Hello.

Someone asked me about installing oracle module on hp-ux. They said always failed in 
install. 
So I let they know how to compile in my experience. But it was same thing.
I tried to fix makefile file and write a documentation about this problems. but 
anything coudn't help them.

I know this problem caused by oracle but hope you don't look on this.

Thank you for your modules, Tim.

Best Regards.


----- Original Message ----- 
From: "Tim Bunce" <[EMAIL PROTECTED]>
To: "Llewellyn Jim" <[EMAIL PROTECTED]>
Cc: "'PD Miller'" <[EMAIL PROTECTED]>; "'Tim Bunce'" <[EMAIL PROTECTED]>; 
<[EMAIL PROTECTED]>; "'Baxter, Lincoln'" <[EMAIL PROTECTED]>
Sent: Friday, June 01, 2001 1:09 AM
Subject: Re: make failure on HP-UX 11


> On Thu, May 31, 2001 at 03:49:33PM +0100, Llewellyn Jim wrote:
> > >
> > >I was able to figure out more about what was going on there by running it
> > >through the preprocessor only, and examining the output.
> 
> Thanks. I didn't have the energy to try to explain how to do that
> and was kind'a hoping someone would do it for me :)
> 
> 
> This is the line > in > >question in the original:
> > >
> > >     DBIc_THR_USER(imp) = DBIc_THR_USER_NONE;
> > >
> > >and this is what it becomes after preprocessing:
> > >
> > >     ((  imp  )->com.  std.thr_user )   = (   4294967295u l     ) ;
> > >
> > >I assume the intent was for that to be an unsigned long constant. What's
> > the
> > >space doing between them though?
> > >
> > >OK, U32_MAX is being turned into that mess. Perl's handy.h turns U32_MAX
> > into
> > >UINT32_MAX. I go and search for that, and find this in perl.h:
> > >
> > >#    ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
> > >#        define UV_MAX UINT32_MAX
> > >#    else
> > >#        define UV_MAX 4294967295U
> > >#    endif
> > >
> > >Obviously Larry Wall knows something about UINT32_MAX being broken 
> > >in HP-UX with
> > >gcc :-) It seems that the Perl makefile sets -DUINT32_MAX_BROKEN in the
> > Perl
> > >"cppflags".
> > >
> > >Therefore, may I suggest changing the definition of DBIc_THR_USER_NONE to
> > >something like this:
> > >
> > >#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
> > >#  define DBIc_THR_USER_NONE    (U32_MAX)
> > >#else
> > >#  define DBIc_THR_USER_NONE    4294967295U
> > >#endif
> > >
> > >With that, DBI compiles perfectly (well, except for some unused variable
> > >warnings from gcc)..
> 
> Actually I've already changed it to
> 
> #define DBIc_THR_USER_NONE (0xFFFFFFFF)
> 
> :)
> 
> Tim.


Reply via email to