Xavier wrote:
2007/6/29, Andrew Fyfe <andrew@neptune-one.net>:
I need to do some more research before playing with the code :) I saw the comment in pacman_deptest() and assumed there was still a problem with alpm_splitdeps().
I just find this code in pacman_deptest rather curious, because it indeed looks out of place. So I think your patch for this function is correct.
Ah one little thing, I wonder if this free(saved_target) that you already removed could be replaced by free(dep). Done :)
Another thing I missed earlier, why is this deps.h needed ? Isn't it only in the backend anyway ? I don't think pacman should have access to it.
I added deps.h so that I could access pmdepend_t->name, at the moment deptest is passing target (which could = foobar>=1.0) where the functions are expecting just the package name. So deptest is useless when dealing with versioned deps at the moment. Looking a little deeper I think pacman_deptest() should be using _alpm_checkdeps(). Are all _* functions private? Andrew