[PATCH] D13501: [ELF2] Add -wrap switch

[Hal Finkel via llvm-commits]

----- Original Message -----

> From: "Igor Kudrin via llvm-commits" <llvm-commits at lists.llvm.org>
> To: "Rui Ueyama" <ruiu at google.com>
> Cc: "llvm-commits" <llvm-commits at lists.llvm.org>,
> reviews+D13501+public+cc48a9c32b74c6f4 at reviews.llvm.org
> Sent: Friday, October 9, 2015 10:13:20 AM
> Subject: Re: [PATCH] D13501: [ELF2] Add -wrap switch

> On 09.10.2015 20:45, Rui Ueyama wrote:

> > On Thu, Oct 8, 2015 at 9:33 PM, Igor Kudrin < ikudrin.dev at gmail.com
> > >
> > wrote:
> 

> > > On 09.10.2015 4:08, Rui Ueyama wrote:
> > 
> 

> > > > On Wed, Oct 7, 2015 at 10:55 PM, Igor Kudrin <
> > > > ikudrin.dev at gmail.com
> > > > > wrote:
> > > 
> > 
> 

> > > > > I've added some examples to show the differences between your
> > > > > patch
> > > > > and GNU ld's behavior, see http://reviews.llvm.org/D13528 .
> > > > > For
> > > > > me,
> > > > > it's not just the different way to implement the switch, but
> > > > > mostly
> > > > > change its meaning.
> > > > 
> > > 
> > 
> 
> > > > It is indeed different from GNU, but I think our behavior makes
> > > > sense
> > > > and I don't think that's a bad difference. LLD's symbol
> > > > handling
> > > > is
> > > > already pretty different in some area, particularly how we
> > > > handle
> > > > archive files and --start-group/--end-group. Just like that,
> > > > this
> > > > is
> > > > something we don't want to aim the exact compatibility.
> > > 
> > 
> 
> > > Any unnecessary difference may lead to unexpected problems for
> > > end
> > > users. According to this, I don't feel like accepting your
> > > approach,
> > > which, compared to ld's behavior, is different in almost all
> > > aspects
> > > I can imagine, whereas my original patch follows ld's and gold's
> > > conduct pretty well. Maybe it's better not implement this switch
> > > at
> > > all.
> > 
> 

> > I'd agree. Maybe we want some implementation of --wrap in future,
> > but
> > probably it's too early. We can do that later after getting better
> > idea what end users would say.
> 
> OK. Let's postpone it for a while.

For the record, I'll want the --wrap feature eventually (we need it to support profiling tools on statically-linked binaries). 

-Hal 

> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits

-- 

Hal Finkel 
Assistant Computational Scientist 
Leadership Computing Facility 
Argonne National Laboratory 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20151009/cf37e99b/attachment.html>