Re: pkg/38387: pkgsrc-2008Q1 fails to bootstrap on OpenSolaris. Compiles OK but ...

["Michael C. Vergallen" <mvergall%telenet.be@localhost>]

The following reply was made to PR pkg/38387; it has been noted by GNATS.

From: "Michael C. Vergallen" <mvergall%telenet.be@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: pkg/38387: pkgsrc-2008Q1 fails to bootstrap on OpenSolaris. 
Compiles
 OK but ...
Date: Sat, 12 Apr 2008 07:50:38 +0200

 -----BEGIN PGP SIGNED MESSAGE-----
 Hash: SHA1
 
 Uwe Klaus wrote:
 | The following reply was made to PR pkg/38387; it has been noted by GNATS.
 |
 | From: Uwe Klaus <uklaus%hgb-leipzig.de@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
 | Subject: Re: pkg/38387: pkgsrc-2008Q1 fails to bootstrap on OpenSolaris.
 |  Compiles OK but ...
 | Date: Wed, 9 Apr 2008 09:31:00 +0200 (CEST)
 |
 |  On Wed, 9 Apr 2008, Michael C. Vergallen wrote:
 |
 |  > With the latest modifcations it seems to build and install correctly.
 |  > Now on the issue of the bmake tests ..the problem when doing the tests
 |  > with the sunpro compiler on solaris sparc is that the dotwait test does
 |  > not function well it goes to sleep and then it just dies.for some
 reason
 |  > unknown to me...I tested each test separately by commenting them
 all out
 |  > but the first one and then gradually uncommenting them.until the
 failure
 |  > ~ took place... now I will test it again later on to see if my tests
 |  > where correct.... but the problem is there abouts...
 |  this problem is subject of PR#38221
 |
 |  Regards,
 |  Uwe
 |
 |
 Uwe,
 
 I have managed to track the failure down to the -j4 flag in the dotwait
 test as being the cause of the failure when compiling with the sunpro
 compiler suite. However I could not find a fix that would produce the
 output as in the test.exp file. When using the -B -j4 flag it does not
 die however  the output is singlty different.The main difference being
 in the output of the cycle test who fails to rebuild cycle the second
 time instead of just finishing.
 
 Hope this is useful to somebody who has a better knowledge of the sunpro
 ~ compiler.Because I do not have enough experience with the internal
 workings of the sunpro compiler to understand the reason why it fails.
 
 Regards,
 
 Michael
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2.0.7 (GNU/Linux)
 
 iD8DBQFIAE2u8DZO4jg1E8kRAvprAJ0Q20QNaSbo0px4oxRNsMxPbPK92gCdFCm5
 KUQQ44xM9OEn/gINmaDow+I=
 =39A8
 -----END PGP SIGNATURE-----