[PATCH] Re: Output more detailed error on corrupt LocalStatus

[chris coleman]

On Wed, May 04, 2011 at 12:55:28PM +0200, Sebastian Spaeth wrote:
> On Tue, 3 May 2011 11:13:28 -0700 (PDT), chris coleman  wrote:
> > It uses some python 2.6 syntax for the exception handling, and compile failed.  
> 
> Ohh right, for python 2.5 you would need to use it as you did. Nice
> catch. 

Could you resend a fixed patch, please?

> > Indeed that corrupt line is there.  but it's a sequence of non-consecutive numbers... Here is an excerpt:
> > 
> > 40213:
> > 40214:
> > 40215:
> > 40217:R
> > 40218:S
> > 40229:S
> > 4023(0x00)(0x00)P(0x00)(0x01)(0x00)(0x00)(0x00)0:S
> > 40261:S
> > 40262:S
> > 40264:
> > 40269:S
> > 40270:
> > 40276:
> > 40279:R
> 
> Actually, all other lines are OK, they just have no flags. The only
> damaged line is:
> 
> 4023(0x00)(0x00)P(0x00)(0x01)(0x00)(0x00)(0x00)0:S
> 
> I suspect that it is supposed to mean:
> 40230:S

I wonder if it would worth to store more information about fetched
mails. If we had the Message-ID header or filename for example, we could
automatically repair the database.

---------------------------------------- 
Nicolas ,

I agree it's wise to store the one additional piece of info (Message-ID appears like it could be a unique key), making it possible to regenerate the damaged file in a few seconds, up to a minute.

Without the Message-ID to identify each message absolutely, is it possible there could be a condition where the data in LocalStatus becomes useless, and that would force re-downloading and re-uploading the entire mailbox. On this particular INBOX there are 12385 messages - at 3-5 seconds per message , it would take ~13 hours to download all messages ( 26 hours if uploading and downloading each message).

Chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/offlineimap-project/attachments/20110504/d94cd871/attachment-0001.html>