I have  some old (and some still current) BP7 code that is being and
will be migrated to other OS', and I have some new work to be done in
some of other these OS'.  I need a suitable migration path.  I try not
to use too many Borland specific extensions, but of course have used
some (wouldn't want to live w/o objects for instance).  I don't mind
writing some routines to replace some extensions: e.g. BlockRead, and I
don't mind coding some fixups on small differences.  I prefer not to
re-write and I prefer not to move to another language (C/C++/Java, don't
suggest it, please).  In addition I prefer not to revert to debuging via
low level de{*word*81}s and writeln statements, I've come to rely on
windowed, stepping de{*word*81}s. I know some of the answers to this
question, what I need are the missing bits.  I need to do some work in
Win95, NT and Unix (I have just setup a Linux station to do my work).  I
know and, already have, Delphi for Win95 and NT, but am not finding much
luck with a complete and ready to use solution for Unix.  I know there
are a few things out there: GPC, and FPK, but neither seems to offer a
de{*word*81} and both admit to being not quite ready for prime time.  Does
anyone know of anything?  Even OS/2 has options (SpeedPascal/Sibyl).
Has not the fertile Linux environment come up with something?  FWIW, if
I must re code I would prefer to move to something like Modula2/3 or
Oberon, but again it must support Dos, Win95, NT and Unix and be
complete in the aspects I have mentioned.  Any chance of my finding
happiness?

Many thanks,

In <345B7161.4...@ibm.net>,

Anyway, I personally prefer a compiler (and OS) that's free and actively
developed, though there might be a few bugs, because I have better chances
not only to get the bugs fixed, but also to get new features that I want.
Sometimes, I can even do these things to the compiler myself. With
proprietary and/or obsolete software, I don't have these chances. That's my
experience with GPC vs. BP, and with Linux vs. Dos.