Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Sat, 15 Jan 2011 15:06:53 +0100] rev 13266
revlog/parseindex: construct the nodemap if it is empty
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Sat, 15 Jan 2011 13:02:19 +0100] rev 13265
revlog: always add the magic nullid/nullrev entry in parseindex
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Sat, 15 Jan 2011 15:04:58 +0100] rev 13264
revlog/parseindex: no need to pass the file around
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Sat, 15 Jan 2011 12:44:28 +0100] rev 13263
parsers.c: fix comment
Benoit Boissinot <benoit.boissinot@ens-lyon.org> [Sat, 15 Jan 2011 12:28:10 +0100] rev 13262
perf: there is no lazy index anymore
Matt Mackall <mpm@selenic.com> [Wed, 12 Jan 2011 13:54:39 -0600] rev 13261
pure: update index parsing
Matt Mackall <mpm@selenic.com> [Wed, 12 Jan 2011 11:34:42 -0600] rev 13260
perf: fix ordering of invalidate in perfindex
Matt Mackall <mpm@selenic.com> [Tue, 11 Jan 2011 21:52:03 -0600] rev 13259
revlog: do revlog node->rev mapping by scanning
Now that the nodemap is lazily created, we use linear scanning back
from tip for typical node to rev mapping. Given that nodemap creation
is O(n log n) and revisions searched for are usually very close to
tip, this is often a significant performance win for a small number of
searches.
When we do end up building a nodemap for bulk lookups, the scanning
function is replaced with a hash lookup.
Matt Mackall <mpm@selenic.com> [Tue, 11 Jan 2011 17:12:32 -0600] rev 13258
revlog: introduce a cache for partial lookups
Partial lookups are always O(n), and often we look up the same
one multiple times.
Matt Mackall <mpm@selenic.com> [Tue, 11 Jan 2011 17:09:06 -0600] rev 13257
tags: avoid a pointless usage of revlog.nodemap