On Dec 17, 2007, at 10:01 AM, Jamie Leech wrote:
>>
>> We have a solaris box and I think Peter has one too.
>>
>> Writing the dump file out as a .sdf is clearly the best thing to
>> do. You would need to have a scheme where specx notices the old
>> fashioned format, reads that but then writes out the new dump file
>> as a simple HDS file. Wouldn't need to be an NDF. Something like
>> the ADAM parameter files (in ~/adam) would be more than enough).
>> This would also allow HDSTRACE to work on the file
>> to list the contents. You would need to first try $SPECX_DUMP as an
>> old file and then try again dropping the .dmp and trying it as a
>> HDS file. In principal you could retain the .dmp suffix but that
>> might get confusing.
>>
>> I'm not planning on spending time on this though :-)
>
> I've done the brute force fix of getting someone with a login to a
> Cambridge Solaris machine to reload all of the .dmp files and save
> them as .sdf files. This seems to have worked.
>
> I assume all of the T_sys info goes into the .sdf when you do this,
> so further coadding works properly. From what I remember SPECX .sdfs
> do do this correctly.
>
Hang on. I'm very confused. I don't think you've been talking about
the same problem I've been talking about. How do you save a .dmp file
as .sdf? What's TSYS got to do with it. What's your input file and
what's your command to write the output? What does hdstrace say for
the output?
--
Tim Jenness
Joint Astronomy Centre
|