Re: compiler back-end development?

  From: "Ian Lance Taylor" <ian@airs.com>
> "BGB / cr88192" <cr88192@hotmail.com> writes:
>
>> yep, one place where I would like to see improvement personally is GCC's
>> Win64 support...
>>
>> it is presently, IMO, to put it nicely, crap...
>
> There has been quite a bit of work recently to improve the win64
> support. The development version of gcc should be better, though I
> don't know how much remains to be done.
>


I don't know either...


the current version I have identifies itself as '4.4.0 20090106
(experimental)'...


I have thus far not succeeded in getting the thing to rebuild from sources
(it was distributed in binary form).


this is typical IME of GCC on Windows: I have, as of yet, never been able to
successfully build gcc on Windows...




>
>> I am almost tempted to join in, apart from the general sense of horror
>> experienced thus far WRT GCC's codebase... (it is funny, I want it
>> improved,
>> but don't want to have to deal with their codebase). it is terrible
>> enough
>> just trying to get it to build...
>
> It's free software. People joining in is the only way it is going to
> get better. The gcc-help@gcc.gnu.org mailing list has plenty of
> people who can help you get started, but you do have to meet us
> halfway.
>


it is also one of the most complicated and terrible codebases I am aware
of...


(FWIW, I have doubts that it would be worth my while, or theirs, for me to
go through the effort of familiarizing myself with the GCC codebase...). by
the time I could likely contribute much of anything of value, probably they
will have gotten the problems fixed on their own.


meanwhile, I also have my own compiler project that I am in the process of
migrating to x86-64 (note: not in direct conflict with GCC, as my project
and gcc fulfill essentially different roles, and I am much more likely to
end up competing against LLVM than I am against GCC...).


after all, the people who started on this front, probably know a whole lot
more about what they are doing than I would.




now, one could argue that competing against LLVM would probably just be a
losing battle, but oh well...


note that I don't really too much try to address the static compilation
front, mostly as I had assumed that GCC and friends would fill this role.
note the less, it would not be impossible by any means, just probably I
might have reason to write my own mini-binutils (probably focussing
specifically on the archs and OS's I care about, AKA, no BFD or similar,
each tool OS-specific and self-contained...).




as-is, there are also MS's tools, but I am not having exactly the maximum of
fondness towards them:
bleh, nmake (sucks so hard I don't use it...);
bleh, link, annoys me as it doesn't produce the sort of EXEs I want;
...


for a fairly long time, I had been tempted to write my own debugger as well,
mostly to address an inherent limitation of gdb:
it can't deal with dynamically generated code, so I just end up with
stackdumps with a whole lot of "??" in them... (still, it is better than no
debugger, which is the present situation...).




but, for now, things work, sort of...




> Ian

Return to the comp.compilers page.
Search the comp.compilers archives again.