[Evolution-hackers] Re: camel-private.h not installed

[Jules Colding <colding omesc com>]

On Wed, 2005-08-31 at 18:07 +0800, Not Zed wrote:
> Right, well, the url never really ever changes - if it does a new store
> object will be created anyway, so I don't think that should be a major
> issue.

OK.

Thanks a lot,
  jules


> 
> On Wed, 2005-08-31 at 09:52 +0200, Jules Colding wrote:
> > On Wed, 2005-08-31 at 13:58 +0800, Not Zed wrote:
> > > > which is awkward, to say the least, and not really an option for
> > > > distributed code.
> > > 
> > > Well you don't need to do that anyway, you could path it in a -I thing.
> > 
> > Sure, much better.
> > 
> > > > Access to a locking mechanism for externally developed components is
> > > > really necessary unless they should use homegrown solutions, which is
> > > > not an option either I guess.
> > > 
> > > Hmm, i'm not sure - it might depend on the particular lock.  It is
> > > probably possible to get away with not using any of those locks but just
> > > defining your own.  That's how camel-imap-* used to work, but because of
> > > other problems (mainly complexity and races due to the the 4 levels of
> > > interested parties, service, store, disco-store and imap-store), that
> > > particular implementation was changed.
> > 
> > Hmm... OK. 
> > 
> > I was of the impression that those particular locks in "camel-private.h"
> > was of mandated use due to design issues up-source, since everybody
> > seemed to use them. Well, I'll think something up myself then.
> > 
> > > > I am really not qualified at all to hack on some way to extract the
> > > > "to-be-public" parts of "camel-private.h". I would bet a really big part
> > > > of my right arm that much in Evolution depend on hard to spot properties
> > > > of the current implementation.
> > > >  
> > > > Any suggestions?
> > > 
> > > Given we are in hard code freeze, not sure; since moving it around
> > > requires some macro changes, which i guess are code.
> > > 
> > > Which locks are you trying ot access?  Do you really need them?
> > > Anything else in private is definitely private.
> > 
> > I am trying to synchronize access to the url in connect() with the store
> > lock. The url is really the only thing that I am currently aware of that
> > I should protect (right?). The backend server is fully thread-safe so
> > only locally shared resources are of any concern to me.
> > 
> > Thanks,
> >   jules
> >