[Gnucash-changes] r14198 - gnucash/trunk - Merge 'misc-backend' -r 14160:14184 into trunk.

[Derek Atkins]

Quoting Chris Shoemaker <c.shoemaker at cox.net>:

> I'm pretty sure that for all working configs, these are identical.
> They are both libdir.

That may be true today, but not something you should assume...

> Not that it matters that much, since the only difference between
> loading qsf and not is the extra schema failure error messages that
> qsf throws, but...

Well, that's only the case until we fix the QSF import/export.

> If you really want to change this, the _real_ problem, IMO, is that
> qsf has an identity-crisis.  If it's a qof module, then it should live
> in QOF's libdir and be loaded by qof, not gnucash.  If it's a gnucash
> backend/module, it should live in gnucash's libdir and be loaded by
> gnucash.  Gnucash should have has no business poking around in qof's
> libdir.

Indeed.  The "bug" is that qofsession.c:qof_session_load_backend() only
loads the qof-known backends if the application hasn't loaded any of its
own qof backends.  I'll go fix that, so we can remove QSF from gnc-engine.

>> I'll go ahead and fix this by changing the list of names into a list
>> of structures where we can define the library name and the directory..
>
> I'm pretty sure that can't make any difference, and in the long term,
> we probably shouldn't be keeping a list of libdirs.

True.  On the other hand, why should gnc-engine even have to know about
all the gnucash sources qof backends?  For example, shouldn't the postgres
qof backend only get loaded if we build with PG?

> -chris

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available