Howdy-- With some of Danek's recent (or maybe still private) updates to webrev.sh, he's addressing
> 6793448 webrev should automatically link to defect.opensolaris.org bugzilla > bugs > http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6793448 Is it time to ask that question more generally? Ie for a webrev, how many different information tracking systems might it reference? What information is pertinent about each? How do you differentiate references? It might be overkill, but I think there's some value in keeping a master list, at least for webrev use, or any use that requires resolving references (comment checkers, cross referencing, etc). I think that (at least) the following information is useful, for each defect tracking system: > unique prefix > info > url > regex > contact ...and possibly also "on-SWAN URL," to be set if different from general URL. Or maybe that's a Sun internal thing and others are offended by the mere mention of SWAN. For example, I already use at least four such systems: > prefix: sun > info: > url: http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=%s > swanurl: http://monaco.sfbay.sun.com/detail.jsf?cr=%s > regex: \([0-9]\{7\}\) > contact: Valerie's e-mail address :) > prefix: doo > info: http://defect.opensolaris.org/ > url: http://defect.opensolaris.org/bz/buglist.cgi?quicksearch=%s > swanurl: > regex: \([0-9]\{1,6\}\) > contact: tools-discuss@opensolaris.org > prefix: arc > info: > url: http://www.opensolaris.org/os/community/arc/caselog/%s > swanurl: http://sac.eng.sun.com/%s > regex:\([A-Z]\{1,2\}ARC\)[ /]\([0-9]\{4\}\)/\([0-9]\{3\}\) > contact: arc-disc...@opensolaris.org > prefix: bgc > info: http://bugs.grommit.com/ > url: http://bugs.grommit.com/show_bug.cgi?id=%s > swanurl: > regex: \([0-9]\{1,4\}\) > contact: stevel The prefixes could be output along with appropriately substituted URLs. Ie, webrev might output in the comments section for a file something like: sun XXXXXXX synopsis arc 2009/YYY case_title doo ZZZZ summary ...where the prefix is a link to the info URL (if provided), and the ids are links to the URL (off or on SWAN). Ok, I know a couple things about this example: some of the regex are not exactly correct, and the substitution may be more complicated than I have implied above (ie the ARC case log URL uses two components on opensolaris, but three on SWAN). And the pattern spaces for some of the regex's overlap, but I don't think that's important, or at least it's not something I care about trying to solve right now. I figured that three letters should provide appropriate differentiation, but I'm not wedded to it. The main advantage to keeping a registry of systems: tools like webrev or comment checkers could be configurable to list and prioritize the valid backends. This would also make it easier to mix references in webrevs and changeset comments, to reference fixes in upstream systems, and to evolve policies and best practicies about embedding references in changeset comments. Thoughts? What other information should be tracked? Where should this go? (I think in a repository underneath the tools community, possibly by itself?) Optimistically, this note will result in a converation. I'll attempt to summarize, file a bug, and followup once it settles. If I'm the only one that cares, that should be pretty simple cut-and-paste... Anyone? --Mark _______________________________________________ tools-discuss mailing list tools-discuss@opensolaris.org
