# [PATCH] Fix minor memory leak in init-db 5 messages from 2009-05-14 to 2009-05-17. Participants: Ammon Riley, Junio C Hamano. Thread: https://gitlist.dev/t/19354 ## Ammon Riley, 2009-05-14 22:22 Subject: [PATCH] Fix minor memory leak in init-db Message-ID: URL: https://gitlist.dev/e/cd3664ac0905141522s3ad78411o9f015d2e7720c8b5%40mail.gmail.com ``` There was an xmalloc() for path, but I didn't see a corresponding free(). Does it happen somewhere else that I'm not expecting? Signed-off-by: Ammon Riley --- builtin-init-db.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/builtin-init-db.c b/builtin-init-db.c index d1fa12a..6969987 100644 --- a/builtin-init-db.c +++ b/builtin-init-db.c @@ -308,6 +308,8 @@ int init_db(const char *template_dir, unsigned int flags) strcpy(path+len, "/info"); safe_create_dir(path, 1); + free(path); + if (shared_repository) { char buf[10]; /* We do not spell "group" and such, so that -- 1.6.3.1.9.g95405b.dirty ``` ## Junio C Hamano, 2009-05-16 19:56 Subject: Re: [PATCH] Fix minor memory leak in init-db Message-ID: <7vzldclt33.fsf@alter.siamese.dyndns.org> URL: https://gitlist.dev/e/7vzldclt33.fsf%40alter.siamese.dyndns.org In-Reply-To: ``` Ammon Riley writes: > There was an xmalloc() for path, but I didn't see a corresponding free(). > Does it happen somewhere else that I'm not expecting? It implicitly happens in exit() in git.c:handle_internal_command() after cmd_init_db() returns the control to it. ``` ## Ammon Riley, 2009-05-17 04:07 Subject: Re: [PATCH] Fix minor memory leak in init-db Message-ID: URL: https://gitlist.dev/e/cd3664ac0905162107t7ebca943p1722685a8484f08e%40mail.gmail.com In-Reply-To: <7vzldclt33.fsf@alter.siamese.dyndns.org> ``` On Sat, May 16, 2009 at 12:56 PM, Junio C Hamano wrote: > Ammon Riley writes: > >> There was an xmalloc() for path, but I didn't see a corresponding free(). >> Does it happen somewhere else that I'm not expecting? > > It implicitly happens in exit() in git.c:handle_internal_command() > after cmd_init_db() returns the control to it. Ah. Naturally. :) So if I were to write a long-lived application (such as a custom UI) that links to libgit, and bypasses those functions to call init_db() (and other functions) directly, all those implicit free-on-exit() turn into memory leaks. Cheers, Ammon ``` ## Junio C Hamano, 2009-05-17 05:13 Subject: Re: [PATCH] Fix minor memory leak in init-db Message-ID: <7veiuojoqy.fsf@alter.siamese.dyndns.org> URL: https://gitlist.dev/e/7veiuojoqy.fsf%40alter.siamese.dyndns.org In-Reply-To: ``` Ammon Riley writes: > On Sat, May 16, 2009 at 12:56 PM, Junio C Hamano wrote: >> Ammon Riley writes: >> >>> There was an xmalloc() for path, but I didn't see a corresponding free(). >>> Does it happen somewhere else that I'm not expecting? >> >> It implicitly happens in exit() in git.c:handle_internal_command() >> after cmd_init_db() returns the control to it. > > Ah. Naturally. :) > > So if I were to write a long-lived application (such as a custom UI) that > links to libgit, and bypasses those functions to call init_db() (and other > functions) directly, all those implicit free-on-exit() turn into memory > leaks. Correct, and there are other much larger issues to worry about. That's why there is a separate libgit2 effort in progress. ``` ## Ammon Riley, 2009-05-17 16:09 Subject: Re: [PATCH] Fix minor memory leak in init-db Message-ID: URL: https://gitlist.dev/e/cd3664ac0905170909l3184c286s7f7d61e108fd5b05%40mail.gmail.com In-Reply-To: <7veiuojoqy.fsf@alter.siamese.dyndns.org> ``` On Sat, May 16, 2009 at 10:13 PM, Junio C Hamano wrote: > Ammon Riley writes: > >> On Sat, May 16, 2009 at 12:56 PM, Junio C Hamano wrote: >>> Ammon Riley writes: >>> >>>> There was an xmalloc() for path, but I didn't see a corresponding free(). >>>> Does it happen somewhere else that I'm not expecting? >>> >>> It implicitly happens in exit() in git.c:handle_internal_command() >>> after cmd_init_db() returns the control to it. >> >> Ah. Naturally. :) >> >> So if I were to write a long-lived application (such as a custom UI) that >> links to libgit, and bypasses those functions to call init_db() (and other >> functions) directly, all those implicit free-on-exit() turn into memory >> leaks. > > Correct, and there are other much larger issues to worry about. > > That's why there is a separate libgit2 effort in progress. Okay, cool! I wasn't aware of that -- I'll take a look at it. In the meantime, are small patches for this type of issue welcome if I run across others, or would you prefer I let them lie? Cheers, Ammon ```