[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Monotone-commits-nodiffs] Revision 47cb4fe79efb6064d0331c6137b5ec175c0c
From: |
monotone |
Subject: |
[Monotone-commits-nodiffs] Revision 47cb4fe79efb6064d0331c6137b5ec175c0c1901 |
Date: |
Fri, 2 May 2008 00:59:43 +0200 |
-----------------------------------------------------------------
Revision: 47cb4fe79efb6064d0331c6137b5ec175c0c1901
Ancestor: c31b7a50c71adf4f537f1ed121c44870a3a418ec
Author: address@hidden
Date: 2008-05-01T22:17:09
Branch: net.venge.monotone.cvsimport-branch-reconstruction
Modified files:
rcs_import.cc
ChangeLog:
* rcs_import.cc: log_path(): added the size of the blob to that debugging
function
Collect the oldest event per fil in the cycle. That helps with
early-aborting breadth first searches in our graph.
Instead of only traversing one level of dependencies, we now check the
full history (optimized with upper early-aborting) when checking if an
event depends on any other event in the cycle.
Dropped the concept of 'strange_blobs', instead we now keep track of
any kind of blob we can split to resolve the cycle. That helps in case
we cannot split any of them by timestamp, because none of them has
a time gap.
Prefer splitting blobs which consist of only type 2 and type 3 events
over others with additional type 1 events (which we don't would have
to randomly put on one side of the split).
Added quite a large portion of code duplication for debugging, but
printing out the cycle which couldn't be split really helps in lots
of cases.
monotone --db={your.database} diff
--revision=c31b7a50c71adf4f537f1ed121c44870a3a418ec
--revision=47cb4fe79efb6064d0331c6137b5ec175c0c1901
----------------------------------------------------------------------
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [Monotone-commits-nodiffs] Revision 47cb4fe79efb6064d0331c6137b5ec175c0c1901,
monotone <=