131831 - Printing does not work in de_AT@euro locale

Status: VERIFIED FIXED

(Core Graveyard :: Printing: Xprint, defect, P3)

Description

[Christian :Biesinger (don't email me, ping me on IRC)]

XPrint printing seems to only work with locales using . as decimal seperator,
e.g. en_US or POSIX or C, but not de and derivatives.

Comment 1

[Roland Mainz]

Ohhh, fun.

I assume this is caused by
http://lxr.mozilla.org/seamonkey/source/gfx/src/xprint/xprintutil.c#669 ...

From the scanf(3S) manual page:
-- snip --
The scanf() function in all its forms allows  for  detection
of a language-dependent radix character in the input string.
The radix character  is  defined  in  the  program's  locale
(category  LC_NUMERIC).  In the POSIX locale, or in a locale
where the radix character is not defined, the radix  charac-
ter defaults to a period (.).
-- snip --

Does anyone have a suggestion how to force sscanf() to use the C/POSIX locate
when parsing the data from Xprt ?

Comment 2

[Roland Mainz]

Swapping QA <--> Owner ...

Comment 3

[Roland Mainz]

CC:'ing [e-mail] ...

Comment 4

[Roland Mainz]

Attachment: Patch for 2002-03-15-08-trunk

I'll take the setlocale()-way - PostScipt module does the same... :)

Comment 5

[John Keiser (jkeiser)]

Comment on attachment 74826 [details] [diff] [review]
Patch for 2002-03-15-08-trunk

- spacing around the sscanf() should be 2-space
- open another bug that our setlocale() code is not threadsafe.  It *is* a
problem, and we shouldn't lose it, but the problem is really the sscanf() API
and we're doing this elsewhere anyway.	No need to hold this up for that.
r=jkeiser with that

Comment 6

[Roland Mainz]

Attachment: New patch for 2002-03-15-08-trunk per jkeiser's comments

Comment 7

[John Keiser (jkeiser)]

Comment on attachment 74854 [details] [diff] [review]
New patch for 2002-03-15-08-trunk per jkeiser's comments

r=jkeiser

Comment 8

[Roland Mainz]

Filed bug 131867 ("Our use of setlocale() is not threadsafe") ...

Comment 9

[Marc Attinasi]

Comment on attachment 74854 [details] [diff] [review]
New patch for 2002-03-15-08-trunk per jkeiser's comments

looks fine to me- sr=attinasi

Comment 10

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 74854 [details] [diff] [review]
New patch for 2002-03-15-08-trunk per jkeiser's comments

In my professional opinion *scanf is a pit of evil, but a=roc+moz

Comment 11

[Christian :Biesinger (don't email me, ping me on IRC)]

Roland, are you sure that this line returns the old locale and not the one just set?
+  cur_locale = setlocale(LC_NUMERIC, "C"); 
From the manpage:
RETURN VALUE
       A  successful call to setlocale() returns an opaque string
       that corresponds to the locale set. 

Comment 12

[Christian :Biesinger (don't email me, ping me on IRC)]

Comment on attachment 74854 [details] [diff] [review]
New patch for 2002-03-15-08-trunk per jkeiser's comments

marking as needs work.
I tested, and setlocale(LC_NUMERIC, "C"); returns "C".
You should change it to:
cur_locale = setlocale(LC_NUMERIC, NULL);
setlocale(LC_NUMERIC, "C");

Comment 13

[Roland Mainz]

xx@@@!!!... ;-(

I'll file a new patch...

Comment 14

[Christian :Biesinger (don't email me, ping me on IRC)]

I just had a look at NSPR. It does not use locale information, so using it would
avoid the setlocale calls; this is probably a better solution than in this patch.

Comment 15

[Roland Mainz]

Attachment: New patch for 2002-03-17-08-trunk per biesi's comment

Comment 16

[Roland Mainz]

Christian Biesinger wrote:
> I just had a look at NSPR. It does not use locale information, so using it 
> would avoid the setlocale calls; this is probably a better solution than in 
> this patch

The problem is that the XprintUtil code is used outside the Mozilla tree, too -
where NSPR is not available.

One item on my ToDo list is to rewrite the parser in the near future (and I
added the item "avoid sscanf()" to that list... :) - the current one has become
a hacked hack with hacked fixes (but it works... :)
But I need a fix for the 1.0 release - which means I need a quick solution for
this issue...

----

biesi, can you r= , please ?

Comment 17

[Christian :Biesinger (don't email me, ping me on IRC)]

Comment on attachment 74960 [details] [diff] [review]
New patch for 2002-03-17-08-trunk per biesi's comment

r=biesi

Comment 18

[Scott Collins]

Comment on attachment 74960 [details] [diff] [review]
New patch for 2002-03-17-08-trunk per biesi's comment

sr=scc, of course, with all the cautions issued above

Comment 19

[Boris Zbarsky [:bzbarsky]]

patch checked in on trunk

Comment 20

[Roland Mainz]

Marking bug as FIXED.

Biesi:
Can you verify, please ?

Comment 21

[Christian :Biesinger (don't email me, ping me on IRC)]

great, works now
marking verified.