Front page | perl.daily-build |
Postings from January 2009
Max Maischein wrote: > Hello Abe, Hi Max, >>> Locally applied patches: >>> SMOKE????? >> >> Your reports seem to miss the rather important information about >> commit id. This seems to point at the fact that you do not have the >> '.patch' file. >> >> Can you tell me how you sync the source-tree so we can find a way to >> fake the .patch file. > Ooops - yes, that was the one issue I wanted to remember before starting > to send reports outside :( > > I'm copying a synced git tree with the following commands. The git repo > doesn't have the .patch and hence, things get interesting. I guess we > could (ab)use the code in make_patchnum.pl to find out the git > patchnumber for the current branch. ... > I haven't yet looked at what .patch should contain. I guess running > make_patchnum.pl with an installed perl instead of the just-built > miniperl should simply work, because all make_patchnum.pl does is run > git several times to extract the current state. Nope, make_patchnum.pl provides de full sha1 for patchlevel.h .patch is generated as part of the hole post-commit stuff for the rsync-archives. you can either rsync it from rsync://perl5.git.perl.org/perl-current/.patch or ftp it with wget from ftp://public.activestate.com/pub/apc/perl-current/.patch The .patch file contains 4 fields (branch, timestamp, sha1, git-describe). My choice to use the timestamp field, to identify the commit in the subject of the reports, might not have been the smartest as the timestamp is generated by the post-commit handler and not reproducible from the git repository (it's not the same stamp as the commit message). I think I will be changing this, so we can reproduce these fields from the git repository. thanks again for your interest in perl-core smoking + good luck, Abe. -- _ | "Pinky, are you pondering what I'm pondering?" ( ) | I think so, Brain, but if the plural of mouse is mice, wouldn't the X | plural of spouse be spice? (27 Sep 97 - Brain Acres) / \ |