On 28/09/2011 20:56, Colomban Wendling wrote:
Le 28/09/2011 19:36, Frank Lanitz a écrit :
On Sun, 25 Sep 2011 20:11:06 +0200 [...]
Just want to bring it up again. Release is near so at least we should add a comment to api documentation about this.
Right. Fixed now in r5965, thanks.
If we want to deprecate the use of NULL here, I think we better do it after the release so plugin developers have a chance to fix their use until next-next release.
Thanks for fixing my mistake. I think there's no need to stop accepting null on utils_string_replace_all. It's bad to change plugin API behaviour unless the old use will get caught by the compiler.