aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorSam Hartman <hartmans@mit.edu>2010-10-19 19:57:56 +0000
committerSam Hartman <hartmans@mit.edu>2010-10-19 19:57:56 +0000
commitedfff4b06ffcb9924885e4790792093349cf2585 (patch)
treed9c4de9b9854e0c09e5c460d4f3d03c6aadd6195
parentb98ac34fe6ef1d00216bdd239d3bd6219e8e6e38 (diff)
downloadkrb5-edfff4b06ffcb9924885e4790792093349cf2585.zip
krb5-edfff4b06ffcb9924885e4790792093349cf2585.tar.gz
krb5-edfff4b06ffcb9924885e4790792093349cf2585.tar.bz2
configure.in: don't force use of cache
Back when hardware was slower and we had lots and lots of imakefile-like configure scripts, forcing the use of config.cache made sense for performance reasons. Now that we have one configure script and we build on modern hardware, config.cache just introduces errors as you are debugging changes to the configure script. This commit stops second guessing autoconf's cache handling. git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@24468 dc483132-0cff-0310-8789-dd5450dbe970
-rw-r--r--src/configure.in7
1 files changed, 0 insertions, 7 deletions
diff --git a/src/configure.in b/src/configure.in
index 8d6bb82..592bdd8 100644
--- a/src/configure.in
+++ b/src/configure.in
@@ -1,11 +1,4 @@
K5_AC_INIT([aclocal.m4])
-#
-# autoconf 2.49 defaults to a /dev/null cache file, which is what we
-# do not want for performance reasons.
-if test "x$cache_file" = "x/dev/null"; then
- cache_file=./config.cache
- AC_CACHE_LOAD
-fi
CONFIG_RULES
KRB5_VERSION=K5_VERSION