tests: make 'f' tool open files in binary mode when hexdumping
On Windows, the various bundle hexdumps were being truncated at the first
occurrence of 0x1a. The line endings in the sequence generation then needs to
be standardized on LF so that the file is the same for the tool tests.
--- a/tests/f Thu Nov 12 14:29:02 2015 -0800
+++ b/tests/f Mon Nov 09 12:49:31 2015 -0500
@@ -44,7 +44,7 @@
if opts.type:
facts.append('file')
if opts.hexdump or opts.dump or opts.md5:
- content = file(f).read()
+ content = file(f, 'rb').read()
elif islink:
if opts.type:
facts.append('link')
--- a/tests/test-tools.t Thu Nov 12 14:29:02 2015 -0800
+++ b/tests/test-tools.t Mon Nov 09 12:49:31 2015 -0500
@@ -43,7 +43,13 @@
foo: mode=644
#endif
+#if no-windows
$ python $TESTDIR/seq.py 10 > bar
+#else
+Convert CRLF -> LF for consistency
+ $ python $TESTDIR/seq.py 10 | sed "s/$//" > bar
+#endif
+
#if unix-permissions symlink
$ chmod +x bar
$ f bar --newer foo --mode --type --size --dump --links --bytes 7