Re: pkg/41240: Perl5 build issue on opensolaris due to getconf

[Jens Rehsack <rehsack%googlemail.com@localhost>]

The following reply was made to PR pkg/41240; it has been noted by GNATS.

From: Jens Rehsack <rehsack%googlemail.com@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: pkg-manager%netbsd.org@localhost, gnats-admin%netbsd.org@localhost, 
 pkgsrc-bugs%netbsd.org@localhost, mvergall%telenet.be@localhost, 
 Havard Eidnes <he%NetBSD.org@localhost>
Subject: Re: pkg/41240: Perl5 build issue on opensolaris due to getconf
Date: Sat, 18 Apr 2009 00:21:47 +0000

 Michael C. Vergallen wrote:
 
 >  Havard Eidnes wrote:
 >  >  Can you please describe in a bit more detail how it fails?  Does
 >  >  it only emit warnings during the build, or does it completely
 >  >  fail to build?  I don't have the environment to try this out
 >  >  myself.
 >  >  
 >  Yes I will,
 >  
 >  Due to the fact that getconf on solaris is totaly broken and the author
 >  does not want to fix the package to give the correct values one has only
 >  one option and that is to
 >  ignore the output from getconf all together...this removes the compiler
 >  warnings all together and makes shure perl5 gets build correctly on the
 >  solaris/opensolaris platform.
 >  Someone did a partial fix in the original  source code but forgot to
 >  finalize the last 3 entries in the script who  get  used  by  solaris 
 >  10  & opensolaris...so I fixed those becuase
 >  the warning messages emitted on opensolaris turn the whole build into a
 >  mess and sometimes cause the program not to build at all...Mostly a
 >  cosmetic fix but it also has implications
 >  in the build process becuase incorrect compiler directives are deduced
 >  witch is a bad thing ... it is not leathel in most cases that the bad
 >  values get used but in certain cases it makes the
 >  build process fail.
 
 I've seen the same when compiling perl on OpenSolaris, but I hadn't time to
 fix it correctly. Michael - I cannot turn on my test machine the next few
 days - can you please attach the work/.work.log, so that someone could take
 a look what fails?
 
 Thanks in advance,
 Jens