[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
spurious "make check" failure in gzip-1.3.6
From: |
Jim Meyering |
Subject: |
spurious "make check" failure in gzip-1.3.6 |
Date: |
Sat, 25 Nov 2006 10:23:53 +0100 |
Hi Paul,
I ran the usual build/check iteration and was surprised to see
a make check failure:
$ make check >& makerr & tail -f makerr
...
m4/xstrndup.m4: 36.9%
m4/yesno.m4: 34.1%
makerr: 69.5%
cmp: EOF on -
make[2]: *** [check-local] Error 1
make[2]: Leaving directory `/t/gzip-1.3.6'
make[1]: *** [check-am] Error 2
make[1]: Leaving directory `/t/gzip-1.3.6'
make: *** [check-recursive] Error 1
That's because the file "makerr", was growing while the pipeline progressed.
One way to fix this is to use e.g., $(DISTFILES) rather than the
output of find. I've included a patch for that below.
A smaller change would be to remove the "-v" option that's responsible
for making "gzip -c" produce the output during that pipeline.
If you opt to keep the "find" command, you may also want to change the "*"
argument to "." -- unless the use of "*" was deliberate -- but the only
name matching ".??*" is ".deps", and there should be no harm in checking
the files in that directory. I suppose it could be deliberate if the
author presumed that build log file names would always start with ".".
2006-11-25 Jim Meyering <address@hidden>
* Makefile.am (check-local): Use the list of files in $(DISTFILES),
rather than running find on the current directory, in case there are
non-distribution files (e.g., build log) changing during the test.
--- Makefile.am.~1~ 2006-11-25 10:04:32.550257498 +0100
+++ Makefile.am 2006-11-25 10:04:50.449843430 +0100
@@ -60,7 +60,7 @@ SUFFIXES = .in
# A simple test, just of gzip -- more of a sanity check than anything else.
check-local:
- for file in `find * -type f -print | sort`; do \
+ for file in $(DISTFILES); do \
test -r "$$file" || continue; \
./gzip -cv $$file | ./gzip -d | cmp - "$$file" || exit; \
done
- spurious "make check" failure in gzip-1.3.6,
Jim Meyering <=