summaryrefslogtreecommitdiffstats
path: root/t/README
diff options
context:
space:
mode:
Diffstat (limited to 't/README')
-rw-r--r--t/README29
1 files changed, 29 insertions, 0 deletions
diff --git a/t/README b/t/README
index ea8fc8f3..9ca9e1c0 100644
--- a/t/README
+++ b/t/README
@@ -1,4 +1,33 @@
+The tests are running the following way:
+
+01compile.t ensures everything is valid perl
+02create_pkgs.t builds the various RPMs needed by the others tests
+cfg.t tests reading/writing the config files
+pod.t check for POD errors in files (optional)
+superuser--*.t
+
+The first tests just ensure that basic things in urpmi works.
+
you need to be root to run superuser--*.t,
with a normal user those tests are simply skipped.
+They actually test both urpmi, URPM _and_ *rpm* extensively.
+It had often found regressions in rpm itself that rpm's own testsuite didn't found.
+
+superuser--file-conflicts.t runs only some patches on Mageia due to having a
+downstream patch there
+
+superuser--mirrorlist.t will only be run on Mageia as it needs a specific patch
+that introduces Time::ZoneInfo->current_zone()
+
+Usage:
+------
+
use "make testall"
+
+You can speed up things by using eatmydata which will prevent all the fsync()
+& fdatasync() calls in librpm:
+eatmydata make testall
+
+If you didn't alter the tests data, you can further speedup a re-run by using:
+[[ -d t/media ]] && mv -f t/02create_pkgs.t{,.i}