[lld] r248389 - [ELF2] - added ignored command line options for compatibility.

Rui Ueyama via llvm-commits llvm-commits at lists.llvm.org
Wed Sep 23 14:58:32 PDT 2015


And as a side note, if we want to add command line options that are just
ignored, we don't want to add help texts just to say that they are ignored.
One day when we have a man page, we would want to write about that in it,
but they probably don't worth to take the real estate in the --help message.

On Wed, Sep 23, 2015 at 11:41 AM, George Rimar via llvm-commits <
llvm-commits at lists.llvm.org> wrote:

> Ok, its sounds reasonable. Sorry for this patch. Please revert then - I
> dont have write permissions for that.
> ________________________________________
> От: davide.italiano at gmail.com [davide.italiano at gmail.com] от имени Davide
> Italiano [davide at freebsd.org]
> Отправлено: 23 сентября 2015 г. 22:35
> Кому: George Rimar
> Копия: Rafael Espíndola; llvm-commits
> Тема: Re: [lld] r248389 - [ELF2] - added ignored command line options for
> compatibility.
>
> On Wed, Sep 23, 2015 at 11:05 AM, George Rimar <grimar at accesssoftek.com
> <mailto:grimar at accesssoftek.com>> wrote:
> > I found the description of that flag in The GNU Linker book -
> http://www.eecs.umich.edu/courses/eecs373/readings/Linker.pdf
> > Page 14:
> > -assert keyword
> > This option is ignored for SunOS compatibility
>
> This was insteresting. Unfortunately, option handling doesn't seem to be
> very consistent between version and documentation.
>
> On binutils 2.25:
> % /usr/local/bin/ld.bfd -assert
> /usr/local/bin/ld.bfd: unrecognized option '-assert'
>
> % /usr/local/bin/ld.gold -assert
> /usr/local/bin/ld.gold: -assert: missing argument
>
> It seems that also older versions of ld don't ignore that version, just
> not recognize it (both single and double dash).
>
> When there's ambiguity, I generally err on the side of being a little bit
> stricter (unless there's a good reason not to).
> In this case in particular, it seems you don't have a real use case for
> this, and there's some contraddiction between what documentation says, what
> ld.bfd does and what gold does.
>
> At the light of these facts, I propose to revert this given it seems
> whatever choice you make you violate something, and given this option has
> probably impact on a negligible amount of users.
>
> Please let me know if you have any objections.
>
> Thanks,
>
> --
> Davide
>
> "There are no solved problems; there are only problems that are more or
> less solved" -- Henri Poincare
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150923/6f64bdfd/attachment.html>


More information about the llvm-commits mailing list