On Sat, Nov 17, 2007 at 12:24:44PM +0100, Nagy Gabor wrote:
diff --git a/lib/libalpm/deps.c b/lib/libalpm/deps.c index 548b643..f545d14 100644 --- a/lib/libalpm/deps.c +++ b/lib/libalpm/deps.c @@ -470,38 +470,36 @@ static int dep_vercmp(const char *version1,
On Fri, Nov 16, 2007 at 11:20:30PM +0100, Nagy Gabor wrote: pmdepmod_t mod,
<snip>
+ /* check provisions */ + for(i = alpm_pkg_get_provides(pkg); i; i = i->next) { + char *provname = i->data; + char *provver = strchr(pkgname, ' ');
I think it should be provname instead of pkgname, which would explain why your broken pactest below passed. It fails now :) And after correcting these little errors, everything looks fine to me, so good job. Oops. Thanks, please correct this, too (I cannot resubmit the patch at this moment). Yes, first I reused the pkgname pkgversion variables in this case too, but gcc forced me to forget it (because of const char definition);-). Probably there is a git issue here (/me doesn't really familiar in git; and atm git makes my life much more difficult instead of easier): at home all the add*.py files passed; probably I forgot a git-add or dunno... [/me is really disappointed now: this patch may be a half-work; I had to review it again; of course I did 'checkout -f' after format-patch... so the final state is lost :-(]<- is this needed with all modifications?
Hm, why did you checkout -f ? Did you want to switch to another branch? If so, you always try with checkout first. and if it isn't happy because of local changes, then you carefully check what these local changes are. And I think git add is only necessary for the untracked files : that is, new files that don't exist yet in the git repo, like new pactests.