First of all, I want to thank MoinMoin developers for the great
software. MoinMoin is the only tool I found suitable for handling
personal notes. I'm using it for a few years and can sey it is one of
the most valuable tool for me.
My usage patter of MoinMoin is rather specific. I keep there random
notes, but because I neet to access them on several computers that do
not have network access to each other, so I have to keep an autonomous
instance of MoinMoin on every host. I've tried to keep them in sync
using the following procedure.
Instances A and B needs to be synchronized, I create auxiliary wiki
instace M on a removable media (flash). Then I configure SyncJobs:
A <-> M and B <-> M. In theory, moving M between sites should allow
keeping A and B in sync. But, it seems, I found one corner case.
Mobile instance M is vulnerable to accidental loss, so I make frequent
backups. I thought, If I lose M, I may recover it from a recent backup
and that is OK, but current WikiSync implementation stores part of M
state (page revision numbers) in peer instances A, B. If M is restored
from backup this state information could become invalid. This may lead
to random stack traces (the best case) or silent data corruption
For example, some revisions that were in M before loss may not exist in
backup, this case leads to invalid FROM_REV and TO_REV stack
The other case, may happen if M loses some page revision, but
then gain it from the other source. This case leads to silent
page data corruption
I think, the problem is hard to resolve and don't expect it will be
fixed soon, but, probably, we can put a word of warning to the docs,
that one should avoid use of WikiSync if he uses Backup/Restore?
Get latest updates about Open Source Projects, Conferences and News.