ESC1111@DDAESA10.BITNET (02/22/86)
VMS Condition Handlers - do they work?
We are just starting to design a moderately complex
interactive cum real-time application to run under VMS 4.2, to
be written in FORTRAN (why FORTRAN? - ask me another!).I
have been reading all sorts of VMS books and am very happy
to discover the Condition Handler support - it seems to me
that this should greatly help the production of reasonably
robust software.
The only problem is that one of my colleagues was told (in a
previous project) to avoid condition handlers like the
plague 'cos they didn't work safely. He's not sure if this
applied only to their use in the particular language (CORAL)
and its run-time library or if this was a general caveat.
Anyone got any bad (or good come to that) experiences with
condition handlers, especially in combination with FORTRAN?
Do you think they are a help or a hindrance to producing
good systems?
------------ standard denial of everything ------
any views expressed are purely my own and may not reflect those
of my employer - the European Space Agency
N. Head, ESC1111@DDAESA10, ESC1111%ESOC.BITNET
RobertBoschstr. 5
6100 Darmstadt
Federal Rep. Germanymerusi@PWA-B.UUCP (Don Merusi) (02/26/86)
We have had substantial experience with condition handlers. The only thing you have to be careful about, is using some of the routines that Digital has provided as part of the FORTRAN language merely as a migration facility. One routine, in particular, ERRSET is not recommended since it does cause inconsistent results. However, using the other condition handler facilities, such as LIB$SIGNAL, LIB$ESTABLISH or LIB$MATCH work quite well. If you have access to the Colorado Support Center System, they provide you with an excellent example of how to write a FORTRAN program using condition handlers.