Thanks Maceij for the information -
>>Currently I had not much time to do something with this project,
>>but I'll try to make it cleaner soon, write some tests (so far I was
>>not able to run original tests...) and to put this into official Zope
I use DCOracle2 with Sun Solaris and Windows XP for direct Oracle
connectivity. I do not use Zope, so I only use the dco2.pyd part only.
The original author of this package was wise enough to keep dco2.pyd
completely separate from the Zope pieces.
I know that dco2.c for Windows is slightly different from the Solaris
dco2.c for Windows adds the line:
right at the start of Module initialization section.
I mention that to turn attention to the need of trying to keep
descriptions clear and the target platform of a package, and to keep
divergences at a minimum.
I wonder if simple IFDEF's can do the trick.
> On a different note:
> The 'official' distribution of DCOracle2 (from zope.org) compiled fine
> with Oracle 9i on Solaris. However, with Oracle 10g on Solaris, I ran
> into a compilation error and I had to add the line:
> typedef unsigned int dword;
> I did not have the time to investigate the reason that this happened
> with Oracle 10g.
>>Thanks for this info. We have this DCOracle2 compiled with Oracle10 on
>>RedHatEnterprise 4. Unfortunatelly I have no Solaris. But would be
>>if you can check dco2.c from this version on Solaris machine. At last
>>I'll add a note about this Solaris issue into INSTALL file.
It will be better to investigate further as to what is causing this
error related to dword to appear on Solaris with Oracle 10g but not
Oracle 9i. Maybe a note in INSTALL file turning attention to this issue
in case it arises would be fine.
By the way, this line:
typedef unsigned int dword;
goes on line 187 right before the line:
#define NATIVENUMERIC 1
Zope-DB mailing list