Comment 1 for bug 632894

I used bzr heads --dead-only to locate the checkin that I had lost... yay. But the mystery deepens.

I found TWO dead heads,
one looks like what I was working on,
and the other looks like what rebase should've generated.
What is very weird is that the rebase-generated one has got a merge on a commit where it wasn't before.

This is the one that I think looks like what I was working on:
---
revno: 2367
timestamp: Wed 2010-09-08 12:00:29 +0800
---
revno: 2366
timestamp: Wed 2010-09-08 12:00:07 +0800
---
revno: 2365
timestamp: Wed 2010-09-08 11:59:34 +0800
---
revno: 2364
timestamp: Wed 2010-09-08 10:29:00 +0800
---
revno: 2363
timestamp: Tue 2010-09-07 22:46:36 +0800
---
revno: 2362
timestamp: Tue 2010-09-07 22:28:04 +0800
---
revno: 2361
timestamp: Fri 2010-09-03 15:05:08 +0800
message:
  - Compile fix
---
revno: 2360 [merge]
timestamp: Fri 2010-09-03 11:28:51 +0800
message:
  - Merge from trunk, bug fixes
---
revno: 2359 [merge]
timestamp: Tue 2010-08-31 20:59:43 +0800
message:
  - Merge from trunk

The possible rebase one:
revno: 2367
timestamp: Wed 2010-09-08 12:00:29 +0800
---
revno: 2366
timestamp: Wed 2010-09-08 12:00:07 +0800
---
revno: 2365
timestamp: Wed 2010-09-08 11:59:34 +0800
---
revno: 2364
timestamp: Wed 2010-09-08 10:29:00 +0800
---
revno: 2363
timestamp: Tue 2010-09-07 22:46:36 +0800
---
revno: 2362
timestamp: Tue 2010-09-07 22:28:04 +0800
---
revno: 2361 [merge] **** WAS NOT A MERGE BEFORE
timestamp: Fri 2010-09-03 15:05:08 +0800
message:
  - Compile fix
---
revno: 2360
committer: Paul Harris <email address hidden>
branch nick: 4.0_clustering
timestamp: Tue 2010-08-31 21:04:59 +0800
message:
  - Compile fixes
---
revno: 2359 [merge]
timestamp: Tue 2010-08-31 20:59:43 +0800
message:
  - Merge from trunk