On Tue, 2005-12-06 at 08:59 -0500, Robert Story wrote:
> On Tue, 06 Dec 2005 09:16:19 + Dave wrote:
> DS> Having done a bit of work on 'get_exec_output', I don't
> DS> think that should be too difficult. But it's definitely
> DS> something for 5.4ff, rather than trying to squeeze it in now!
On Tue, 06 Dec 2005 09:16:19 + Dave wrote:
DS> Having done a bit of work on 'get_exec_output', I don't
DS> think that should be too difficult. But it's definitely
DS> something for 5.4ff, rather than trying to squeeze it in now!
So what do you thing is reasonable to fit into 5.3?If there i
On Mon, 2005-12-05 at 20:07 -0800, Wes Hardaker wrote:
> Dave> (Which then also
> Dave> raises the question of whether "pass" output should be
> Dave> cached or not).
>
> Caching... sigh... needed and a pain at the same time.
Yup.
> I think caching should
> On Fri, 02 Dec 2005 16:34:17 +, Dave Shield <[EMAIL PROTECTED]> said:
Dave> But that's definitely something for a later date.
Dave> As far as 5.3 is concerned, I'd just be looking at
Dave> making 'get_exec_output' invoke 'run_exec_command',
Dave> and keep the same cache behaviour. (Whi
Dave Shield wrote:
On Fri, 2005-12-02 at 17:14 +0100, Thomas Anders wrote:
Are there any drawbacks (=change in behaviour other than
bug fixing) of this approach?
OK - there's one other drawback that I've only just twigged.
The 'get_exec_output' routine directs the output from the
external c
Dave Shield wrote:
I would like to:
a) Convert these remaining directives to
use 'run_exec_command' directly, and
b) Convert 'get_exec_output' to be a simple
wrapper round 'run_exec_command' (to catch
any other private uses of this routine)
But I'm not sure
On Fri, 2005-12-02 at 17:14 +0100, Thomas Anders wrote:
> Are there any drawbacks (=change in behaviour other than
> bug fixing) of this approach?
OK - there's one other drawback that I've only just twigged.
The 'get_exec_output' routine directs the output from the
external command to a temporar
On Fri, 2005-12-02 at 17:14 +0100, Thomas Anders wrote:
> Sounds fine. Are there any drawbacks (=change in behaviour
> other than bug fixing) of this approach?
I don't think so - assuming everything works correctly,
of course!The most likely area would be caching.
The 'run_exec_command' routi
Dave Shield wrote:
I would like to:
a) Convert these remaining directives to
use 'run_exec_command' directly, and
b) Convert 'get_exec_output' to be a simple
wrapper round 'run_exec_command' (to catch
any other private uses of this routine)
But I'm not sure w
I've just applied a patch to 'utilities/execute.c' so that
it uses the library routine 'copy_nword' when tokenizing
commands for execution, rather than handling this internally.
This means that it'll parse quoted strings in the natural
manner, rather than breaking on internal spaces.
The reason f
10 matches
Mail list logo
