Old C Compiler Lives Again Under GPL
JordanH writes "The DeSmet-C compiler, a commercial C compiler from the '80s, is being released under the GPL. Yet another alternative C compiler implementation available for your coding pleasure."
← Back to Stories (view on slashdot.org)
I do think it's good to bring new life to old code by open-sourcing it, but this one seems a little problematic. Playing Doom or using an old OS for a special purpose can be fun and useful, but why would you want to chance your code on an old and questionably maintained compiler?
If I'm missing something, let me know, but this just looks a bit more useless than the average open source project. I know it's educational, but is there anything real anyone would chance on it?
Don't get me wrong, I love good compilers, but with all the changes in the standard, and current, good, optimizing compilers, why would we want this?
antipaucity
Pretty nifty story behind the open-sourcing. I wonder how many other applications have received this sort of treatment and how many more wait in anticipation for freedom.
First post?
How is something from the 80's considered OLD?! It's not even vintage yet. I have underwear from the 80's that are just fine.
From the FA:
"""
There are other, smaller options like TCC that is a complete C compiler, but it's too geared to 386+ and Linux to be a good playground. Other open-source C compilers tend to be variations of Small C that, while understandable, don't implement the entire language.
"""
I remember seeing this in shareware catalogs. Didn't it generate assembler code, and you had to have MASM to build your program?
Anyone remember MIX C? They used to have the full-page ads in Computer Shopper, you got the compiler, book, and everything for $59.95.
I never got to use either of them, I was lucky enough to get a copy of Borland Turbo C.
I started feeling the usual insane urge to play with a C compiler source.
It seems it's more for people who just want to poke at the source of a real compiler, without having to deal with the mind-boggling complexities of GCC. I might take a look myself.
LOAD "SIG",8,1
Id like to see the hisoft compiler open sourced.
DICE C, wildly popular on the Amiga is free and available here. DICE was how Matt Dillon (As in DragonFly BSD) made his name as a programmer.
VBCC, also popular on the Amiga, is free.
I thought they didn't even give copyright to software until the 90s.
I should have added that I saw this on the excellent PLNews: Programming Language News site.
<PLUG>
If you want to try something altogether larger, more powerful and more flexible, then check out the ACK --- this is a compiler toolchain written by Andy Tanenbaum and Ceriel Jacobs that was released under a BSD license a few years ago. It supports K&R C, ANSI C, Pascal, Modula 2, Occam, Basic and Fortran, and supports a whole bunch of (slightly elderly) architectures. A subset of the compiler comes with Minix, if you've ever used that. Ever wanted to run Occam programs on your Apple I? Yep, you can do that.
It's way, way smaller than gcc, astonishingly faster, much easier to port new architectures for, and produces adequate if not brilliant code.
</PLUG>
1) I survived a slashdotting! Yah!! (at least it wasn't the front page!)
2) Where I expect the source code to show up most often is in someone's compiler class homework (unattributed, of course...)
3) People have been asking "what's the use of an old compiler"? As I mentioned on the website, it's small enough so that a noobie (like me) can get their teeth into it. The code itself may never go anywhere, but the lessons learned will be around for a long time.
Bill
I've got software I wrote in the 80's that is still in use, which I still use that compiler for. It never was worth porting to a newer compiler, but sometimes I need to make cosmetic changes and recompile.
unfortunate choice of domain on slashdot. :)
..don't panic
What's wrong with newer compilers being geared for 386+?
What about MIPS, SPARC, ARM, PowerPC? A retargetable compiler may prove useful to more people than one that makes too many x86-specific assumptions.