Re: Issues with compiler failures

[Chris Hanson <cmhanson%eschatologist.net@localhost>]

On May 31, 2019, at 6:14 AM, Paul Koning <paulkoning%comcast.net@localhost> wrote:

>> On May 31, 2019, at 3:16 AM, Chris Hanson <cmhanson%eschatologist.net@localhost> wrote:
>> 
>> On Mar 17, 2019, at 6:33 AM, Paul Koning <paulkoning%comcast.net@localhost> wrote:
>>> 
>>> Is there no one willing to take on the gcc maintainer role for vax?  It's not rocket science.
>> 
>> Honestly, for many people it might be less daunting to consider putting together an LLVM backend and clang support for VAX.
> 
> Less daunting?  Hm.  The GCC backend for VAX exists, it has a couple of bugs and needs an update.  Would LLVM involve a brand new backend?  How does the LLVM backend complexity, and the documentation about writing one, compare to the GCC case?  Not all GNU projects have good internals documentation, but GCC's does, and in particular it covers back ends rather well.

LLVM would need a new backend, yes.

I’ve found the LLVM documentation and structure rather clear and straightforward when I’ve worked with it, compared to what I remember of GCC from the 4.2.1-and-earlier days.

I can’t have anything to do with GPLv3 code for professional reasons, so I haven’t looked at GCC since then. This is the other reason I prefer LLVM/clang whenever possible, but in this case I think it’s more about what a new contributor would find approachable.

  — Chris