Thread: Driver manager detection in 08.02.0400 is broken
The detection of the driver manager in 08.02.0400 (unixODBC vs iODBC) doesn't work at all. No matter what you say as option, it always picks unixODBC. Apparently, somebody didn't understand what the AC_ARG_WITH macros do. That's why we had used PGAC_ARG_* before. What was the point of that change anyway? Somebody has been creating a big mess in configure.ac without much explanation. -- Peter Eisentraut http://developer.postgresql.org/~petere/
From: "Peter Eisentraut" <peter_e@gmx.net> To: <pgsql-odbc@postgresql.org> Sent: Friday, May 04, 2007 7:14 PM Subject: [ODBC] Driver manager detection in 08.02.0400 is broken > The detection of the driver manager in 08.02.0400 (unixODBC vs iODBC) doesn't > work at all. No matter what you say as option, it always picks unixODBC. > Apparently, somebody didn't understand what the AC_ARG_WITH macros do. > That's why we had used PGAC_ARG_* before. What was the point of that change > anyway? Somebody has been creating a big mess in configure.ac without much > explanation. Eh? , I use iODBC on FreeBSD. I have the positive proof of the test that uses it. How does it make trouble? P.S) I had propriety and confirmed it to you first. However, I think your opinion and correction to be a surprise attack. ..... Regards, Hiroshi Saito
Am Freitag, 4. Mai 2007 13:37 schrieb Hiroshi Saito: > Eh? , I use iODBC on FreeBSD. I have the positive proof of the test that > uses it. How does it make trouble? I said --without-unixodbc and it still turned it on. -- Peter Eisentraut http://developer.postgresql.org/~petere/
> Am Freitag, 4. Mai 2007 13:37 schrieb Hiroshi Saito: >> Eh? , I use iODBC on FreeBSD. I have the positive proof of the test that >> uses it. How does it make trouble? > > I said --without-unixodbc and it still turned it on. Ahhh, I understood. I use this. configure --enable-unicode --with-iodbc --with-odbcver=0x0300 --enable-pthreads It is necessary to adjust it. or write the document. Thanks! Regards, Hiroshi Saito
Am Freitag, 4. Mai 2007 13:37 schrieb Hiroshi Saito: > Eh? , I use iODBC on FreeBSD. I have the positive proof of the test that > uses it. How does it make trouble? It seems it does end up working; just the selection logic is faulty in some cases. I'm working on a fix. -- Peter Eisentraut http://developer.postgresql.org/~petere/
Am Freitag, 4. Mai 2007 13:58 schrieb Hiroshi Saito: > > Am Freitag, 4. Mai 2007 13:37 schrieb Hiroshi Saito: > >> Eh? , I use iODBC on FreeBSD. I have the positive proof of the test that > >> uses it. How does it make trouble? > > > > I said --without-unixodbc and it still turned it on. > > Ahhh, I understood. > > I use this. > configure --enable-unicode --with-iodbc --with-odbcver=0x0300 > --enable-pthreads > > It is necessary to adjust it. or write the document. I've cleaned up the configure.ac file so that the --without- and --disable- options actually mean "turn off". I'm not really familiar with the changed driver manager detection logic yet, but at least it should have less surprises now. -- Peter Eisentraut http://developer.postgresql.org/~petere/
From: "Peter Eisentraut" <peter_e@gmx.net>
> Am Freitag, 4. Mai 2007 13:58 schrieb Hiroshi Saito:
>> > Am Freitag, 4. Mai 2007 13:37 schrieb Hiroshi Saito:
>> >> Eh? , I use iODBC on FreeBSD. I have the positive proof of the test that
>> >> uses it. How does it make trouble?
>> >
>> > I said --without-unixodbc and it still turned it on.
>>
>> Ahhh, I understood.
>>
>> I use this.
>> configure --enable-unicode --with-iodbc --with-odbcver=0x0300
>> --enable-pthreads
>>
>> It is necessary to adjust it. or write the document.
>
> I've cleaned up the configure.ac file so that the --without- and --disable-
> options actually mean "turn off". I'm not really familiar with the changed
> driver manager detection logic yet, but at least it should have less
> surprises now.
Umm, Problem of my autoconf?
Putting files in AC_CONFIG_AUX_DIR, `config'.
configure.ac:64: error: possibly undefined macro: AC_DEFINE
If this token and others are legitimate, please use m4_pattern_allow.
See the Autoconf documentation.
configure: error: cannot find install-sh or install.sh in config ./config
inet% autoconf -V
autoconf (GNU Autoconf) 2.59
Regards,
Hiroshi Saito
Am Freitag, 4. Mai 2007 15:19 schrieb Hiroshi Saito: > Umm, Problem of my autoconf? > > Putting files in AC_CONFIG_AUX_DIR, `config'. > configure.ac:64: error: possibly undefined macro: AC_DEFINE > If this token and others are legitimate, please use m4_pattern_allow. > See the Autoconf documentation. > configure: error: cannot find install-sh or install.sh in config ./config Try running autoreconf -i The order in which you run the autotools and the options you pass can be pretty complicated, and I could reproduce your problem by running them manually, but the above should work and run all the necessary tools. -- Peter Eisentraut http://developer.postgresql.org/~petere/
From: "Peter Eisentraut" <peter_e@gmx.net> > Am Freitag, 4. Mai 2007 15:19 schrieb Hiroshi Saito: >> Umm, Problem of my autoconf? >> >> Putting files in AC_CONFIG_AUX_DIR, `config'. >> configure.ac:64: error: possibly undefined macro: AC_DEFINE >> If this token and others are legitimate, please use m4_pattern_allow. >> See the Autoconf documentation. >> configure: error: cannot find install-sh or install.sh in config ./config > > Try running > > autoreconf -i > > The order in which you run the autotools and the options you pass can be > pretty complicated, and I could reproduce your problem by running them > manually, but the above should work and run all the necessary tools. cvs HEAD passed all. Thanks.! Regards, Hiroshi Saito