summaryrefslogtreecommitdiff
path: root/README-release
diff options
context:
space:
mode:
Diffstat (limited to 'README-release')
-rw-r--r--README-release26
1 files changed, 17 insertions, 9 deletions
diff --git a/README-release b/README-release
index 7a73c9dcd..465bedab6 100644
--- a/README-release
+++ b/README-release
@@ -13,21 +13,29 @@ Here are most of the steps we (maintainers) follow when making a release.
* Ensure that you're on "master" with no uncommitted diffs.
This should produce no output: git checkout master; git diff
-* Run bootstrap: ./bootstrap
+* Run bootstrap one last time. This downloads any new translations:
+
+ ./bootstrap
FIXME: enable excluded programs like arch? to get their manual pages?
* Pre-release testing:
- Ensure that make check syntax-check succeeds.
- On at least one SELinux-enabled (enforcing) and one non-SELinux system,
- run all tests, both root-only and regular.
- Run *all* non-root tests, including expensive and very expensive ones i.e.,
- run this: make -j1 check RUN_VERY_EXPENSIVE_TESTS=yes RUN_EXPENSIVE_TESTS=yes
- Run the root-only tests:
- sudo env PATH="$PATH" NON_ROOT_USERNAME=$USER make -k check-root
+ Run the following on at least one SELinux-enabled (enforcing) and
+ one non-SELinux system:
+
+ make distcheck
+ make -j1 check RUN_VERY_EXPENSIVE_TESTS=yes RUN_EXPENSIVE_TESTS=yes
+ sudo env PATH="$PATH" NON_ROOT_USERNAME=$USER make -k check-root
+
+ Note the -j1 above. If you use -jN, for larger N, some of the expensive
+ tests are likely to interfere with concurrent performance-measuring or
+ timing-sensitive tests, resulting in spurious failures.
+
+ If "make distcheck" doesn't run "make syntax-check" for you, then run
+ it manually:
-* Run "make distcheck"
+ make syntax-check
* Set the date, version number, and release type [stable/alpha/beta] on
line 3 of NEWS, commit that, and tag the release by running e.g.,