parker@zaphod.mpr.ca (Ross Parker) (03/24/90)
Does anyone know if ANSI header files are available for UNIX (specifically Ultrix 3.0 & 3.1)? I.e. header files to replace the standard /usr/include/... but complete with function prototypes? Thanks, Ross Parker parker@mpre.mpr.ca (604)293-5495 uunet!ubc-cs!mpre!parker
gwyn@smoke.BRL.MIL (Doug Gwyn) (03/25/90)
In article <2105@kiwi.mpr.ca> parker@zaphod.mpr.ca (Ross Parker) writes: >Does anyone know if ANSI header files are available for UNIX >(specifically Ultrix 3.0 & 3.1)? I.e. header files to replace >the standard /usr/include/... but complete with function prototypes? You don't want to do this! A C implementation consists of standard headers, a standard run-time library, and a compiler. All parts have to work together, so it is a mistake to replace just one of them. In particular, if the compiler doesn't handle prototypes the headers had better not use them. If it does handle prototypes, I would expect the implementation to have used them in its standard headers.
jsnyder@lehi3b15.csee.Lehigh.EDU (Jay Snyder) (03/26/90)
I'd be interested in ANSI header files to be used with GCC on my SCO Xenix system, as GCC and SCO's headers don't get along. -- ==================================================== Jay A. Snyder jsnyder@lehi3b15.csee.lehigh.edu CSEE Dept. lehi3b15!jsnyder Lehigh University
darcy@druid.uucp (D'Arcy J.M. Cain) (03/28/90)
In article <2105@kiwi.mpr.ca> parker@zaphod.mpr.ca (Ross Parker) writes: >Does anyone know if ANSI header files are available for UNIX >(specifically Ultrix 3.0 & 3.1)? I.e. header files to replace >the standard /usr/include/... but complete with function prototypes? > I have been modifying the standard files for SV3.2 (ESIX) and would not mind posting the results but I am not sure what the copyright situation would be. I don't think diffs would be as universally useful as the full files. I suppose I could simply extract the function prototypes and post them but that seems too much like work. Any ideas? -- D'Arcy J.M. Cain (darcy@druid) | Thank goodness we don't get all D'Arcy Cain Consulting | the government we pay for. West Hill, Ontario, Canada | (416) 281-6094 |
root@nebulus.UUCP (Dennis S. Breckenridge) (03/28/90)
In article <1990Mar27.181713.7875@druid.uucp> darcy@druid.UUCP (D'Arcy J.M. Cain) writes: > useful as the full files. I suppose I could simply extract the function > prototypes and post them but that seems too much like work. Any ideas? > I would not release any of the ESIX code to the net unless you want to get sued by Everex for copywrite infringment. Any code that belongs to Everex cannot be freely distributed without express permission in writing from the vendor. Be careful with context diffs as well. Darcy tsk tsk tsk! :-) -- ----------------------------------------------------------------------------- Dennis S. Breckenridge (604) 277-7413 dennis@nebulus.uucp VE7TCP EMACS: Eight Megabytes And Constantly Swapping! -----------------------------------------------------------------------------
darcy@druid.uucp (D'Arcy J.M. Cain) (03/28/90)
In article <12431@smoke.BRL.MIL> gwyn@brl.arpa (Doug Gwyn) writes: >In article <2105@kiwi.mpr.ca> parker@zaphod.mpr.ca (Ross Parker) writes: >>Does anyone know if ANSI header files are available for UNIX >>(specifically Ultrix 3.0 & 3.1)? I.e. header files to replace >>the standard /usr/include/... but complete with function prototypes? > >You don't want to do this! Well I do. I have installed gcc which doesn't come with header files. It uses the standard AT&T library so the standard header files are sort of OK but they don't have prototypes. I have been slowly upgrading the header files to be more ANSI compliant so that they work with gcc better. -- D'Arcy J.M. Cain (darcy@druid) | Thank goodness we don't get all D'Arcy Cain Consulting | the government we pay for. West Hill, Ontario, Canada | (416) 281-6094 |
root@nebulus.UUCP (Dennis S. Breckenridge) (03/29/90)
In article <1990Mar28.144531.12008@druid.uucp> darcy@druid.UUCP (D'Arcy J.M. Cain) writes: > sort of OK but they don't have prototypes. I have been slowly upgrading --------> WOW <---------- > the header files to be more ANSI compliant so that they work with gcc > better. Darcy, if you arbitrarily stomp on /usr/include with the ANSI C or gcc header files you probably will get to a point where your kernel will not compile and link. There are *MANY* dependencies on the various XENIXisms found all over the place. TIOCGWINSZ is the one that frags me all the time. If you use some of the net programs like "rn" that take best guess at thier environment they find all kinds of axymorons and try to build a XENIX - SYSV version of rn. I would put the gcc headers in ~include/gcc and then add -I/usr/include/gcc to the compile line. -- ----------------------------------------------------------------------------- Dennis S. Breckenridge (604) 277-7413 dennis@nebulus.uucp VE7TCP EMACS: Eight Megabytes And Constantly Swapping! -----------------------------------------------------------------------------
darcy@druid.uucp (D'Arcy J.M. Cain) (03/30/90)
In article <1990Mar28.041844.8655@nebulus.UUCP> root@nebulus.UUCP (Dennis S. Breckenridge) writes: >In article <1990Mar27.181713.7875@druid.uucp> darcy@druid.UUCP (D'Arcy J.M. Cain) writes: >> useful as the full files. I suppose I could simply extract the function >> prototypes and post them but that seems too much like work. Any ideas? >> >I would not release any of the ESIX code to the net unless you want to >get sued by Everex for copywrite infringment. Any code that belongs to >Everex cannot be freely distributed without express permission in writing >from the vendor. Be careful with context diffs as well. Actually they are copyright Eh TNT. The thing is that the code I want to release is written by me (more or less). example: the header file may have originally said extern int foo(); and I changed it to read extern int foo(const char *bar); and that's the line I really want to release. The problem is that it is surrounded by a lot of scary legal stuff and I really don't want to get involved with it. >Darcy tsk tsk tsk! :-) Dennis - have you been gone so long that you forget how to spell my name? 8^) -- D'Arcy J.M. Cain (darcy@druid) | Thank goodness we don't get all D'Arcy Cain Consulting | the government we pay for. West Hill, Ontario, Canada | (416) 281-6094 |
moraes@cs.toronto.edu (Mark Moraes) (03/30/90)
By default, gcc will look for headers in $(libdir)/gcc-include before /usr/include. ($(libdir) is defined in the Makefile) If you want to put prototyped headers on your system for gcc, gcc-include is probably a better place to do it. gcc -traditional won't work with those headers, though, and -traditional is often useful for compiling non-ANSItized code. (As more vendors start stuffing prototypes and volatile and const in their /usr/include, using gcc -traditional becomes harder; soon we'll have to maintain a separate include directory of traditional non-prototyped headers:-) Yeah, I know; we should all bite the ANSI bullet... Mark.