On Thu, Mar 20, 2014 at 10:26:31AM +1300, Karl Tomlinson wrote: > Erik Rose writes: > > > Once we have request-time rendering of indexed content, it will > > be a lot more straightforward to fetch and render arbitrary revs > > out of version control the same way. We just won't have > > analysis—only syntax coloring—since we won't have time to build > > the whole project during the request. ;-) MXR's ability to > > half-ass an analysis at request time is a deeper question with > > no clear answer, short of implementing an entire second analysis > > engine that operates more heuristically. DXR's strength right > > now is its exactness. Ideas are most welcome. > > Yes, I think there are two different purposes here. > > 1) For development, usually I just want the latest code, to chase > references/overrides/definitions/declarations/documentation > etc. around the code. > > 2) For permalinks to code from bugzilla, usually the main aim is > to identify what the code that is being discussed in a comment. > > There are occasions, such as constructing branch patches or > understanding how things worked in the past, where it would be > neat to have all the power of DXR in past snapshots, but this is > less common. I understand that the cost of performing and/or > archiving an analysis for every revision means this is not > feasible.
Note that for this, analysis of the tip of each of -release, -beta, -aurora and -esr* would be enough. That's significantly more than just doing -central, but it's also significantly less than doing every single revision. Mike _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform