Bruce McAlister wrote:
Rick Widmer wrote:


I am working on making all the back ends work the same way, but not
being a PostgreSQL user, and not seeing all that many people using it,
it isn't a high priority for me.  I'll get around to it in or soon after
version 6 hits the street.  That may be a year or more out.

WOW, thats a long time to wait :)

We just started discussing version 6, so it is very early in the cycle, and I still have a list of things that go into 5.4 before I start on 6. Too bad I can't put much more than a couple weekends a month into it on a regular basis.


Have you considered the possibilty of unixODBC/iODBC, or an ODBC manager
of sorts? I would have thought that it would be easier to point it to an
OdBC manager and let it take care of the specific driver details, or is
there something else I am missing :)

Yes.  That is pretty much how I plan to handle database independence in 6.


I'm no C expert, but I will try to look around to see if I can do
anything that may help my cause :)

Neither am I. I'm better at PHP, and Delphi but I need vpopmail, and I've been using it nearly a decade, so it is time I do my part supporting the Open Source software I use.


I did manage to get a little further with the sql logging
option, it appears that the "sqlBufUpdate" issue was a typo, because it
is defined in vpgsql.c but is defined as "SqlBufUpdate". I just changed
the 2 places in vauth.c to be an uppercase S and it compiled further

That is the basic idea but it was the wrong approach. Would you please change vauth.c back the way it was, then make changes in vpgsql.c to make it work. Once that is done, then do a diff -u oldfile newfile and either post the results to this list, or upload it to the tracker on SourceForge. I'll patch CVS with it, and add you to the list of contributors in Changelog. It's really pretty easy.

The thing to remember, the rest of vpopmail is probably correct, all problems should be fixed in vpgsql.c. As long as you are testing and using your patches to vpgsql.c or vpgsql.h I'll apply them to the next release. If you think you need changes to any other files, you'll have to convince us why
..

now I'm getting the following error:

--------------------------------------------------------------------------
/usr/sfw/bin/gcc  -I/usr/sfw/include -Wall   -o vchkpw  vchkpw.o md5.o
hmac_md5.o libvpopmail.a -L/usr/lib -lpq  -lcrypt
Undefined                       first referenced
 symbol                             in file
PGresultStatus                      libvpopmail.a(libvpopmail_a-vauth.o)
ld: fatal: Symbol referencing errors. No output written to vchkpw
collect2: ld returned 1 exit status
gmake[2]: *** [vchkpw] Error 1
gmake[2]: Leaving directory `/export/home/user/build/qmail/vpopmail-5.4.25'
gmake[1]: *** [all-recursive] Error 1
gmake[1]: Leaving directory `/export/home/user/build/qmail/vpopmail-5.4.25'
gmake: *** [all] Error 2
--------------------------------------------------------------------------

Is this a library issue?

That is weird. PGresultStatus is the proper PostgreSQL function for returning query status, and if it is ever found, it should always be found. It is declared in libpq-fe.h somewhere in your system include files. On OpenSuSE 10.3 mine is in /usr/include/pgsql/.

Did you do a make clean before switching between compiling for MySQL and PostgreSQL? I suspect vpopmail has some dependencies that are not properly checked, so make clean is always a good idea when you change ./configure options.


Rick

!DSPAM:473c3c2b32001579221220!

Reply via email to