llvm dev - May 2016 - Automake Assembler Assumptions with LLVM-MC

All, we recently had a user report an issue associated with utilizing LLVM-MC as
the desired assembler in place of GNU’s GAS.  They are utilizing a tangled mess
of automake scripts to build this respective software package and have decided
to manually override the “CCAS” macro to point directly to LLVM-MC (rather than
using Clang).  Apparently, automake assumes that whatever “compiler” (as it is
referenced in the automake documentation) is used for assembling files must
accept the “-c” and “-o” options.  LLVM-MC correctly yells at the user given
that “-c” is not a valid option (as it should).  However, GAS does not.  It
happily accepts “-c” as a valid option.  We peered into the Bintutils 2.24
source and there is no explicit “-c” option support.  This appears to be an
option parsing bug in Binutils that either 1) collides with a known long option
or 2) somehow wanders into the “-ac” parsing logic.  Either way, its not a valid
option that does anything.

Given that LLVM-MC has basic support for handling GNU GAS options, what’s the
policy on being “bug for bug” compatible with GAS?  My personal opinion is the
user should fall back to using the compiler frontend to build their assembly
options, but this is a valid question nonetheless.  Thoughts?

Automake documentation on this topic is here: 
https://www.gnu.org/software/automake/manual/html_node/Assembly-Support.html

- John D. Leidel

Hi,

It's my understanding that llvm-mc is intended to be a testing tool for LLVM
developers rather than an assembler for end users. Users should be assembling
with clang.
> -----Original Message-----
> From: llvm-dev [mailto:llvm-dev-bounces at lists.llvm.org] On Behalf Of
John
> Leidel (jleidel) via llvm-dev
> Sent: 18 May 2016 13:35
> To: llvm-dev at lists.llvm.org
> Subject: [llvm-dev] Automake Assembler Assumptions with LLVM-MC
> 
> All, we recently had a user report an issue associated with utilizing
LLVM-MC
> as the desired assembler in place of GNU’s GAS.  They are utilizing a
tangled
> mess of automake scripts to build this respective software package and have
> decided to manually override the “CCAS” macro to point directly to LLVM-MC
> (rather than using Clang).  Apparently, automake assumes that whatever
> “compiler” (as it is referenced in the automake documentation) is used for
> assembling files must accept the “-c” and “-o” options.  LLVM-MC correctly
> yells at the user given that “-c” is not a valid option (as it should). 
However,
> GAS does not.  It happily accepts “-c” as a valid option.  We peered into
the
> Bintutils 2.24 source and there is no explicit “-c” option support.  This
appears
> to be an option parsing bug in Binutils that either 1) collides with a
known
> long option or 2) somehow wanders into the “-ac” parsing logic.  Either
way,
> its not a valid option that does anything.
> 
> Given that LLVM-MC has basic support for handling GNU GAS options, what’s
> the policy on being “bug for bug” compatible with GAS?  My personal opinion
> is the user should fall back to using the compiler frontend to build their
> assembly options, but this is a valid question nonetheless.  Thoughts?
> 
> Automake documentation on this topic is here:
> https://www.gnu.org/software/automake/manual/html_node/Assembly-
> Support.html
> 
> - John D. Leidel
> 
> 
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev

On Wed, May 18, 2016 at 01:10:50PM +0000, Daniel Sanders via llvm-dev
wrote:
> It's my understanding that llvm-mc is intended to be a testing tool
> for LLVM developers rather than an assembler for end users. Users
> should be assembling with clang.
That's my understanding and position as well.

Joerg