Jean, On 9/6/05, Jean Delvare <[EMAIL PROTECTED]> wrote: > > Hi John, > > [Jean Delvare] > > > Also, your extension of the sort test looks interesting, so I plan to > > > commit it at the same time. > > [John Vandenberg] > > that extension to the sort test is necessary in order to ensure sort > > does the right thing wrt the added files. At present, the test is > > only verifying that the "files_in_patch" come back sorted; whereas I > > think it is supposed to also sort the "filenames_in_patch" that are > > not part of "files_in_patch". > > files_in_patch is actually supposed to come back in order, not sorted > (unless --sort is used) and that's what the test was checking, together > with the fact that added files are put after the existing ones (but that > one would be quite hard to fail.)
Argh. that would be what I meant to say. > ... I fully agree that this wasn't > sufficient. Even with your extension it hardly is, as having two values > come back in the expected order has a 50% probability to happen. We > probably should add even more files in this test to make it more rebust. Adding more files to the test sounds good. > > Do your changes pass my test? I will test your change today to > > confirm it works on OS X, Solaris and MSYS. > > Yes, I tested my code with your extended test case and it passed OK. Great. I have tested your change on all but OSX; there is no point waiting, as I can always use gawk in the unlikely event that OSX awk doesn't like your replacement code. Thanks for tackling this compatibility problem, and coming up with a better solution. -- John _______________________________________________ Quilt-dev mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/quilt-dev
