[Evolution] Z's camel status for the week

[NotZed <notzed helixcode com>]

I can hardly tell when the week began, really ...

But I think, since Monday ...

Camel:
 - Fixed the flags api and implemented it for mbox.
 - Implemented expunge for the mbox provider.
 - Moved the camel-mbox-search into camel-folder-search class
   and made it re-usable for other provider implementations.
 - Started moving the camel-mbox-summary into a new (old)
   class camel-folder-summary, which handles loading/saving
   compressed summaries, storing and accessing memory summaries,
   and generating the summaries (+ indexes) from raw data.
 - Some profiling and small speed improvements in the
   summariser (summarising only - maybe 20% improvement), indexing
   speed is now the slowest part of the process.
 - A few parser bugfixes and memory leak fixes.

Mail gui
 - Added a quick-search function.
 - Major peformance improvements in the handing of data to
   the e-table model used for the message list
 - Helped Chris with some other etable issues
 - Combined results is roughly 2 orders of magnitude
   faster startup and column sort times (really!).

Next week ...?

 - Track down some memory corruption issues (i'm hoping this
   just 'vanishes', as they appear to be in the soon to be
   redone mbox summary).
 - Setup a camel-mbox-summary which subclasses camel-folder-summary
   and controls summary stuff.
 - Robustification and structure changes to the way the mbox
   provider does things.

  I've created a new branch NEW_SUMMARY to perform most of this
  work on; the current head code should be 'stable enough' I
  hope, for most cases (it still has some issues though).

Basically, all of the tricky stuff that the mbox folder used
to do will have been reorganised into libcamel as reusable objects.
So mbox can be simplified more, and have it concentrate
on the job of making an mbox a fast and reliable disk-based
message database, rather than having it worry about message
structure, summarising details, and so forth.

 Michael