On 12/26/2011 11:26 PM, Jim Meyering wrote:
> FAIL: cond29
> ============
>
> Running from installcheck: no
> Using TAP: no
> PATH =
> /h/j/w/co/automake/tests:/opt/cov-sa-linux64-5.4.0/bin:/usr/lib64/ccache:/h/j/bin/perl:/h/j/bin:/p/p/git/bin:/p/bin:/sbin:/usr/sbin:/bin:/usr/bin:/usr/local:/usr/local/bin
> ++ pwd
> /h/j/w/co/automake/tests/cond29.dir
> + echo AC_PROG_CC
> + cat
> + ulimit -v 1
> + sh -c :
> ./cond29.test: line 35: /bin/sh: Argument list too long
> + ulimit -v 20000
> + sh -c :
> + ulimit -v 20000
> + for i in 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22
> + unindent
> + test x = x
> ++ printf '%s\n' '
> /^$/b # Nothing to do for empty lines.
> x # Get x<indent> into pattern space.
> /^$/{ # No prior x<indent>, go prepare it.
> g # Copy this 1st non-blank line into pattern space.
> s/^\([ ]*\).*/x\1/ # Prepare x<indent> in pattern space.
> } # Now: x<indent> in pattern and <line> in hold.
> G # Build x<indent>\n<line> in pattern space, and
> h # duplicate it into hold space.
> s/\n.*$// # Restore x<indent> in pattern space, and
> x # exchange with the above duplicate in hold space.
> s/^x\(.*\)\n\1// # Remove leading <indent> from <line>.
> s/^x.*\n// # Restore <line> when there is no leading <indent>.
> '
> ++ sed -e 's/ *# .*//'
> + sed_unindent_prog='
> /^$/b
> x
> /^$/{
> g
> s/^\([ ]*\).*/x\1/
> }
> G
> h
> s/\n.*$//
> x
> s/^x\(.*\)\n\1//
> s/^x.*\n//'
> + sed '
> /^$/b
> x
> /^$/{
> g
> s/^\([ ]*\).*/x\1/
> }
> G
> h
> s/\n.*$//
> x
> s/^x\(.*\)\n\1//
> s/^x.*\n//'
> + echo 'AM_CONDITIONAL([C01], [:])'
> + for i in 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22
>
[SNIP repeated traces]
>
> + for i in 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22
> + unindent
> + test 'x
> /^$/b
> x
> /^$/{
> g
> s/^\([ ]*\).*/x\1/
> }
> G
> h
> s/\n.*$//
> x
> s/^x\(.*\)\n\1//
> s/^x.*\n//' = x
> + sed '
> /^$/b
> x
> /^$/{
> g
> s/^\([ ]*\).*/x\1/
> }
> G
> h
> s/\n.*$//
> x
> s/^x\(.*\)\n\1//
> s/^x.*\n//'
> + echo 'AM_CONDITIONAL([C22], [:])'
> + aclocal-1.11a -Werror
> /usr/bin/perl: error while loading shared libraries:
> libc.so.6: failed to map segment from shared object:
> Cannot allocate memory
>
Ah-ah, memory ulimit contraints imposed by ulimit above:
ulimit -v 20000
seems to be too tight. I can reproduce a similar failure on the Fedora
16 system (on pcc64 machine) at the GCC compile farm, and fix it with
the attached patch. Does that patch works for you as well?
> + exit_status=127
> + set +e
> + cd /h/j/w/co/automake/tests
> + test no = yes
> + case $am_explicit_skips in
> + test 127 -eq 0
> + keep_testdirs=yes
> + am_keeping_testdirs
> + case $keep_testdirs in
> + return 0
> + set +x
> cond29: exit 127
>
Thanks,
Stefano
>From 073711d00759683895805b31fd6dad76d2c26aec Mon Sep 17 00:00:00 2001
Message-Id: <073711d00759683895805b31fd6dad76d2c26aec.1324978874.git.stefano.lattar...@gmail.com>
From: Stefano Lattarini <stefano.lattar...@gmail.com>
Date: Tue, 27 Dec 2011 10:40:57 +0100
Subject: [PATCH] tests: fix spurious failure of cond29.test
* tests/cond29.test: Limit the amount of virtual memory available
to the automake process to ~ 50 MB, rather than only ~ 20 MB, to
account for higher (but still acceptable) memory usages on some
systems (in this case, a Fedora 16 distro on ppc64). To be
sure not to reduce coverage, increase the number of potential
combinations of automake conditionals from 2**22 = 4194304 to
2**24 = 16777216.
Reported by Jim Meyering in automake bug#10374.
---
ChangeLog | 12 ++++++++++++
tests/cond29.test | 14 +++++++++-----
2 files changed, 21 insertions(+), 5 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index dfe666a..8f332ad 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,15 @@
+2011-12-27 Stefano Lattarini <stefano.lattar...@gmail.com>
+
+ tests: fix spurious failure of cond29.test
+ * tests/cond29.test: Limit the amount of virtual memory available
+ to the automake process to ~ 50 MB, rather than only ~ 20 MB, to
+ account for higher (but still acceptable) memory usages on some
+ systems (in this case, a Fedora 16 distro on ppc64). To be
+ sure not to reduce coverage, increase the number of potential
+ combinations of automake conditionals from 2**22 = 4194304 to
+ 2**24 = 16777216.
+ Reported by Jim Meyering in automake bug#10374.
+
2011-12-24 Stefano Lattarini <stefano.lattar...@gmail.com>
gitignore: use only one .gitignore file, in the top-level directory
diff --git a/tests/cond29.test b/tests/cond29.test
index 3361081..3971c08 100755
--- a/tests/cond29.test
+++ b/tests/cond29.test
@@ -27,16 +27,20 @@ a02_LDADD =
a03_SOURCES =
EOF
-# On this test, Automake 1.7.x would compute all 2**22 = 4194304
+# On this test, Automake 1.7.x would compute all 2**24 = 16777216
# possible combinations of conditionals (it would do this five times,
# to define a01_DEPENDENCIES, a02_DEPENDENCIES, a03_OBJECTS,
# a04_OBJECTS, and to rewrite bin_PROGRAM), eating all memory, swap,
# or cpu time it can found.
(ulimit -v 1; sh -c ":") && skip_ "no adequate 'ulimit' builtin found"
-(ulimit -v 20000; sh -c ":") || skip_ "no adequate 'ulimit' builtin found"
-ulimit -v 20000
-
-for i in 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22; do
+(ulimit -v 50000; sh -c ":") || skip_ "no adequate 'ulimit' builtin found"
+ulimit -v 50000
+
+for i in \
+ 01 02 03 04 05 06 07 08 09 10 \
+ 11 12 13 14 15 16 17 18 19 20 \
+ 21 22 23 24
+do
unindent >>Makefile.am <<EOF
if C$i
bin_PROGRAMS += a$i
--
1.7.7.3
