check tests for minor psLib updates (for rel 11)

[eugene]

these mods were made to psLib in order to sync it with psModules for release 11

• psFitsImage
  • dropped output from psFitsReadImage
  • added Cube functions

• psMetadata
  • added psMetadataItemCopy
  • mods to psMetadataCopy to use psMetadataItemCopy
  • added polynomial to MD header entries
  • added psRegion, List, Time to psMetadataPrint

• psType.h:
  • added PS_DATA_REGION

• psRegion / psRegionForImage : moved from psImage to src/math

• added psRegion includes:

./src/astro/psCoord.h:#include "psRegion.h"
./src/pslib_strict.h:#include "psRegion.h"
./src/pslib_strict.h:#include "psRegionForImage.h"
./src/imageops/psImageStats.c:#include "psRegion.h"
./src/imageops/psImageStats.c:#include "psRegionForImage.h"
./src/imageops/psImageStats.h:#include "psRegion.h"
./src/imageops/psImageStructManip.h:#include "psRegion.h"
./src/math/psRegion.c:#include "psRegion.h"
./src/math/psRegionForImage.c:#include "psRegion.h"
./src/math/psRegionForImage.c:#include "psRegionForImage.h"
./src/types/psPixels.h:#include "psRegion.h"

• psString.h:
  • added psStringSubstitute

[Paul Price]

Updated MAX_STRING_LENGTH in psMetadataConfig.c to 1024.

[Paul Price]

psFitsReadHeader: added code to strip trailing spaces from string

I suspect there may be several problems in psLib now that the 'n' element of
vectors and arrays are not set to nalloc at the time of allocation. I can see
this with psBinaryOp, as follows. If we do:

psVector *vector3 = (psVector*)psBinaryOp(NULL, vector1, "+", vector2);

Then vector3->n == 0, not vector1->n as expected.

I have fixed this behaviour in psBinaryOp and psUnaryOp, but a careful search
should be made for other instances.

[Paul Price]

Found similar problem in psVectorSortIndex.

I suggest searching for all instances of psVectorRecycle and psArrayRecycle and
examining those to make sure that the length is set.

[Paul Price]

I think the tests are up to date following the API changes.