I think it is probably an hp-ux thing rather than an aix thing... 

Not sure if the the suggested patch is correct.  I searched for
hpbig5 on the web and came across this reference:
http://h21007.www2.hp.com/dspp/tech/tech_TechDocumentDetailPage_IDX/1,1701,3974,00.html

which shows and entry 
    -arph-armingtil-medium-r-normal--0-0-0-0-m-0-hpbig5-
note it ends in '-'.  I am not enough of a font expert to even
guess at this thought.

Comment 4

[Roland Mainz]

Jim Dunn wrote:
[snip]
> which shows and entry 
>     -arph-armingtil-medium-r-normal--0-0-0-0-m-0-hpbig5-
> note it ends in '-'.  I am not enough of a font expert to even
> guess at this thought.

Is it possible to attach such a "*-hpbig5-" font here (*.pck/*.bdf and the
matching fonts.dir, please) that we can have a look at it, please ?

Comment 5

[Jungshik Shin]

I guess timeless' patch will work, but I'm not sure whether Xlib will match
'hpbig5-' with 'hpbig5-*' (i.e. whether it'll return with 'hpbig5-' fonts when
asked for 'hpbig5-*'). Our FT2 code has no problem with that.

Kishan or Mi-Young, can you try attachment 123639 [details] [diff] [review] on HP/UX with 'hpbig5-' fonts?
 I wish HP/UX had something more reasonable Like 'hpbig5-0' ;-)

Comment 6

[Jungshik Shin]

> whether it'll return with 'hpbig5-' fonts when asked for 'hpbig5-*'

  It seems like it doesn't return 'hpbig5-' when we ask for '*-....*-hpbig5-*'
according to bug 214769 comment #10 (and bug 214769 comment #11). The test was
done with 'hpbig5-*' (note '*') and 'hphkbig5-' (no '*' here). The fact that
'hphkbig5-' fonts were listed for zh-TW (bug 152264) while 'hpbig5-' fonts were
not listed strongsly suggests that at least on HP/UX, XListFonts (sp?) doesn't
return 'hpbig5-' when asked for 'hpbig5-*'. I'll write a simple X11 program and
test it on Linux.  

 It's likely that we need a different fix than attachment 123639 [details] [diff] [review].

Comment 7

[Jim Dunn]

I am probably not the right person to review this patch.
Since I don't have an real review authority... it would
probably be better to get a review Jungshik or Kishan
to check whether it is a workable fix.

Comment 8

[tenthumbs]

I'm now seeing this 7 times on <http://www.w3.org/Style/CSS/>. The bogus
"hpbig-", et. al.,  really needs to be dealt with.

Comment 9

[Mats Palmgren (inactive)]

mozilla/gfx/src/xlib and mozilla/gfx/src/gtk have been removed on trunk.
This bug doesn't look like a branch candidate to me.

-> WONTFIX