Re: [PATCH v9 5/5] Speed up log -L... -M

2013-03-24 Thread Eric Sunshine
On Sat, Mar 23, 2013 at 5:04 AM, Jeff King wrote: > On Sat, Mar 23, 2013 at 06:58:48AM +0100, Thomas Rast wrote: >> Eric Sunshine writes: >> > On Thu, Mar 21, 2013 at 8:52 AM, Thomas Rast wrote: >> >> This is a bit hacky and should really be replaced by equivalent >> >> support in --follow, and

Re: [PATCH v9 5/5] Speed up log -L... -M

2013-03-23 Thread Jeff King
On Sat, Mar 23, 2013 at 06:58:48AM +0100, Thomas Rast wrote: > Eric Sunshine writes: > > > On Thu, Mar 21, 2013 at 8:52 AM, Thomas Rast wrote: > >> This is a bit hacky and should really be replaced by equivalent > >> support in --follow, and just using that. However, in the meantime it > > > >

Re: [PATCH v9 5/5] Speed up log -L... -M

2013-03-22 Thread Thomas Rast
Eric Sunshine writes: > On Thu, Mar 21, 2013 at 8:52 AM, Thomas Rast wrote: >> This is a bit hacky and should really be replaced by equivalent >> support in --follow, and just using that. However, in the meantime it > > s/using/use/ I'm not a native speaker, but I really think 'using' is more

Re: [PATCH v9 5/5] Speed up log -L... -M

2013-03-21 Thread Eric Sunshine
On Thu, Mar 21, 2013 at 8:52 AM, Thomas Rast wrote: > This is a bit hacky and should really be replaced by equivalent > support in --follow, and just using that. However, in the meantime it s/using/use/ > speeds up 'log -M -L' by an order of magnitude. -- To unsubscribe from this list: send the

[PATCH v9 5/5] Speed up log -L... -M

2013-03-21 Thread Thomas Rast
So far log -L only used the implicit diff filtering by pathspec. If the user specifies -M, we cannot do that, and so we simply handed the whole diff queue (which is approximately 'git show --raw') to diffcore_std(). Unfortunately this is very slow. We can optimize a lot if we throw out files tha