pidgin3: src/gaim-client.h@7ecd4441fdc7 (annotated)

src/gaim-client.h@7ecd4441fdc7 (annotated)

src/gaim-client.h

Wed, 21 Dec 2005 18:36:19 +0000

author: Richard Laager <rlaager@pidgin.im>
date: Wed, 21 Dec 2005 18:36:19 +0000
changeset 12600: 7ecd4441fdc7
parent 11241: f62a71823fc3
child 12627: fa4c64a821d2
permissions: -rw-r--r--

[gaim-migrate @ 14934]
Enable the extra warnings regardless of --enable-debug.

Enable FORTIFY_SOURCE regardless of --enable-debug, adding a --disable-fortify flag to configure.

Enable (well, stop disabling) the missing initializer warnings.
This leads to warnings with: GValue v = {0,}; that must be worked around.
Basically, instead of:
GValue v = {0,};
...
g_value_init(&v, G_TYPE_FOO); /* or other use of the GValue */
We'd need to do:
GValue v;
...
v.g_type = 0;
g_value_init(&v, G_TYPE_FOO); /* or other use of the GValue */

Fix several cases of missing initializers. I don't think any of these are bugs, but having this warning seems like a good idea. It might prevent us from making a mistake in the future.

While I was fixing missing initializers, I optimized substitute_simple_word in plugins/spellchk.c, in the same way as I did substitute_word before. Yes, I'm bad for committing these together.

Added a --enable-fatal-asserts flag to configure. As the name implies, this makes g_return_... guards fatal. This is a useful flag to run on a debug copy of Gaim. It will make it very clear if your changes have triggered one of these guards. It's also useful in detecting g_return_... abuse, which helps prevent crashes if Gaim is compiled with G_DISABLE_ASSERT defined.

 #ifndef _LIBGAIM_H_INCLUDED_
 #define _LIBGAIM_H_INCLUDED_
 #include <glib.h>
 #include "gaim-client-bindings.h"
 G_BEGIN_DECLS
 void gaim_init(void);
 G_END_DECLS
 #endif