# HG changeset patch
# User Bryan O'Sullivan <bos@serpentine.com>
# Date 1174971504 25200
# Node ID 7f07aca44938d38b30ae8713946346123cdf97b6
# Parent  745ff473c8c466c4e09d2594b1a837691a847293
Write up the unpleasant effects of change ef1f1a4b2efb in the hg tree.

diff -r 745ff473c8c4 -r 7f07aca44938 en/examples/filenames.glob.range.out
--- a/en/examples/filenames.glob.range.out	Mon Mar 26 21:25:34 2007 -0700
+++ b/en/examples/filenames.glob.range.out	Mon Mar 26 21:58:24 2007 -0700
@@ -1,3 +1,7 @@
 $ \textbf{hg status 'glob:**[nr-t]'}
+A examples/simple.py
 ? MANIFEST.in
+? examples/performant.py
+? src/watcher/_watcher.c
+? src/watcher/watcher.py
 ? src/xyzzy.txt
diff -r 745ff473c8c4 -r 7f07aca44938 en/filenames.tex
--- a/en/filenames.tex	Mon Mar 26 21:25:34 2007 -0700
+++ b/en/filenames.tex	Mon Mar 26 21:58:24 2007 -0700
@@ -152,6 +152,15 @@
 token.  This small example illustrates the difference between the two.
 \interaction{filenames.glob.star-starstar}
 
+When you're writing a glob pattern, bear in mind that Mercurial will
+treat a pattern that matches a directory name as ``match every file
+under that directory''.  For example, a glob pattern of
+``\texttt{**c}'' means \emph{both} ``match files ending in
+`\texttt{c}''' ``any file under all directories that end in
+`\texttt{c}'''.  I personally find this behaviour counterintuitive.
+If you need to write a pattern that means ``match \emph{only} files'',
+you'll need to express it as a regular expression instead; see below.
+
 \subsection{Regular expression matching with \texttt{re} patterns}
 
 Mercurial accepts the same regular expression syntax as the Python