Surely Brian would know how to close and reopen the temp file system? Does hdsShow list many long lived temp files? I wonder why they never get cleaned up.
On Jul 21, 2010, at 5:47 AM, Malcolm J. Currie wrote:
>> Sorry Malcolm, you're in uncharted territory as far as I am concerned.
>> Some experimentation required I think.
>
> If we had an up-to-date internals manual for HDS I might know where to start looking. As I said to Brian, I'm quite happy to answer some of the questions empirically, if I knew how to access the top-level locator for a monolith's scratch container file.
>
> I could create a temporary file near the start of the monolith routine, get its locator, CALL DAT_PAREN, annul the workspace, and then at the end of the monolith call HDS_ERASE on the parent locator. OK, let's try that. If this or some variant works, we should consider including in the Fixed Part.
>
> Malcolm
--
Tim Jenness
Joint Astronomy Centre
|