On Wed, Feb 28, 2007 at 12:23:18PM +0000, Philip Hazel wrote:
> On Tue, 27 Feb 2007, Daniel Richard G. wrote:
>
> > Round three of The Patch(tm). This one is uuencoded for your
> > whitespace-preserving pleasure.
> But I won't do this until I've received an "OK" from both Bob and
> Daniel, to both of whom I owe a big debt of thanks for taking on this
> extensive restructuring work.
I don't think you owe us anything. :) My contribution is a very small
thanks in response to you giving me a great regular expression library!
Daniel, after I applied your patch I configured pcre with out any
special options. I ran 'make check' and got,
--- testtry 2007-02-28 10:34:45.000000000 -0500
+++ ./testdata/grepoutput 2007-02-28 10:34:04.000000000 -0500
@@ -342,8 +342,8 @@
./testdata/grepinputx
RC=0
---------------------------- Test 36 -----------------------------
-./testdata/grepinput8
./testdata/grepinputx
+./testdata/grepinput8
RC=0
---------------------------- Test 37 -----------------------------
aaaaa0
FAIL: RunGrepTest
===================
1 of 5 tests failed
===================
make[1]: *** [check-TESTS] Error 1
make[1]: Leaving directory `/home/bob/rcs/svn/pcre/builds/new-default/build'
make: *** [check-am] Error 2
This was working before your changes. Do you have any idea what would
have caused this?
Here is part of the change I got from you,
Index: RunGrepTest.in
+# Set the C locale, so that sort(1) behaves predictably.
+LC_ALL=C
+export LC_ALL
+
I suppose this is fine, I guess you said it helps you pass the test
suite.
@@ -10,7 +14,7 @@
# itself. What we are checking here is the file handling and options that are
# supported by pcregrep.
-cf=diff
+cf="diff -u"
Is this also a change you wanted to make to get the test suite passing?
@@ -19,7 +23,7 @@
while [ $# -gt 0 ] ; do
case $1 in
- valgrind) valgrind="valgrind -q --leak-check=no";;
+ valgrind) valgrind="valgrind -q --leak-check=no";;
Why did you change the above line?
@@ -127,19 +131,19 @@
$valgrind ./pcregrep -L 'fox' $testdata/grepinput $testdata/grepinputx >>testtry
echo "---------------------------- Test 33 -----------------------------" >>testtry
-$valgrind ./pcregrep 'fox' $testdata/grepnonexist >>testtry 2>&1
+$valgrind ./pcregrep 'fox' $testdata/grepnonexist >>testtry 2>&1
Same here, and there are many others. Why change this in this patch?
-$valgrind ./pcregrep --newline=cr -F "def
jkl" $testdata/grepinputx >>testtry
+pattern=`printf 'def\rjkl'`
+$valgrind ./pcregrep --newline=cr -F "$pattern" $testdata/grepinputx >>testtry
Now, did you mean to make this change? Was the ^M in there on purpose?
Same for the others in the patch.
-
+
Why make a change like the one above?
These change effect the diff'ing that I'm doing. I have to look at each
configuration, and visibly confirm that these changes were intentional.
It takes me a long time. That's one reason I wanted to keep the changes
as minimal as possible. I'm not sure if these were because of your
editor, or if you made them on purpose.
Anyways, I figured something from above might have helped shed some
light on why the test suite was failing. Did it pass for you?
Thanks,
Bob Rossi
