X-Git-Url: https://git.cworth.org/git?p=apitrace-tests;a=blobdiff_plain;f=cli%2FREADME.markdown;h=4f49822a768006d33fdec56bd6448c1e681870ce;hp=3d49c7c9a355758746de5bb0966eb55a262447f8;hb=8351c1e187d8f65477db05a07bacd5d184c2b950;hpb=283dea13d9305a10cc309eb7b524b514012f4be7

diff --git a/cli/README.markdown b/cli/README.markdown
index 3d49c7c..4f49822 100644
--- a/cli/README.markdown
+++ b/cli/README.markdown
@@ -10,8 +10,10 @@ test in the traces directory. Otherwise, a new test program can be
 written in this directory.
 
 The tests in this directory are found in files with names matching
-*.script. The scripts are simple line-based commands with the
-following meanings based on the first word of each line:
+*.script by convention. The scripts must be listed explicitly in the
+CMakeLists.txt file. Each script consists of simple line-based
+commands with the following meanings (based on the first word of each
+line):
 
   apitrace:     Execute the current apitrace executable being tested
   		with the given arguments. If apitrace returns a
@@ -26,6 +28,12 @@ following meanings based on the first word of each line:
                 interpreted locally. If this fails for any reason
                 other than "file does not exist" the test will fail.
 
-If none of the commands in the script cause the test to fail, then the
-test will pass.
+Note: Blank lines and lines beginning with '#' are ignored.
+
+Commands can be prefixed with "EXPECT_FAILURE:" to indicate that a
+command is expected to return a non-zero value. In this case, a return
+value of zero from the command will cause the test to fail.
+
+If none of the commands in the script cause the test to fail (as
+described above), then the test will pass.