On Jan 10, 2011, at 10:18 AM, Andy Gibb wrote:
>
> MAKEMAP: Map-maker will use no more than 57947 MiB of memory
> !! Truncation during copy (destination 132 < source 255) of
> ! '^/export/data/visitors/agibb/reduced/scuba2_850/20091205/fakesrc/dimmcon
> ! fig_bright_extended2.lis,fakemap=s8d20091205_00022_fakemap.sdf'
> ! DAT_PUT: Error writing value(s) to an HDS primitive.
> ! GRP_GROUP: Unable to associate a group of names with parameter CONFIG
> ! Application exit status DAT__TRUNC, Text truncated
> ! in=^in22.lis
> ! config='^/export/data/visitors/agibb/reduced/scuba2_850/20091205/fakesrc/
> ! dimmconfig_bright_extended2.lis,fakemap=s8d20091205_00022_fakemap.sdf'
> ! out=junk method=iter
>
> If I put the config file somewhere with a shorter path then I don't
> get the error.
SUBPAR seems to have a hard-coded limit (in many places) of 132 characters. Going so far as writing _CHAR*132 as the HDSTYPE despite HDS being perfectly capable of being given a dynamic size. Changing every 132 in SUBPAR to 255 would help but that seems to be a bit of a sledge hammer because it seems to use those buffers even for reading numbers. The only short term fix is to make the pipeline write a config file with each entry in separately.
PS I've fixed the error message so that it is accurate about the source length:
$ makemap in=a123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012
!! Truncation during copy (destination 132 < source 133) of
! 'a12345678901234567890123456789012345678901234567890123456789012345678901
! 2345678901234567890123456789012345678901234567890123456789012'
! DAT_PUT: Error writing value(s) to an HDS primitive.
--
Tim Jenness
Joint Astronomy Centre
|