home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: gnu.misc.discuss
- Path: sparky!uunet!mcsun!news.funet.fi!funic!nntp.hut.fi!nntp!sja
- From: sja@snakemail.hut.fi (Sakari Jalovaara)
- Subject: Re: Public Domain C Compiler?
- In-Reply-To: mcastle@cs.umr.edu's message of Wed, 6 Jan 1993 16:12:45 GMT
- Message-ID: <SJA.93Jan8182853@lk-hp-19.hut.fi>
- Sender: usenet@nntp.hut.fi (Usenet pseudouser id)
- Nntp-Posting-Host: lk-hp-19.hut.fi
- Organization: Helsinki University of Technology, Finland
- References: <2B4A0FD5.FD6B@tct.com> <C0FuFu.o41@austin.ibm.com> <1993Jan6.161245.4926@umr.edu>
- Date: 08 Jan 93 16:28:52 GMT
- Lines: 64
-
-
- > I've seen a few posts here [...] I have doubts to their dedication
- > to the project.
-
- Well, "dedication"... I've been writing a free C compiler as a
- hobby. I can't guarantee eternal "dedication"; so far I've stuck
- with it. It really is an interesting and challenging project.
-
- The front end is 95% done (doesn't it always seem like your programs
- are 95% finished?); a back end is early stages of construction.
-
- (With suitable definitions of "back end" & "front end".)
-
- An ANSI C test suite (way over 50% of my effort so far) is shaping
- up very nicely.
-
- If you want a compiler in a shiny plastic wrapper, a glossy
- manual with pictures of cute furry animals explaining its use,
- delivered to your door by Federal Express -- shop elsewhere.
-
- > Writing a compiler is a daunting project
-
- "Daunting" is a bit strong; it's not _that_ difficult. Depends;
- if you want to put in every conceivable optimization and port
- it to everything from vector processors to toasters...
-
- A reasonably reliable (better than what you usually get in UN*X)
- compiler with a couple of back ends is certainly doable.
-
- > Writing a portable compiler, even more so.
-
- This isn't *too* bad either. Depends on the machine specific
- optimizations you want to do; code generation is not that
- difficult given the research that has been published over the
- last 10..15 years.
-
- Debugging information is a pain. I don't think I'll be able
- to do that for more than a few machines by myself.
-
- Otherwise, I'd venture to guess that the portability problems
- of e.g. tcsh are far greater than those of a well written compiler.
-
- > it becomes quite a logistic problem to try to incorporate changes
- > into the original compiler in a coherent manner.
-
- Depends... If you just want to merge machine dependent changes (code
- generator generator descriptions etc) it isn't too bad. If you
- incorporate optimizations and language extensions you'll have more
- work.
-
- Incorporating changes you can't test is a lot of work but others
- seem to be coping (tcsh? elm?) I'm looking forward to it if I
- ever get there.
-
- You know, there is a theory that says that free software tends to be
- easy to maintain because there are so many people who can help with it.
- If the original authors can't continue working on it, the theory says
- someone else can adopt it. Maybe that is even true!
-
- > I just really doubt one person could do it and remain sane
-
- Thanks a lot, sunshine :-) I've always been halfway over the hill
- so this doesn't bother me.
- ++sja
-