[cfe-dev] Status of SEH?

[mfaithfull]

I would obviously like clang SEH to be compatible with my client side SEH implementation for MSVC so they can share code. I do agree that some of the MSVC implementation seems overly convoluted. Either way keep me in the loop. I should have a reasonably complete SEH client library that works when built with MSVC and builds with clang online within the next 10 days as part of the QOR project Strata-1.

Matthew F.


Sent from Samsung Mobile

-------- Original message --------
From: Dennis Luehring <dl.soluz at gmx.net> 
Date:30/01/2014  20:18  (GMT+00:00) 
To: cfe-dev at cs.uiuc.edu 
Subject: Re: [cfe-dev] Status of SEH? 

>Does anyone feel strongly that we should follow
>the Visual C++ SEH approach

not by default but as an option if its needed to get closer to
full interop with vstudio compiled libs

see Reid Kleckner talk "Bringing clang and LLVM to Visual C++ users" 
(http://llvm.org/devmtg/2013-11/#talk11)

Am 30.01.2014 20:37, schrieb Jb Feldman:
> So, as I'm looking at this, thinking about trying to get it implemented by
> June, I came up with a question for the list. Does anyone feel strongly
> that we should follow the Visual C++ SEH approach or just make sure it is
> functional. Visual C++ uses a crazy fancy extended structure to make sure
> that everything goes back to the same function, but the operating system
> doesn't mandate that we take that approach.
>
> Opinions?
>
> JB
>
>
> On Thu, Jan 30, 2014 at 9:18 AM, Jb Feldman <jb.feldman at kyrus-tech.com>wrote:
>
> > Yes, it looks like the patent will expire in June, so that's exciting.
> >
> >
> > On Thu, Jan 30, 2014 at 12:57 AM, Jacob Carlborg <doob at me.com> wrote:
> >
> >> On 2014-01-30 04:41, Jb Feldman wrote:
> >>
> >>> I was wondering what the current state of SEH and planned SEH support
> >>> is. The things I've found in mailing list/forum history indicate that
> >>> parsing and AST are now supported, but the back end is not in place.
> >>>
> >>> Is the problem a lack of implementation, or the referenced patent
> >>> issues? Are there any paths forward identified? I know that C++
> >>> exceptions work in user mode, but that won't cut it in Windows drivers.
> >>>
> >>
> >> As far as I understand, from reading this mailing list, is that nothing
> >> will happen as long as there is a patent on the technology. But if I recall
> >> correctly that patent should end sometime this year.
> >>
> >> --
> >> /Jacob Carlborg
> >>
> >> _______________________________________________
> >> cfe-dev mailing list
> >> cfe-dev at cs.uiuc.edu
> >> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev
> >>
> >
> >
>
>
>
> _______________________________________________
> cfe-dev mailing list
> cfe-dev at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

_______________________________________________
cfe-dev mailing list
cfe-dev at cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20140130/84159688/attachment.html>