On Sep 8, 2009, at 12:33 PM, Adam Kocoloski wrote:
On Sep 8, 2009, at 10:59 AM, Brian Candler wrote:
Extra data point:
I've now recompiled using git sources instead of svn. I picked the
following commit, and it works:
commit 897dd975f8a1e97f05e4289a1b65984a1a640fc8
Author: jan <j...@13f79535-47bb-0310-9956-ffa450edef68>
Date: Sat Aug 29 16:49:22 2009 +0000
remove legacy erlang tests, etap covers these
git-svn-id: http://svn.apache.org/repos/asf/couchdb/tr...@809158
13f79535-47bb-0310-9956-ffa450edef68
i.e. couchdb talks HTTP, and /usr/local/etc/init.d/couchdb stop
works too.
But if I go to the next commit it fails as before:
commit 975d416bc86fb579d834edb6468dcfc533fa6b89
Author: kocolosk <kocol...@13f79535-47bb-0310-9956-ffa450edef68>
Date: Sun Aug 30 17:56:56 2009 +0000
refactor CouchDB startup procedure, see COUCHDB-216 for details
git-svn-id: http://svn.apache.org/repos/asf/couchdb/tr...@809392
13f79535-47bb-0310-9956-ffa450edef68
i.e. no listener on 5984; /usr/local/etc/init.d/couchdb stop
doesn't stop
any processes
Regards,
Brian.
Hi Brian, this is really useful. 975d41 should've resulted in a
broken build on Ubuntu; the next commit (r809407 or 7b1019 in the
apache git repo, not sure what it is in yours) fixed that problem
with a POSIX-compatible invocation of basename.
Nevertheless, I am worried that there is an R12B versus R13B issue
in that COUCHDB-216 refactoring. I vaguely recall a problem with
starting linked-in drivers in the init <0.1.0> process on R12B.
I'll try to dig into it a bit more if I can find an R12B build
somewhere.
I rebuilt R12B-5 on my OS X 10.6 laptop and couchdb trunk started fine.
Adam
