136743 - xlib complains a lot about fonts with '*' in the XLFD string

Status: RESOLVED FIXED

(Core Graveyard :: GFX, enhancement)

Description

[timeless]

xlib to qnx xphoton

WEBSHELL+ = 3
###!!! ASSERTION: atoi(resolutionX) failure.: '!(resolutionX[0] != '0' && resX == 0)', file /home/timeless/mozilla/gfx/src/xlib/nsFontMetricsXlib.cpp, line 4109
###!!! Break: at file /home/timeless/mozilla/gfx/src/xlib/nsFontMetricsXlib.cpp, line 4109
Suspending process; attach with the debugger.

Program received signal SIGSTOP, Stopped (signal).
0x284527d0 in kill () from /usr/lib/libc_r.so.4
(gdb)
Continuing.
###!!! ASSERTION: atoi(resolutionY) failure.: '!(resolutionY[0] != '0' && resY == 0)', file /home/timeless/mozilla/gfx/src/xlib/nsFontMetricsXlib.cpp, line 4117
###!!! Break: at file /home/timeless/mozilla/gfx/src/xlib/nsFontMetricsXlib.cpp, line 4117
Suspending process; attach with the debugger.

Comment 1

[timeless]

Attachment: xlsinfo

Comment 2

[timeless]

Ok, this appears to be a bug in mozilla xlib

#3  0x28d92501 in GetFontNames (aPattern=0xbfbfd11c "-*-helvetica-*-*-*-*-*-*-*-*-*-*-*-*", aAnyFoundry=1, aNodes=0x8675440)
    at /home/timeless/mozilla/gfx/src/xlib/nsFontMetricsXlib.cpp:4117
(gdb) p list[i]
$3 = 0x86a3001 "-*-helvetica-medium-i-normal-*-12-*-*-*-*-*-iso8859-1"

patch coming

Comment 3

[timeless]

Attachment: check for * instead of just 0

this fixes the assertion, i'm guessing it's the right fix too :)
i'm attaching the patch using an xlib mozilla running with the patch :)

Comment 4

[Roland Mainz]

Looking at attachment 78624 [details]:
I never saw things like this in a "xlsfonts" output:
-- snip --
-*-helvetica-medium-i-normal-*-12-*-*-*-*-*-iso8859-1
-*-helvetica-medium-i-normal-*-7-*-*-*-*-*-iso8859-1
-*-lucidasans-medium-r-normal-*-10-*-*-*-*-*-iso8859-1
-*-lucidasans-medium-r-normal-*-12-*-*-*-*-*-iso8859-1
-*-lucidasans-medium-r-normal-*-14-*-*-*-*-*-iso8859-1
-*-lucidasans-medium-r-normal-*-18-*-*-*-*-*-iso8859-1
-*-lucidasans-medium-r-normal-*-24-*-*-*-*-*-iso8859-1
-*-lucidasans-medium-r-normal-*-8-*-*-*-*-*-iso8859-1
-- snip --

bstell:
Are this legal XLFDs ?

Comment 5

[kill this account]

The wildcard character '*' is allowed in the pattern used to query for fonts
but not in the font name. See:

http://www.x-docs.org/XLFD/xlfd.pdf

    3.1.1. FontName Syntax
    The FontName is a structured, parsable string ...

      Field values are constructed as strings of ISO 8859-1 graphic 
      characters, excluding the following:
        • ‘‘-’’ (HYPHEN), the XLFD font name delimiter character
=>      • ‘‘?’’ (QUESTION MARK) and ‘‘*’’ (ASTERISK), the X protocol font name 
                wildcard characters
        • ‘‘,’’ (COMMA), used by Xlib to separate XLFD font names in a font set.
        • ‘‘"’’ (QUOTATION MARK), used by some commercial products to quote a 
                font name.

Comment 6

[kill this account]

Since there is no way for the font code to specify these fonts I suggest
we consider them invalid.

Comment 7

[Roland Mainz]

Brian Stell wrote:
> Since there is no way for the font code to specify these fonts I suggest
> we consider them invalid.

OK, I'll file a new patch to reject these fonts...

Comment 8

[Roland Mainz]

Attachment: Patch for 2002-04-09-08-trunk

The new patch rejects the font if certain values contain '*' and adjust the
assertion to stop complaining (since the font is rejected anyway...) ...

Comment 9

[Roland Mainz]

Requesting r=/sr=/a= ...

Comment 10

[Roland Mainz]

Attachment: Patch for 2002-04-09-08-trunk (2nd attempt)

Same patch as previous one, but I fixed a minor typo - sorry...

Comment 11

[tenthumbs]

Don't you also want to reject '?' as well? In fact, wouldn't it be 
better to test and reject for any non-digit character?

In nsFontMetricsXlib.cpp, it might be a good idea to replace the atoi
calls with strtol so you can use strtol's extra functionality. On Linux,
atoi is just a wrapper for strtol anyway. I wouldn't be surprised if
that's also true for other platforms.

Comment 12

[kill this account]

This problem could also be solved by checking the whole string for '?' and '*' 
using strchr/strpbrk.

Comment 13

[Roland Mainz]

Attachment: New patch for 2002-04-14-08-trunk which uses strpbrk()

Comment 14

[Roland Mainz]

Requesting r=/sr= ...

Comment 15

[kill this account]

The code in the patch looks good but the comment need correction:

> +    /* Check if we can handle the font name ('*' and '?' are valid in XLFD

Both '*' and '?' are valid in the _input_ pattern to XListFonts (XLFD) but are 
_invalid_ in font names (ie: the results of XListFonts). See the comment #5
http://bugzilla.mozilla.org/show_bug.cgi?id=136743#c5?id=136743

Comment 16

[Roland Mainz]

Attachment: New patch for 2002-04-21-08-trunk per bstell's comment

Comment 17

[Roland Mainz]

Requesting r=/sr= ...

Comment 18

[kill this account]

r=[e-mail]

Comment 19

[Marc Attinasi]

Comment on attachment 80429 [details] [diff] [review]
New patch for 2002-04-21-08-trunk per bstell's comment

sr=attinasi

Comment 20

[Roland Mainz]

Patch checked in
(http://bonsai.mozilla.org/cvsquery.cgi?module=MozillaTinderboxAll&branch=HEAD&cvsroot=/cvsroot&date=explicit&mindate=1020129780&maxdate=1020130320&who=bzbarsky%25mit.edu),
marking bug as FIXED.