Matt Turner <matts...@gmail.com> writes: > On Mon, Sep 30, 2013 at 1:44 PM, Eric Anholt <e...@anholt.net> wrote: >> The previous interface relied on a static struct, which meant tha the >> driver didn't get a chance to edit the struct before the struct got used. >> For megadrivers, I want to return a variable struct based on what driver >> is getting loaded.
>> diff --git a/src/glx/dri_common.c b/src/glx/dri_common.c
>> index f1d1164..16f820f 100644
>> --- a/src/glx/dri_common.c
>> +++ b/src/glx/dri_common.c
>> @@ -188,9 +188,25 @@ driOpenDriver(const char *driverName)
>> }
>>
>> _X_HIDDEN const __DRIextension **
>> -driGetDriverExtensions(void *handle)
>> +driGetDriverExtensions(void *handle, const char *driver_name)
>> {
>> const __DRIextension **extensions = NULL;
>> + const __DRIextension **(*get_extensions)(void);
>> + char *get_extensions_name;
>> +
>> + asprintf(&get_extensions_name, "%s_%s",
>> + __DRI_DRIVER_GET_EXTENSIONS, driver_name);
>
> My gcc complains about this:
>
> dri_common.c: In function 'driGetDriverExtensions':
> dri_common.c:197:12: warning: ignoring return value of 'asprintf',
> declared with attribute warn_unused_result [-Wunused-result]
> asprintf(&get_extensions_name, "%s_%s",
> ^
>
> Kind of silly, but whatever.
Silly as far as any malloc checking is silly. Wow, what a bad API:
... If memory allocation wasn't possible, or some
other error occurs, these functions will return -1, and the contents of
strp is undefined.
Seriously, not setting it to NULL on failure?
Applied this diff in v3:
- asprintf(&get_extensions_name, "%s_%s",
- __DRI_DRIVER_GET_EXTENSIONS, driver_name);
- if (get_extensions_name) {
+ if (asprintf(&get_extensions_name, "%s_%s",
+ __DRI_DRIVER_GET_EXTENSIONS, driver_name) != -1) {
pgpzenHsfMqDm.pgp
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
