Done.
On 3 May 2016 at 16:26, Tim Jenness <[log in to unmask]> wrote:
> I agree. Just remove that old code that is now irrelevant (and clearly
> causes trouble).
>
> On Tue, May 3, 2016 at 4:26 AM, Peter W. Draper <[log in to unmask]>
> wrote:
>>
>> On Tue, 3 May 2016, David Berry wrote:
>>
>>> The prologue for dtask_main.f says:
>>>
>>> * Tune the SHELL parameter of HDS - Allow environment variable to
>>> * override the ADAM default of C-shell (Note the HDS default is
>>> Bourne
>>> * shell which will not translate ~)
>>>
>>> Is that true? As a life long user of tcsh, simple tests using sh and
>>> bash suggests that "ls ~" works fine in both.
>>
>>
>> Tilde expansion is supported by the POSIX shell, so should work for all
>> modern variants. I'm sure it wasn't part of the earlier UNICES (like early
>> Solaris where we developed much of this), but we should just move on and
>> dump this code...
>>
>>
>>
>>> Tim - I though the purpose of HDS__NOSHELL was as a value that could
>>> be set to indicate that no shell expansion should be done, rather than
>>> a default value that indicates that a new value is required? A test
>>> using the following code:
>>>
>>> program ff
>>> implicit none
>>> integer ival, status
>>> status = 0
>>> call hds_gtune( 'SHELL', ival, status )
>>> write(*,*) ival, status
>>> end
>>>
>>> indicates that hds_gtune is returning zero (i.e. "sh") if HDS_SHELL is
>>> unset (as suggested by the comment in the dtask_main.f prologue), not
>>> -1 (HDS__NOSHELL). So shouldn't dtask_main.f just be changed by
>>> simply removing the block of code Graham pointed out?
>>>
>>> David
>>>
>>
>> --
>> Peter W. Draper, http://astro.dur.ac.uk/~pdraper
>
>
|