Issue 647: include files (cxx_revision) Source: (, ) Nature: Uncategorized Severity: Summary: Summary: There seems to be nothing in the spec that specifies the name of the include files for things in the CORBA module (e.g. type definitions). Add such a requirement to each language mapping Resolution: Revised Text: Actions taken: August 1, 1997: received issue Discussion: End of Annotations:===== Return-Path: X-Sender: rabin@postman.opengroup.org Date: Fri, 1 Aug 1997 15:08:47 -0400 To: cxx_revision@omg.org From: Paul Rabin Subject: An issue for the RTF Cc: issues@omg.org >>From andy@aptest.com Wed Jul 30 18:33:11 1997 >Date: Wed, 30 Jul 1997 15:32:42 -0700 (PDT) >From: andy@aptest.com (Andy Silverman) >To: rabin@opengroup.org >Subject: An issue for the RTF >Cc: ahby@themacs.com > >Hi Paul- > >Stephen and I have been discussing an issue which I'd like to request be >added to the list of things your RTF is considering. > >The CORBA spec defines the names of the include files produced by >IDL compilers, (as .h for C for example). However there seems >to be nothing in the spec that specifies the name of the include >files for things in the CORBA module (type definitions, etc.). >We would like to see such a requirement added to each language mapping >to which it applies, i.e. that CORBA.h for C and CORBA.hh for C++ shall be >what an application needs to include to get definitions for the >CORBA module. > >As vendors now all have vastly different schemes for naming >their include files now, I expect there might be some pushback on >this proposal based on compatibility with existing applications. >In this regard its worth pointing out that that just adding >a CORBA.h and .hh containing #include directives for an implementation's >existing include files would satisfy the proposed requirement, maintain >backward compatibility, and involve minimal development effort. Thus >we see this as something with minimal impact on ORB vendors, but which >will address a major impediment to the ability of application vendors >to create portable applications under the CORBA spec. > >Sound like a resonable topic for the RTF? > >- AS > >-- > >Andy Silverman Phone: 408-399-1930 x203 >Applied Testing and Technology Inc. Fax: 408-399-1931 >59 North Santa Cruz Avenue, Suite U Email: andy@aptest.com >Los Gatos, CA 95030 USA WWW: http://www.aptest.com/ > Paul Rabin tel: +1 617 621 8873 The Open Group - Open Software Foundation fax: +1 617 621 0584 11 Cambridge Center email: p.rabin@opengroup.org Cambridge, MA 02142-1405 USA Return-Path: Sender: jon@sems.com Date: Fri, 01 Aug 1997 13:03:28 -0700 From: Jonathan Biggar Organization: Seagate Software NMSG To: Paul Rabin Cc: cxx_revision@omg.org, issues@omg.org, port-rtf@omg.org Subject: Re: An issue for the RTF References: Another thing to consider at the same time is whether the new classes defined by the POA spec should be in this ".hh" file or whether they should be in a separate file. I would argue to separate them, so that client programs don't need the overhead of compiling class definitions they will never need. Of course, we are likely to run afowl of the "religious" wars over what is the appropriate file extension for a C++ header file: ".h", ".hh", ".hxx", ".hpp", ".H", etc, etc, etc. Of course if we are to follow the style of the new C++ standard, it should just be: #include or is that: #include Jon Paul Rabin wrote: > > >>From andy@aptest.com Wed Jul 30 18:33:11 1997 > >Date: Wed, 30 Jul 1997 15:32:42 -0700 (PDT) > >From: andy@aptest.com (Andy Silverman) > >To: rabin@opengroup.org > >Subject: An issue for the RTF > >Cc: ahby@themacs.com > > > >Hi Paul- > > > >Stephen and I have been discussing an issue which I'd like to > request be > >added to the list of things your RTF is considering. > > > >The CORBA spec defines the names of the include files produced by > >IDL compilers, (as .h for C for