On Tuesday, August 31, 2004, at 04:37PM, Douglas B Rupp <[EMAIL PROTECTED]> wrote:
> >> The existence of a good port of GNU C/C++ for OpenVMS would make porting Open >> Source tools to OpenVMS easier, which would benefit some enterprise development. > >Well you are making exactly my point which is that a Unix-like interface >might be nice for porting Open Source tools, but I still don't see the need >for the Gnu C compiler itself. There are not many tools/applications that >have to be compiled with Gnu C in order to work. I think I mostly agree with Douglas on this. Most of the major open source packages (including Perl) have benefitted greatly from being thrown up against every commercial compiler on the market as well as GCC, so from a language point of view I don't see much benefit to open source porting that would result from having GCC available. There are more likely to be dependencies on GNU libraries than on the GNU compiler per se. What would help a lot with open source porting in general would be having a sufficiently robust GNV package with all the utilities a Configure script is likely to need and a cc command that is compatible with what unixy packages expect (regardless of what compiler that cc command invokes). The wrappers around cc and gcc in GNV are certainly headed in this direction; I haven't tried them recently enough to know how good they are. Another possibility, one that only HP could do, would be to implement an alternative cc command around the HP C compiler that would use Unix syntax (getopt() rather than CLI$DCL_PARSE). Since this already exists on Tru64, it might not be that hard. The need for a compiler to build open source packages on sites that do not normally develop in C is a genuine one. Single-user license prices for HP C are not quite as outrageous as unlimited licenses and I have seen various sites handle the problem that way. Perhaps a case could be made for including a single-user C license with the OS. Having said all this, I would be happy to see an up-to-date and useable GCC port on OpenVMS just for the niftiness factor if nothing else. It might bring with it some other capabilities, like Objective C, that are otherwise unavailable. But given the lack of a business case, it would require a serious (*really* serious) volunteer effort. A nice Ph.D. dissertation might be to port it to the GEM back end, then port to the Intel IA64 back end, then write up a comparison.
