> First question is: does anyone know of a reason why this should *not*
> work, has it already been tried and abandonded?
No, all info in the porting comments are much older than MacOS-X. I'm
not sure how different it is from other *nixes, but if it is close you
should be able to make this work.
> I've gotten it to compile, partly by making changes correctly, and partly
> using the "bull in a china shop" technique of directly hacking source
> files so that they'll work on this platform but not others.
You can always clean this up later.
> I have gotten it down to 1 linker error, the TAS function is missing. (I
> have to stop for a few days because of a deadline I've got on a project.)
> I haven't had time to investigate what TAS is; I know where the source
> files are; I know there's a couple of platform-specific ones and a dummy
> one. I know that there's a need_tas flag in configure. I have tried
> building with that flag set to both yes and no; it made no difference.
> But I had trouble with make not rebuilding after some other changes to
> configure, which I believe was due to a bug (I think) in OS X having to
> do with links and modification dates.
TAS is "test and set", and is used to delimit critical sections of
shared data. You should have appropriate TAS functions available, since
we do run on Linux-PPC machines, as long as your assembler is compatible
with the code as currently written.
If it is not, then you may need to look into touching it up.
In either case, the ./configure tests are not identifying your platform
and matching up the right TAS code, and/or your compiler is missing a
predefined variable (e.g. "__GNU__" or "__PPC__" or ??) which the TAS
code uses to #define the right code to compile.
In response to
pgsql-ports by date
|Next:||From: Peter Eisentraut||Date: 2000-10-31 17:31:13|
|Subject: Re: pgsql on Mac OS X?|
|Previous:||From: Lamar Owen||Date: 2000-10-31 16:37:50|
|Subject: Re: Re: [HACKERS] Re: [GENERAL] 7.0 vs. 7.1 (was: latest