You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Matthew Braithwaite <ma...@braithwaite.net> on 2000/11/02 07:17:16 UTC
Should `make check' work?
Several previous postings have indicated that people have been able to
run `make check' -- for example Branko's postings about builds on
various platforms, and Karl's Milestone 1 posting. I have not had
such luck.
I wanted to incorporate `make check' into the automatic builds, but
since it gave me some problems (starting from a clean checkout and
build), I thought it'd be worthwhile to ask whether it's *supposed* to
work, and whether spamming dev or individual developers when the tests
fail is desirable at this time.
The two tests that failed were:
gmake[2]: Entering directory `/usr/home/mab/subversion-stage/subversion/subversion/libsvn_ra_dav/tests'
Running all sub-tests in ra-dav-test...usage: ./ra-dav-test REPOSITORY_URL TARGET_DIR
at least one sub-test FAILED, check tests.log:
gmake[2]: *** [check] Error 1
and
gmake[2]: Entering directory `/usr/home/mab/subversion-stage/subversion/subversion/client/tests'
sh ./svn-test.sh
Checking out t1.
A t1/iota
A t1/A
A t1/A/mu
A t1/A/B
A t1/A/B/lambda
A t1/A/B/E
A t1/A/B/E/alpha
A t1/A/B/E/beta
A t1/A/B/F
A t1/A/C
A t1/A/D
A t1/A/D/gamma
A t1/A/D/G
A t1/A/D/G/pi
A t1/A/D/G/rho
A t1/A/D/G/tau
A t1/A/D/H
A t1/A/D/H/chi
A t1/A/D/H/psi
A t1/A/D/H/omega
Modifying t1/A/D/G/pi.
Modifying t1/A/mu.
Adding t1/newfile1.
This is added file newfile1.
Adding t1/A/B/E/newfile2.
This is added file newfile2.
Deleting versioned file A/D/H/omega, with --force.
apr_error: #2, src_err 0, canonical err 2 : No such file or directory
svn_client_delete: error deleting t1/A/D/H/omega
Committing changes in t1.
apr_error: #2, src_err 0, canonical err 2 : No such file or directory
contents_identical_p: apr_open failed on `./A/B/lambda'
Updating t2 from changes in t1.
svn_error: #21008 no element found at line 15
gmake[2]: Leaving directory `/usr/home/mab/subversion-stage/subversion/subversion/client/tests'
Re: Should `make check' work?
Posted by Karl Fogel <kf...@galois.collab.net>.
Greg Stein <gs...@lyra.org> writes:
> > May I, then, fix it so that ra-dav-test isn't invoked (without arguments)
> > by `make check', so that I can get the all-important 0 return value from
> > a top-level make check?
>
> Oh geez... of course! (IMO) No need to ask me about it.
Seconded. Don't be afraid to commit, Matt -- that's why we have
version control! :-)
Re: Should `make check' work?
Posted by Greg Stein <gs...@lyra.org>.
On Wed, 1 Nov 2000, Matthew Braithwaite wrote:
> On Wed, Nov 01, 2000 at 11:29:36PM -0800, Greg Stein wrote:
> > The ra-dav-test program won't work without a target server (specified on the
> > cmd line). We aren't really set up to automate that one right now, so I'd
> > say don't worry about it.
>
> May I, then, fix it so that ra-dav-test isn't invoked (without arguments)
> by `make check', so that I can get the all-important 0 return value from
> a top-level make check?
Oh geez... of course! (IMO) No need to ask me about it.
Cheers,
-g
--
Greg Stein, http://www.lyra.org/
Re: Should `make check' work?
Posted by Matthew Braithwaite <ma...@braithwaite.net>.
On Wed, Nov 01, 2000 at 11:29:36PM -0800, Greg Stein wrote:
> The ra-dav-test program won't work without a target server (specified on the
> cmd line). We aren't really set up to automate that one right now, so I'd
> say don't worry about it.
May I, then, fix it so that ra-dav-test isn't invoked (without arguments)
by `make check', so that I can get the all-important 0 return value from
a top-level make check?
Re: Should `make check' work?
Posted by Greg Stein <gs...@lyra.org>.
The ra-dav-test program won't work without a target server (specified on the
cmd line). We aren't really set up to automate that one right now, so I'd
say don't worry about it.
Cheers,
-g
On Wed, Nov 01, 2000 at 11:17:16PM -0800, Matthew Braithwaite wrote:
> Several previous postings have indicated that people have been able to
> run `make check' -- for example Branko's postings about builds on
> various platforms, and Karl's Milestone 1 posting. I have not had
> such luck.
>
> I wanted to incorporate `make check' into the automatic builds, but
> since it gave me some problems (starting from a clean checkout and
> build), I thought it'd be worthwhile to ask whether it's *supposed* to
> work, and whether spamming dev or individual developers when the tests
> fail is desirable at this time.
>
> The two tests that failed were:
>
> gmake[2]: Entering directory `/usr/home/mab/subversion-stage/subversion/subversion/libsvn_ra_dav/tests'
>
> Running all sub-tests in ra-dav-test...usage: ./ra-dav-test REPOSITORY_URL TARGET_DIR
>
> at least one sub-test FAILED, check tests.log:
>
> gmake[2]: *** [check] Error 1
>
> and
>
> gmake[2]: Entering directory `/usr/home/mab/subversion-stage/subversion/subversion/client/tests'
> sh ./svn-test.sh
>
> Checking out t1.
> A t1/iota
> A t1/A
> A t1/A/mu
> A t1/A/B
> A t1/A/B/lambda
> A t1/A/B/E
> A t1/A/B/E/alpha
> A t1/A/B/E/beta
> A t1/A/B/F
> A t1/A/C
> A t1/A/D
> A t1/A/D/gamma
> A t1/A/D/G
> A t1/A/D/G/pi
> A t1/A/D/G/rho
> A t1/A/D/G/tau
> A t1/A/D/H
> A t1/A/D/H/chi
> A t1/A/D/H/psi
> A t1/A/D/H/omega
> Modifying t1/A/D/G/pi.
> Modifying t1/A/mu.
> Adding t1/newfile1.
> This is added file newfile1.
> Adding t1/A/B/E/newfile2.
> This is added file newfile2.
> Deleting versioned file A/D/H/omega, with --force.
>
> apr_error: #2, src_err 0, canonical err 2 : No such file or directory
> svn_client_delete: error deleting t1/A/D/H/omega
> Committing changes in t1.
>
> apr_error: #2, src_err 0, canonical err 2 : No such file or directory
> contents_identical_p: apr_open failed on `./A/B/lambda'
> Updating t2 from changes in t1.
>
> svn_error: #21008 no element found at line 15
> gmake[2]: Leaving directory `/usr/home/mab/subversion-stage/subversion/subversion/client/tests'
--
Greg Stein, http://www.lyra.org/
Re: Should `make check' work?
Posted by Ben Collins-Sussman <su...@newton.collab.net>.
Branko =?GB2312?B?uDBpYmVq?= <br...@xbc.nu> writes:
>
> That's weird... A checkout about three hours ago magically passed "make
> check" on Solaris, not counting the libsvn_ra_dav tests. Some (APR?)
> fixes in the last 24 hours must've fixed the Solaris bug.
Indeed, we fixed it this morning.
Re: Should `make check' work?
Posted by Branko 0ibej <br...@xbc.nu>.
Ben Collins-Sussman wrote:
> Karl and I accidentally broke the milestone 1 test (which lives in
> client/tests); that's what you're seeing up there. We're gutting
> some major code bits right now; this will be fixed soon. :)
That's weird... A checkout about three hours ago magically passed "make
check" on Solaris, not counting the libsvn_ra_dav tests. Some (APR?)
fixes in the last 24 hours must've fixed the Solaris bug.
--
Brane
Re: Should `make check' work?
Posted by Ben Collins-Sussman <su...@newton.collab.net>.
matt@braithwaite.net (Matthew Braithwaite) writes:
> Deleting versioned file A/D/H/omega, with --force.
>
> apr_error: #2, src_err 0, canonical err 2 : No such file or directory
> svn_client_delete: error deleting t1/A/D/H/omega
> Committing changes in t1.
>
> apr_error: #2, src_err 0, canonical err 2 : No such file or directory
> contents_identical_p: apr_open failed on `./A/B/lambda'
> Updating t2 from changes in t1.
