Re: Cobol compiler date

> Didn't mean to confuse anyone with this.  IFOX, which came
> with 3.8j, had various PTF updates in the DLIBs without the
> source being updated.  Paul G (?) spent some time on updating
> the IFOX source we had to get it to generate the distributed
> modules.

A side note, while I think of it:  I had a heck of a time
getting a version of IFOX00 that would run on MVT.  Any
modern version will at some point exhibit an error in
which the location counter goes nuts and is suddenly
incremented by a very large amount, such that everything
becomes unaddressable.  Something introduced by some PTF
at some point did something incompatible with MVT.  The
version I've found to work consistently under MVT is the
version supplied in object deck form with VM/370 R6.  That's
the level on the ASP system.

> A side effect of the compiler date zap reminded me that the
> OS/360 libraries might suffer from the same type thing.  Also
> that they were assembled on OS/360 and not on MVS3.8j.  There
> are probably some MAClib changes between the two environments. 
> Whether it would benefit the compilers and run time libraries
> to be generated from source on MVS3.8j is unknown at present.

Mostly what gets assembled in MVT stage 2 for compilers
and libraries are small modules containing default options.
ANS COBOL is an exception; for some reason, the default
options are included as part of a full-size module.
Except for things that get assembled in MVT stage 2, we
have the same problem with MVT that we have with MVS:
load modules with PTFs are newer than the source we've
got.

--