H Christian, great work! This really helps. Thanks for supporting the migration efforts.
Regards, Heiner Christian Lohmaier wrote: > Hi *, > > I'm proud to announce that tinderbox will now also show checkin-data > for mercurial based cws. > > If you experience any problems, have general feedback on the feature, > please don't hesitate to write a mail/ping me on IRC. > > What is the feature about? > > Tinderbox checks for changes in the cws. If a change did occur after a > build has started, the results of that build is invalidated, the tree > is flagged as dirty. This is represented with the purple color on the > status-pages. > > This allow builders to requeue/rebuild cws based on that status and > even more important allows QA and DEV to judge whether the build > results actually reflect the current status of the code that was > checked in. > > See for example > > http://tinderbox.go-oo.org/fwk116/status.html > As of now, the MacIntel result is flagged as "dirty" with the purple color. > The previous build result (build failed) is no longer valid, since > there have been commits in the meantime. > The column on the left shows the commit-times along with the committing user. > > To view the actual checkin-data either click on the times on the very > left, or open the popup by clicking the name of the committer and > choose the links there. > > The VC-page will then show detailed info - (commit-message, affected > files, link to changeset on hg.services.openoffice.org) > > Please report any problems, be it with the HG based cvs and also with > the vc-display for svn based cws (where it is unfortunately not > possible to provide links to a site with the actual diff, since > svn.services.ooo doesn't support it). Could be that I introduced a > regression or two :-) > > ciao > Christian > PS: the login data is gathered via a local repository. > * the hg id <cwsurl> is queried and compared to the last seen > revision. If both are the same, processing ends. > * the cws is pulled into the local repository (hg pull <cwsurl>) > * the log is requested using the template mechanism (thanks to Heiner > for that hint) > hg --cwd $hgrepodir log --follow -P $old_id -r $current_id:$old_id -d > ">$lastcheckeddate" --template $template > (where lastcheckddate is mainly used to limit the results when there > is no old_id yet (then old_id is 0 - lastcheckdate is then set to 14 > days in the past) > And for the really curious: the template is as follows: > my $template = > "'{node|short}".$sep.'{date|hgdate|user}'.$sep.'{author|user}'.$sep.'{desc|escape|addbreaks}'.$sep.'{file_adds}'.$sep.'{file_dels}'.$sep.'{file_mods}'.$record_sep."'"; > > PPS: anyone else finds it strange, that there is a |firstline filter, > but no filter like |all_but_firstline? > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
