On Dec 9, 2008, at 3:57 AM, Malcolm J. Currie wrote:
>> I can't help but feel this would be a lot easier if fitsmod simply
>> called KPG1_GTFTS, AST_SETFITSU, KPG1_PTFTS. ie we just did it all
>> in a fitschan
>
> Who dat?
Who "we"? General inclusive expression. ;-) The alternative is to
remove FTS completely and use ATL instead but I'm sure that is a
little bit too much work.
>
>
>> that can already handle everything fitsmod does (except maybe the
>> POSITION parameter)...
>>
>> All the linked list stuff in fts1_edkey is making my head hurt.
>
> It handles the movement of cards (e.g. POSITION) so that you only
> deal in order once after the shuffle. This Fortran predates AST
> and structures in Fortran,
>
> Did you want me to add that new option? Have you decided upon a name?
Clear or Null was fine. Did not like Blank.
Spent another half hour looking at this yesterday but my brain started
hurting again. I kept wondering whether I was supposed to be tweaking
the "what type is this" routine to recognize UNDEF as a type and which
logic to change for storing an undef. I half came to the conclusion
that a hack to call UNDEF mode an UPDATE of a Double with VAL__BADD
was the quickest option. (a bit like the atlPtftx routines).
if you wouldn't mind putting something together and if it's not a lot
of work then that would be great.
--
Tim Jenness
Joint Astronomy Centre
|