[lld] r193301 - [Driver] Change UniversalDriver to use TD file.

Rui Ueyama ruiu at google.com
Thu Oct 24 15:06:33 PDT 2013


On Thu, Oct 24, 2013 at 12:38 AM, Alex Rosenberg <alexr at ohmantics.com>wrote:

> On Oct 23, 2013, at 11:56 PM, Rafael EspĂ­ndola <rafael.espindola at gmail.com>
> wrote:
>
> >> core was supposed to be the "one true" flavor that would potentially be
> the "native" command line for lld in the absence of a need for legacy
> compatibility and for the future.
> >
> > Is that realistic? On the clang side there was some discussion about
> > adding a clean command line interface, but the added value of doing so
> > seemed small.
>
> It's where we started.
>
> >> From
> http://lists.cs.uiuc.edu/pipermail/cfe-dev/2011-November/018588.html
> >
> > However, the challenge then becomes "why would anyone use the 'nice'
> > one"?  It is likely to get less development and features, and users of
> > the compiler don't care about academic purity of the implementation
> > details.  Any features of the 'nice' driver would also need to
> > supported by the 'ugly' driver also, so I don't see that introducing a
> > new driver is really all that useful.
>
> Some flavors can't express everything in the command line, e.g. GNU ld
> uses linker scripts, whereas ld64 can do everything on the command line.
> The hope was that a single universal command line language would be the
> best future direction.
>

I'm not sure if we really want to support everything that the linker script
can express in the command line. Some linker script directives, such as
ENTRY() or TARGET() can easily be translated to command line, but some
features like the variable look so complicated that it's not realistic to
do it in command line. If we try to do, we'd ended up having large set of
command line options that is complex as the linker script. Do we want it?

Also, the other flavors were internally translated to core command lines,
> so it always got tested and always had every feature.
>
> Alex
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20131024/e9698403/attachment.html>


More information about the llvm-commits mailing list