Bug#322713: (fwd) Bug#322713: dvipdfm doesn't understand gs anymore
On Sep 5, 2005, at 4:56 PM, Hilmar Preusse wrote:
Great! Just a short question: Do you know about
http://bugs.debian.org/291438 and is it eventually fixed in DVIPDFMx?
I've just read the bug report. The first report from Prabhu Ramachandran
(on Jan 22 15:31:54 +0530) was related to cmp_dest() function. But the
second report (on Jan 24 16:12:27 +0530) was realted to asn_date()
funtion.
I have no idea on the first report because the report could not give me
much information.
However, for the second report, I may guess what happened.
The problem was that his local timezone is special (+0530). Usually the
time difference happens only in the hour part, e.g., -09:00, -06:00,
etc.
But his one was +05:30. If so, the current source code of DVIPDFM might
cause the segmentation fault, because of memory allocation.
He solved the problem by increasing the size of date_string to 32 (from
24). But it is not the exact solution. PDF Reference requires the format
of date_string as D:YYYYMMDDHHmmSSOHH'mm' (23-characters). The problem
was in the last mm'. PDF Reference says as follows:
mm followed by ' is the absolute value of the offset from UT in
minutes (00-59)
Therefore the current source of DVIPDFM must be fixed as follows:
(original; line 250 in pdfdoc.c)
sprintf (date_string, "D:%04d%02d%02d%02d%02d%02d%+03ld'%02ld'",
bd_time -> tm_year+1900, bd_time -> tm_mon+1, bd_time ->
tm_mday,
bd_time -> tm_hour, bd_time -> tm_min, bd_time -> tm_sec,
-timezone/3600, timezone%3600);
(corrected; line 250 in pdfdoc.c)
sprintf (date_string, "D:%04d%02d%02d%02d%02d%02d%+03ld'%02ld'",
bd_time -> tm_year+1900, bd_time -> tm_mon+1, bd_time ->
tm_mday,
bd_time -> tm_hour, bd_time -> tm_min, bd_time -> tm_sec,
-timezone/3600, (timezone%3600)/60);
Then, why not happened the problem to others? My guess is that he
(Prabhu)
might fix the source code (line 237 in pdfdoc.c) to correct one, and
might use the line
#define timezone (bd_time->tm_gmtoff)
If anyone does not fix the bug in line 237, there are two cases. One is
using the line
#define timezone 0l
so that no problem. The other case is getting the compilation error.
Actually DVIPDFMx had the same problem, and I fixed it today.
Thanks for letting me know the bug.
Best, ChoF.
~~~~~~~~~~~~~~~~~~~~~~~~~ ***
| Cho, Jin-Hwan == ChoF | ^ ^
~~~~~~~~~~~~~~~~~~~~~~~~~~~ o
| Dept. of Mathematics | ---
| The University of Suwon | |
~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| Korean TeX Users Group | |
| http://www.ktug.or.kr | |
~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| chofchof@ktug.or.kr | ***
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Reply to: