On 29.07.2013 05:49, Daniel Shahaf wrote:
> Will this patch break anything? I assume it might break running
> davautocheck against an installed svn/apxs pair --- i.e., running
> basic_tests.py against /usr/local/bin/apxs and /usr/local/bin/svn, out
> of a working copy (or tarball) that not even 'configure' had been run in.
>
> Thoughts?
>
> [[[
> Followup to r1507889:
>
> * subversion/tests/cmdline/davautocheck.sh
> (APXS): Stop guessing where APXS might be. configure will have found
> it for us.
> ]]]
>
> Index: subversion/tests/cmdline/davautocheck.sh
> ===================================================================
> --- subversion/tests/cmdline/davautocheck.sh (revision 1507891)
> +++ subversion/tests/cmdline/davautocheck.sh (working copy)
> @@ -190,10 +190,6 @@ if [ ${APXS:+set} ]; then
> elif APXS=$(grep '^APXS' $ABS_BUILDDIR/Makefile | sed 's/^APXS *= *//') && \
> [ -n "$APXS" ]; then
> :
> -elif APXS=$(which apxs); then
> - :
> -elif APXS=$(which apxs2); then
> - :
> else
> fail "neither apxs or apxs2 found - required to run davautocheck"
> fi
As far as I can see, the code just before the deleted hunk already
checks for whatever APXS configure found -- and skips the bits you
deleted if configure actually did find it. So I think you'r patch would
introduce a bug for exactly the reasons you stated.
-- Brane
--
Branko Čibej | Director of Subversion
WANdisco // Non-Stop Data
e. br...@wandisco.com
