Mailing List Archive: exim: dev

Possible runtest fix?

Index | Next | Previous | View Threaded

tlyons at ivenue

May 22, 2012, 9:05 AM

Post #1 of 3 (174 views)
Permalink

Possible runtest fix?

I have Ubuntu and it uses a version of 'ls' that causes test 345 to
fail. Is the following addition to runtest too hackish or is it ok to
commit? It will fix Ignore the probable line wrapping at the end. I
attempted to make it reuse as much of the values as possible (year and
date) so that it would be somewhat flexible should anything in the
date setting process change in the future, and if the time is off
(normally it would be 00:00).

diff --git a/test/runtest b/test/runtest
index dac0277..e4951a6 100755
--- a/test/runtest
+++ b/test/runtest
@@ -450,6 +450,9 @@ RESET_AFTER_EXTRA_LINE_READ:
# Date of message arrival in spool file as shown by -Mvh
s/^\d{9,10}\s0$/ddddddddd 0/;

+ # Date of mailbox files, non-standard format
+ s/ 0 (2002)-05-(10) \d\d:\d\d / 0 May $2 $1 /;
+
# Date/time in mbx mailbox files
s/\d\d-\w\w\w-\d\d\d\d\s\d\d:\d\d:\d\d\s[-+]\d\d\d\d,/06-Sep-1999
15:52:48 +0100,/gx;

...Todd
--
Always code as if the guy who ends up maintaining your code will be a
violent psychopath who knows where you live. -- Martin Golding

--
## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##

pdp at exim

May 22, 2012, 5:00 PM

Post #2 of 3 (162 views)
Permalink

Re: Possible runtest fix? [In reply to]

On 2012-05-22 at 09:05 -0700, Todd Lyons wrote:
> I have Ubuntu and it uses a version of 'ls' that causes test 345 to
> fail. Is the following addition to runtest too hackish or is it ok to
> commit?

Does setting LC_ALL=C in the environment before invoking runtest fix
things?

I wonder if runtest should simply do:
$ENV{'LC_ALL'} = 'C';
at startup?

-Phil

--
## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##

tlyons at ivenue

May 23, 2012, 6:34 AM

Post #3 of 3 (162 views)
Permalink

Re: Possible runtest fix? [In reply to]

On Tue, May 22, 2012 at 5:00 PM, Phil Pennock <pdp [at] exim> wrote:
> On 2012-05-22 at 09:05 -0700, Todd Lyons wrote:
>> I have Ubuntu and it uses a version of 'ls' that causes test 345 to
>> fail. Is the following addition to runtest too hackish or is it ok to
>> commit?
> Does setting LC_ALL=C in the environment before invoking runtest fix
> things?

Yes.

> I wonder if runtest should simply do:
> $ENV{'LC_ALL'} = 'C';
> at startup?

Applied locally to my runtest script, tested, and confirmed it fixes
test 345, so I ran tests 1 - 565. Now test 345 passes, and only the
same 4 tests as normal fail (152, 358, 373, and 548). I can find no
tests that this causes a failure in.

That was a great catch, Phil. You can see it quite plainly:

[exim-build [at] tlyon ~/projects/exim/test (master)]$ vdir test-mail/exim-build
-rw------- 1 exim-build exim-build 0 2012-05-23 06:30 test-mail/exim-build
[exim-build [at] tlyon ~/projects/exim/test (master)]$ LC_ALL=C vdir
test-mail/exim-build
-rw------- 1 exim-build exim-build 0 May 23 06:30 test-mail/exim-build

I'll commit and push.

...Todd

--
Always code as if the guy who ends up maintaining your code will be a
violent psychopath who knows where you live. -- Martin Golding

--
## List details at https://lists.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads