GnuCOBOL partially supports quite a few COBOL dialects and that support will continue to increase. One problem this causes is that different COBOL compilers support different reserved words. A valid identifier in one dialect may be a verb phrase in another. A configuration feature "specify-all-reserved: yes/no" was added, along with a list of the words that the compiler treats as reserved. Edward Hart did a spectacular job at setting up this powerful feature, allowing for very fine grained control over dialect support. But, this also causes a usability issue. Code that is "mostly" IBM would break when a valid standard clause of END-DISPLAY was used (for instance) as that is not a supported clause in IBM compilers. There is a maintained list of reserved words in each .conf file. The default was going to be "specify-all-reserved: yes" but decisions have been made to change that default to "specify-all-reserved: no". That means existing code that worked before will continue to compile. Which is a good thing. We'll need to document and propagate information for those rarer times when someone is porting code from a different compiler that may use a normally reserved word as an identifier. Or for those developers that want GnuCOBOL to bark when a feature may not compile with a different platform compiler. For GnuCOBOL compiler authors there is a new feature change request at https://sourceforge.net/p/open-cobol/feature-requests/212/ which is asking for opinions on how a final implementation should be managed. For everyone else on the gnucobol-users list, this note is just part of the information propagation so you are aware that when using -std=ibm or =mf or others that you may need to tweak the reserved word settings if porting to or from a different compiler. The feature is powerful, but also very rich in the small details as there are lots of variants in dialect syntax. So, author's take a look at wish list 212 and please add your august opinions. Cheers, Brian
