From 7e1d0d87ea2112f66f63f560f750e8b566d13f82 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Peter=20Ljunglo=CC=88f?= <peter.ljunglof@heatherleaf.se>
Date: Thu, 22 Aug 2019 11:32:35 +0200
Subject: [PATCH] Updated README

---
 unittest/README.md | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/unittest/README.md b/unittest/README.md
index fdfa5e80e..430694489 100644
--- a/unittest/README.md
+++ b/unittest/README.md
@@ -3,7 +3,7 @@
 ## Usage
 
 ```
-python path-to-script.py path/to/testfile.gftest (...)
+python path/to/unittest.py [-h] [-v] [--no-pmcfg] path/to/testfile.gftest (...)
 ```
 
 The script must be located in a sibling directory
@@ -26,3 +26,10 @@ Every line starts with a language, followed by ":" and the sentence
 (or the abstract syntax tree if the abstract grammar is specified).
 
 You can also see an example in the file [`unittest-example.gftest`](unittest-example.gftest).
+
+## No PMCFG
+
+If your grammar is complex and takes long time to compile, you can try 
+the option `--no-pmcfg`, which tells GF to not build the parsing grammar.
+
+Note however that in this case, every test case needs to contain a parse tree.