[Patch] timezone/zic.c: Fix file handle leak in dolink()
Hi all,
I noticed a small issue and made a tiny patch to fix it — it addresses a file handle leak in the dolink() function of src/timezone/zic.c.
--
Regards,
Man Zeng
www.openhalo.org
Attachments:
zic_file_leak_fix.diffapplication/octet-stream; charset=gb18030; name=zic_file_leak_fix.diffDownload+1-0
Furthermore, I noticed that some of the test code in `contrib/pg_trgm/trgm_regexp.c` does not check the return value of the `fopen()` call.
I'm not sure if it's necessary to make the modification.
--
Regards,
Man Zeng
www.openhalo.org
Attachments:
pg_trgm-add-fopen-null-check.diffapplication/octet-stream; charset=ISO-8859-1; name=pg_trgm-add-fopen-null-check.diffDownload+6-0
"=?gb18030?B?emVuZ21hbg==?=" <zengman@halodbtech.com> writes:
I noticed a small issue and made a tiny patch to fix it ¡ª it addresses a file handle leak in the dolink() function of src/timezone/zic.c.
Huh? The program is going to exit() two lines further down.
What's the point of closing here?
Even if there were a plausible argument for issuing close(),
I wouldn't care to diverge from upstream zic. If you like
you can go try to persuade the tzdb crew that this is a
useful change.
regards, tom lane
"=?ISO-8859-1?B?emVuZ21hbg==?=" <zengman@halodbtech.com> writes:
Furthermore, I noticed that some of the test code in `contrib/pg_trgm/trgm_regexp.c` does not check the return value of the `fopen()` call.
I'm not sure if it's necessary to make the modification.
Probably not worth the trouble. This is barely even debug-grade
code, seeing that there's no provision to change the target file
name. It's clearly not meant for use in unfriendly environments.
regards, tom lane