>
> svn_error: #21008 no element found at line 15
> gmake[2]: Leaving directory `/usr/home/mab/subversion-stage/subversion/subversion/client/tests'
Karl and I accidentally broke the milestone 1 test (which lives in
client/tests); that's what you're seeing up there. We're gutting
some major code bits right now; this will be fixed soon. :)
Re: Should `make check' work?
Posted by Karl Fogel <kf...@galois.collab.net>.
matt@braithwaite.net (Matthew Braithwaite) writes:
> Actually the build has been running continuously since Sunday. It
> just hasn't had anything to complain about yet. Sorry, I suppose I
> never announced this.
Matt, that's wonderful! You rock.
(And not announcing it is perfectly in the Unix spirit. After all,
there were no errors, so why should you say anything? :-) )
Re: Should `make check' work?
Posted by Matthew Braithwaite <ma...@braithwaite.net>.
On 02 Nov 2000 08:22:10 -0600, Karl Fogel <kf...@galois.collab.net> said:
>
> In general, people have been running "make check" in their own
> subdirs, not from the top level. I guess once Matt has the
> autobuild stuff running, we should be thinking of "make check" from
> the top level.
Actually the build has been running continuously since Sunday. It
just hasn't had anything to complain about yet. Sorry, I suppose I
never announced this.
Re: Should `make check' work?
Posted by Karl Fogel <kf...@galois.collab.net>.
"make check" in the client directory is broken right now, I broke it
(knowingly, due to work in progress) a couple of days ago.
Yes, keeping "make check" working will be a goal just like "make sure
it builds", once the autobuilds are running. If someone has to commit
with a broken "make check" in some subdir, the thing to do is tweak
the "check" rule to not run the broken code until it's ready. But I
didn't do that this time, sorry. I should have it working again today.
I don't know about the ra_dav tests.
In general, people have been running "make check" in their own
subdirs, not from the top level. I guess once Matt has the autobuild
stuff running, we should be thinking of "make check" from the top
level.
-K
matt@braithwaite.net (Matthew Braithwaite) writes:
> Several previous postings have indicated that people have been able to
> run `make check' -- for example Branko's postings about builds on
> various platforms, and Karl's Milestone 1 posting. I have not had
> such luck.
>
> I wanted to incorporate `make check' into the automatic builds, but
> since it gave me some problems (starting from a clean checkout and
> build), I thought it'd be worthwhile to ask whether it's *supposed* to
> work, and whether spamming dev or individual developers when the tests
> fail is desirable at this time.
>
> The two tests that failed were:
>
> gmake[2]: Entering directory `/usr/home/mab/subversion-stage/subversion/subversion/libsvn_ra_dav/tests'
>
> Running all sub-tests in ra-dav-test...usage: ./ra-dav-test REPOSITORY_URL TARGET_DIR
>
> at least one sub-test FAILED, check tests.log:
>
> gmake[2]: *** [check] Error 1
>
> and
>
> gmake[2]: Entering directory `/usr/home/mab/subversion-stage/subversion/subversion/client/tests'
> sh ./svn-test.sh
>
> Checking out t1.
> A t1/iota
> A t1/A
> A t1/A/mu
> A t1/A/B
> A t1/A/B/lambda
> A t1/A/B/E
> A t1/A/B/E/alpha
> A t1/A/B/E/beta
> A t1/A/B/F
> A t1/A/C
> A t1/A/D
> A t1/A/D/gamma
> A t1/A/D/G
> A t1/A/D/G/pi
> A t1/A/D/G/rho
> A t1/A/D/G/tau
> A t1/A/D/H
> A t1/A/D/H/chi
> A t1/A/D/H/psi
> A t1/A/D/H/omega
> Modifying t1/A/D/G/pi.
> Modifying t1/A/mu.
> Adding t1/newfile1.
> This is added file newfile1.
> Adding t1/A/B/E/newfile2.
> This is added file newfile2.
> Deleting versioned file A/D/H/omega, with --force.
>
> apr_error: #2, src_err 0, canonical err 2 : No such file or directory
> svn_client_delete: error deleting t1/A/D/H/omega
> Committing changes in t1.
>
> apr_error: #2, src_err 0, canonical err 2 : No such file or directory
> contents_identical_p: apr_open failed on `./A/B/lambda'
> Updating t2 from changes in t1.
>
> svn_error: #21008 no element found at line 15
> gmake[2]: Leaving directory `/usr/home/mab/subversion-stage/subversion/subversion/client/tests'