Re: Parsing regression with gmime-2.6?

[Jameson Graef Rollins <jrollins@finestructure.net>]

[-- Attachment #1: Type: text/plain, Size: 1138 bytes --]

On Thu, 08 Mar 2012 16:40:00 -0500, Daniel Kahn Gillmor <dkg@fifthhorseman.net> wrote:
> i think the point here is that gmime used to support parsing them and 
> now it doesn't.
> 
> So if you've got a pre-existing notmuch index (built from when gmime 
> *did* parse them) and you're now running notmuch linked against a gmime 
> that *doesn't* parse them, then notmuch search will produce references 
> to messages that notmuch show can't produce output for.

This is unfortunate for those that suffer from this but again I don't
think the solution is a partial parsing of mbox files.

If this really is the situation, it seems like the *previous* behavior
was actually the problematic one, which has now been fixed, even though
it appears to be a regression.

I think I would argue that users who do fall in to this situation should
just simply convert all of their mbox files into proper messages.
That's the position that we've actually held all along: notmuch does not
support mbox and those with mbox files should just convert them to
individual messages with something like mbox2maildir or the like.

jamie.

[-- Attachment #2: Type: application/pgp-signature, Size: 835 bytes --]