[dev] horde/lib/Test.ph

[steve]

I though about it.
I don't think that $result = file_exists('./' . $key); is a good way to do, 
because you are not sure of the directory (in my case, I am calling horde 
librairies from a complete different directory tree), maybe we should 
grab   horde/config/registry.php, and do the loop with the fileroot . What 
do you think?

At 21:00 18/08/2003, you wrote:
>Message: 7
>Date: Mon, 18 Aug 2003 14:40:43 -0400
>From: John Morrissey <jwm at horde.net>
>Subject: Re: [dev] Re: [cvs] commit: nag tasklists.php
>To: dev at lists.horde.org
>Message-ID: <20030818184043.GA6410 at boost.horde.net>
>Content-Type: text/plain; charset=us-ascii
>
>On Mon, Aug 18, 2003 at 04:31:25PM +0200, Jan Schneider wrote:
> > That didn't change your last patch. The current logic is: "if setting the
> > tasklist description fails, show the error message: the tasklist couldn't
> > be saved." Huh?
>
>I figure you're referring to tasklist.php:85? I'm not sure what you mean -
>what should the code be doing? The existing success message was 'The task
>list has been saved.' To create a corresponding failure message, I used 'The
>tasklist couldn't be saved.' The changes are batched up in the share and
>comitted by the setDescription() call, so either all of it succeeds, or all
>of it fails.
>
>john
>Quoting steve <mailling at gmx.co.uk>:
>| Here is a little patch that makes horde/lib/Test.php wirks a little bit
>| better
>|
>| diff -w -b -r1.9 Test.php
>| 355c355
>| <             $result = file_exists('./' . $key);
>| ---
>|  >             $result = file_exists(HORDE_BASE. '/' . $key);
>
>This would break the application testing; e.g. running test.php for imp
>would check for the existence of './config/foo.php' =
>'horde/imp/config/foo.php' in the current code, but would check for
>'/path/to/horde/config/foo.php' = 'horde/config/foo.php' in your patch
>which is obviously not correct.
>
>michael