User-Agent: Mozilla/5.0 (BeOS; U; BeOS BePC; en-US; rv:1.9a1) Gecko/20060325 Firefox/1.6a1 Build Identifier: Mozilla/5.0 (BeOS; U; BeOS BePC; en-US; rv:1.9a1) Gecko/20060325 Firefox/1.6a1 Build broken sometime in the last 24 hours. Suspect bug 301249 checkin as culprit. Reproducible: Always Steps to Reproduce: 1. build Firefox 2. 3. Actual Results: build breaks (see attachment for console log) Expected Results: completed build Since BeOS tinerboxen are offline at the moment, we undertake builds daily to catch issues like this.

Hmm, I have no idea how to fix this.

I have deja vu feeling about such bustage. But cannot recall details.

(In reply to comment #3) > I have deja vu feeling about such bustage. But cannot recall details. > I got the same feeling as soon as I read your comment. I believe it goes back to late December when post-build lib reordering stopped working.

Thank you, fyysik! How could I have forgotten that unlucky Friday the 13th? Please see https://bugzilla.mozilla.org/show_bug.cgi?id=323379. Error messages are pretty much identical. I'll try to dig into it later tonight.

this affects SeaMonkey as well.

This is a simple regression. Somehow, the patch from bug 323379 was backed out of the current gversion of sign.sh. I applied this against trunk and build is now successful.

Doug, nothing has been backed out of the trunk (of NSS) for this file. The tip of the trunk is rev 1.17, which is your patch, checked in by Wan-Teh, months ago. What version of this file do you have? By what cvs tag did you pull it?

(In reply to comment #8) > Doug, nothing has been backed out of the trunk (of NSS) for this file. > The tip of the trunk is rev 1.17, which is your patch, checked in by Wan-Teh, > months ago. > > What version of this file do you have? > By what cvs tag did you pull it? > CVS "Entries" shows version 1.16 for this file. CVS Tag is NNSS_3_11_RTM. To my knowledge, I have not made any specific attempts to pull anything other than the most current. The problem started sometime between 3/26 and 3/27. When I did a "make -f client.mk", somehow version 1.17 was replaced with 1.16.

Perhaps a consequence of the fix for bug 301249 ?

(In reply to comment #10) > Perhaps a consequence of the fix for bug 301249 ? > That was my thought when I opened the bug and researched a bit.

well, looks like we build NSS 3.11 RTM, while the fix was made later on the branch. maybe we should pull the branch instead?

looks like bug 332050 is where the real discussion is

There are 3 or 4 bugs that all seem to be duplicates, except that they report the problem in the context of different products and/or OSes. This bug is one of those duplicates. As reported in bug 301249, the tag NSS_3_11_20060331_TAG has been created as a replacement for NSS_CLIENT_TAG, and IINM, benjamin has modified client.mk to use the new tag. BTW, note that in the new scheme, unlike NSS_CLIENT_TAG (which was freqently changed and moved), NSS_3_11_20060331_TAG will be immutable, unmoving. When the time comes for a newer tag, a NEW dated tag will be created, and will be used in client.mk or whereever. So, is this bug now resolved/fixed?

Yes, I believe this bug can be marked "RESOLVED".

*** This bug has been marked as a duplicate of 301249 ***