RE: gitweb: please show tags file history browsing
- From
- Tjernlund <tjernlund@tjernlund.se>
- Date
- Sep 12, 2010, 16:12 UTC
- Message-ID
- <001201cb5295$5c055850$141008f0$@se>
- In-Reply-To
- <201009121719.32390.jnareb@gmail.com>
Show 38 quoted lines
> -----Original Message----- > On Sun, 12 Sep 2010, Tjernlund wrote: > >> -----Original Message----- > >> From: Jakub Narebski [mailto:jnareb@gmail.com] > >> On Sun, 12 Sep 2010, Tjernlund wrote: > >>>> -----Original Message----- > >>>> From: Jakub Narebski [mailto:jnareb@gmail.com] > >>>> > >>>> "Tjernlund" <tjernlund@tjernlund.se> writes: > >>>> > >>>>> I would really like to see where the tags are when browsing > >>>>> file or directory history in gitweb. Would that be possible? > >> > >>>> While it probably be possible to show tags in 'history' view, it would > >>>> be not easy. The problem is that 'history' view shows only commits > >>>> that touch specified file or directory, and tagged commits usually do > >>>> not touch those files (at least if one is using "bump version number" > >>>> commits to tag them). > >>>> > >>>> So you would have: > >>>> 1. Design where to show those tags - they would be between shown > >>>> commits. > >>>> 2. Create code which shows some/all tags that are between commits in > >>>> the presence of nonlinear history, without affecting performance > >>>> too badly. > >>> > >>> Ah, that is too bad because I think it would really useful. > >>> Image browsing a drivers history in the linux kernel. Then it would be > >>> really nice to see what changes/bug fixes went into what release. > >> > >> First, you can help with the first issue even if you can't help with > >> the coding itself. > > > > I like the gitk way of showing tags, won't that work in gitweb too? > > Err... "gitk" shows tags, but so does 'shortlog' and 'log' view in > gitweb. The 'history' view in gitweb doesn't show intermediate tags, > but neither does "gitk -- <path>".
I know, I just meant the yellow note with tag name in it, how tags are displayed. When they are display is another matter :)