khb@fatcity.Sun.COM (fatcity) (03/18/89)
Richard Maine reccently suggested that f88 (to use the wg5 term :>) compilers will be required to diagnose anything that is illegal. I quote, from version 111 of the draft (to be voted on in May): " Note that a standard-conforming program must not use any forms or relationships that are prohibited by this standard, but a standard-conforming processor may allow such forms and relationships if they do not change the proper interpretation of a standard conforming program. For example, a standard conforming processor may allow a nonstandard data type " This is from page 1-2 of version 111. I do not know of any counter-example to suggest otherwise. It makes perfect sense to allow useful extensions. I (personally) would be happier if vendor's don't allow strange do-loop nesting simply because it would not invalidate their compiler's f88 (wg5 :>) compliance. kbierman@sun.com Tactical Engineering Keith H. Bierman It's Not My Fault ---- I Voted for Bill & Opus
hirchert@uxe.cso.uiuc.edu (03/21/89)
Keith H. Bierman (khb@fatcity.Sun.COM) writes >Richard Maine reccently suggested that f88 (to use the wg5 term :>) >compilers will be required to diagnose anything that is illegal. >I quote, from version 111 of the draft (to be voted on in May): >" >Note that a standard-conforming program must not use any forms or >relationships that are prohibited by this standard, but a >standard-conforming processor may allow such forms and relationships >if they do not change the proper interpretation of a standard >conforming program. For example, a standard conforming processor may >allow a nonstandard data type >" >This is from page 1-2 of version 111. >I do not know of any counter-example to suggest otherwise. 1. Keith is right in that Fortran 8x, like FORTRAN 77, allows a standard- conforming processor to support extensions. 2. Richard is right that Fortran 8x, unlike FORTRAN 77, requires a standard- conforming processor to offer the OPTION of identifying nonstandard usage. Note that this option need not be turned on by default, and that even when turned on, it is only necessary to identify nonstandard usage, not to consider it an error. Thus, compiler writers are free to add whatever goodies they wish, but people trying to write portable programs can expect assistance in getting the syntactic parts right. Kurt W. Hirchert hirchert@ncsa.uiuc.edu National Center for Supercomputing Applications