doc: describe test control variables

* HACKING (Add tests): Mention the variables and default values.
* README-release (Pre-release testing): Mention that setting
the SHELL variable may be required.
Suggested by Bruno Haible.

2 files changed, 14 insertions, 0 deletions

diff --git a/HACKING b/HACKING
index 51c4acdc8..0adec7911 100644
--- a/HACKING
+++ b/HACKING
@@ -430,6 +430,16 @@ as possible. Note to run tests/misc/new-test in isolation you can do:
 
   (cd tests && make check TESTS=misc/new-test VERBOSE=yes)
 
+Variables that are significant for tests with their default values are:
+
+  VERBOSE=yes
+  RUN_EXPENSIVE_TESTS=no
+  RUN_VERY_EXPENSIVE_TESTS=no
+  SHELL=/bin/sh
+  NON_ROOT_USERNAME=nobody
+  NON_ROOT_GROUP=$(id -g $NON_ROOT_USERNAME)
+  COREUTILS_GROUPS=$(id -G)
+
 There are hundreds of tests in the tests/ directories.  You can use
 tests/sample-test as a template, or one of the various Perl-based ones
 in tests/misc.
diff --git a/README-release b/README-release
index 729ab898b..af2610a8c 100644
--- a/README-release
+++ b/README-release
@@ -34,6 +34,10 @@ FIXME: enable excluded programs like arch? to get their manual pages?
       && make distcheck \
       && make -j$n check RUN_VERY_EXPENSIVE_TESTS=yes RUN_EXPENSIVE_TESTS=yes
 
+  If testing on systems with a non standard default shell, spurious failures
+  may occur.  Often there are other shells available, and you can select
+  those by using for example, SHELL=bash in the commands above.
+
   Note that the use of -j$n tells make to use approximately half of the
   available processing units.  If you use -jN, for larger N, some of the
   expensive tests are likely to interfere with concurrent performance-measuring