example). However there > seems > >to be nothing in the spec that specifies the name of the include > >files for things in the CORBA module (type definitions, etc.). > >We would like to see such a requirement added to each language > mapping > >to which it applies, i.e. that CORBA.h for C and CORBA.hh for C++ > shall be > >what an application needs to include to get definitions for the > >CORBA module. > > > >As vendors now all have vastly different schemes for naming > >their include files now, I expect there might be some pushback on > >this proposal based on compatibility with existing applications. > >In this regard its worth pointing out that that just adding > >a CORBA.h and .hh containing #include directives for an > implementation's > >existing include files would satisfy the proposed requirement, > maintain > >backward compatibility, and involve minimal development effort. > Thus > >we see this as something with minimal impact on ORB vendors, but > which > >will address a major impediment to the ability of application > vendors > >to create portable applications under the CORBA spec. > > > >Sound like a resonable topic for the RTF? > > > >- AS > > > >-- > > > >Andy Silverman Phone: 408-399-1930 x203 > >Applied Testing and Technology Inc. Fax: 408-399-1931 > >59 North Santa Cruz Avenue, Suite U Email: andy@aptest.com > >Los Gatos, CA 95030 USA WWW: > http://www.aptest.com/ > > > > Paul Rabin > tel: > +1 617 621 8873 > The Open Group - Open Software Foundation fax: +1 617 > 621 0584 > 11 Cambridge Center email: > p.rabin@opengroup.org > Cambridge, MA 02142-1405 > USA Return-Path: From: Jeffrey Mischkinsky Subject: Re: An issue for the RTF To: rabin@opengroup.org (Paul Rabin) Date: Fri, 1 Aug 1997 13:05:00 -0700 (PDT) Cc: cxx_revision@omg.org, issues@omg.org i'd suggest that we get #include semantics agreed to first. This was supposed to be on the portability group's plate, but it was dropped. There's not much point in getting agreement on the file names until there is agreement as to how they should be used. Almost noone likes the required semantics as defined in the spec. Some folks implement it anyway, others take other approaches. jeff 'Paul Rabin' writes: > > >>From andy@aptest.com Wed Jul 30 18:33:11 1997 > >Date: Wed, 30 Jul 1997 15:32:42 -0700 (PDT) > >From: andy@aptest.com (Andy Silverman) > >To: rabin@opengroup.org > >Subject: An issue for the RTF > >Cc: ahby@themacs.com > > > >Hi Paul- > > > >Stephen and I have been discussing an issue which I'd like to request be > >added to the list of things your RTF is considering. > > > >The CORBA spec defines the names of the include files produced by > >IDL compilers, (as .h for C for example). However there seems > >to be nothing in the spec that specifies the name of the include > >files for things in the CORBA module (type definitions, etc.). > >We would like to see such a requirement added to each language mapping > >to which it applies, i.e. that CORBA.h for C and CORBA.hh for C++ shall be > >what an application needs to include to get definitions for the > >CORBA module. > > > >As vendors now all have vastly different schemes for naming > >their include files now, I expect there might be some pushback on > >this proposal based on compatibility with existing applications. > >In this regard its worth pointing out that that just adding > >a CORBA.h and .hh containing #include directives for an implementation's > >existing include files would satisfy the proposed requirement, maintain > >backward compatibility, and involve minimal development effort. Thus > >we see this as something with minimal impact on ORB vendors, but which > >will address a major impediment to the ability of application vendors > >to create portable applications under the CORBA spec. > > > >Sound like a resonable topic for the RTF? > > > >- AS > > > >-- > > > >Andy Silverman Phone: 408-399-1930 x203 > >Applied Testing and Technology Inc. Fax: 408-399-1931 > >59 North Santa Cruz Avenue, Suite U Email: andy@aptest.com > >Los Gatos, CA 95030 USA WWW: http://www.aptest.com/ > > > > Paul Rabin tel: > +1 617 621 8873 > The Open Group - Open Software Foundation fax: +1 617 621 0584 > 11 Cambridge Center email: > p.rabin@opengroup.org > Cambridge, MA 02142-1405 > USA > > > -- Jeff Mischkinsky jmischki@dcn.davis.ca.us +1 916-758-9850 jeffm@visigenic.com +1 415-312-5158 Return-Path: Date: Fri, 1 Aug 1997 15:03:51 -0700 (PDT) From: andy@aptest.com (Andy Silverman) To: Jeffrey Mischkinsky , rabin@opengroup.org (Paul Rabin) Subject: Re: An issue for the RTF Cc: cxx_revision@omg.org, issues@omg.org Wouldn't the IDL #include semantics debate and the issue raised here (names for ORB supplied include files for the CORBA interfaces (DII, IR, etc.) in the C and C++ bindings) be separate problems? [In "Re: An issue for the RTF" on Aug 1, 1997 Jeffrey Mischkinsky wrote:] > i'd suggest that we get #include semantics agreed to first. > This was supposed to be on the portability group's plate, but it was > dropped. > There's not much point in getting agreement on the file names until > there is agreement as to how they should be used. > Almost noone likes the required semantics as defined in the spec. > Some folks implement it anyway, others take other approaches. > jeff > 'Paul Rabin' writes: > > > > >>From andy@aptest.com Wed Jul 30 18:33:11 1997 > > >Date: Wed, 30 Jul 1997 15:32:42 -0700 (PDT) > > >From: andy@aptest.com (Andy Silverman) > > >To: rabin@opengroup.org > > >Subject: An issue for the RTF > > >Cc: ahby@themacs.com > > > > > >Hi Paul- > > > > > >Stephen and I have been discussing an issue which I'd like to request be > > >added to the list of things your RTF is considering. > > > > > >The CORBA spec defines the names of the include files produced by > > >IDL compilers, (as .h for C for example). However there seems > > >to be nothing in the spec that specifies the name of the include > > >files for things in the CORBA module (type definitions, etc.). > > >We would like to see such a requirement added to each language mapping > > >to which it applies, i.e. that CORBA.h for C and CORBA.hh for C++ shall be > > >what an application needs to include to get definitions for the > > >CORBA module. > > > > > >As vendors now all have vastly different schemes for naming > > >their include files now, I expect there might be some pushback on > > >this proposal based on compatibility with existing applications. > > >In this regard its worth pointing out that that just adding > > >a CORBA.h and .hh containing #include directives for an implementation's > > >existing include files would satisfy the proposed requirement, maintain > > >backward compatibility, and involve minimal development effort. Thus > > >we see this as something with minimal impact on ORB vendors, but which > > >will address a major impediment to the ability of application vendors > > >to create portable applications under the CORBA spec. > > > > > >Sound like a resonable topic for the RTF? > > > > > >- AS > > > > > >-- > > > > > >Andy Silverman Phone: 408-399-1930 x203 > > >Applied Testing and Technology Inc. Fax: 408-399-1931 > > >59 North Santa Cruz Avenue, Suite U Email: andy@aptest.com > > >Los Gatos, CA 95030 USA WWW: http://www.aptest.com/ > > > > > > > Paul Rabin tel: > > +1 617 621 8873 > > The Open Group - Open Software Foundation fax: +1 617 > 621 0584 > > 11 Cambridge Center > email: > > p.rabin@opengroup.org > > Cambridge, MA 02142-1405 > > USA > > > > > > > > > -- > Jeff Mischkinsky > jmischki@dcn.davis.ca.us +1 916-758-9850 > jeffm@visigenic.com +1 415-312-5158 -- Andy Silverman Phone: 408-399-1930 x203 Applied Testing and Technology Inc. Fax: 408-399-1931 59 North Santa Cruz Avenue, Suite U Email: andy@aptest.com Los Gatos, CA 95030 USA WWW: http://www.aptest.com/