Corruption of string variable in IDL V5.3

[landsman(at)my-deja.com]

I like the upgrade of the STRMID function in IDL V5.3 so that it can now
handle vector arguments.   But the following code shows how in certain
cases, STRMID can also now corrupt a string variable:

IDL >print,!version
{ sparc sunos unix 5.3 Nov 11 1999}

     st = 'idl40\avg.pro'
     g =  intarr(1)
     st[0] = strmid(st[g], 6, 7)

;     print,'st = ',st,byte(st)

     name = st
     name[0] = strmid(st[g],0,3 )

     print,'st = ',st,byte(st)

If the first print statement is uncommented, then the output is as
expected:

st = avg.pro  97 118 103  46 112 114 111
st = avg.pro  97 118 103  46 112 114 111

But if the first print statement is left as a comment, then a byte(0)
gets inserted into the string, st

st = avgt,'  97 118 103   0 112 114 111

(You can imagine my frustration during debugging, not realizing that a
diagnostic print statement was changing the results.)

--Wayne Landsman                       landsman@mpb.gsfc.nasa.gov

P.S. A related wish for IDL V5.4 is that the STR_SEP function allow
vector arguments -- or at the very least, that it be rewritten
internally to use the vector features of STRMID.


Sent via Deja.com http://www.deja.com/
Before you buy.