[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Monotone-devel] Failure To Complete Tests With Monotone Version 0.3
Anthony Edward Cooper
Re: [Monotone-devel] Failure To Complete Tests With Monotone Version 0.34 And 0.35
Thu, 31 May 2007 10:18:07 +0100
Thunderbird 22.214.171.124 (Windows/20070221)
I have tried to compile Monotone 0.35 for Solaris 9.0. I must be
going mad! I distinctly remember it successfully compiling the test
harness, running it and then getting a few tests that failed. However
when I did a clean build yesterday the test harness failed to build due
to unset_env not being ported to Solaris. I provided an implementation
in tester.cc (borrowed straight out of glibc) and this time it compiled.
However when ever the tester program is run with a test script it core
dumps (no symbol table info I am afraid). make check passes with flying
colours on the first two tests but when it gets to run_lua_tests it just
Unfortunately this is beginning to take up too much time... I only
want a certified (known to work) binary that will run on Solaris 9.0.
However the only binaries I could find were version 0.2x and we would
like at least 0.34 or ideally 0.35. Has anyone else managed to get it to
compile for Solaris 9.0? Is there a way of running the lua tests from a
Linux box but using the solaris binary over the network?
In order to get some debug info out of tester, where do I turn on -g
in the make file?
Many thanks in advance.
Thomas Keller wrote:
Many thanks for your quick response. I have put together a tar ball
of all the test directories that failed, minus the large test file
and the databases.
Looking at the logs they all fail with something like this:
address@hidden --dump=_MTN/server_dump --bind=localhost:31781
--rcfile=netsync.lua serve address@hidden
mtn: beginning service on localhost : 31781
/home/aecoope/monotone/monotone-0.35/mtn: operation canceled: Terminated
I'm not experienced enough to explain why the service fails immediately
after starting to serve; maybe others here have an idea what could cause
this SIGTERM signal?
I take your point about gcc version 3.2.2, however the limiting
factor here is Solaris 9.0. I had the devil's own job compiling gcc
3.4.4 for Solaris, I managed it in the end but I don't know what made
it work:-~ Anyway my main concern is that the monotone tests failed
for SPARC Solaris 9.0.
Again, 3.2.2 might be perfectly fine, I just stated that we do not have
build bots for it, thus can't see / test if anything breaks. You're of
course invited to provide such a build bot - Nathaniel will happily add
it to our setup =)
I am trying to compile monotone for work as we currently use CVS and
wish to switch to monotone for the branch management.
As a final note: Please target your next mails directly to
monotone-devel, otherwise the other people can't read and respond to it.
I'm just a part time hacker and translator - there are certainly more
experienced users / developers around here who could help you more than
I ever could.