Re: c++/3099: cygwin: Many g++ EH failures

Re: c++/3099: cygwin: Many g++ EH failures

[Gerald Pfeifer]

The following reply was made to PR c++/3099; it has been noted by GNATS.

From: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
To: "Joseph S. Myers" <jsm28@cam.ac.uk>, Jason Molenda <jason@molenda.com>
Cc: Christopher Faylor <cgf@redhat.com>, <rodrigc@gcc.gnu.org>,
   <David.Billinghurst@riotinto.com>, <gcc-bugs@gcc.gnu.org>,
   <gcc-gnats@gcc.gnu.org>, <rth@gcc.gnu.org>, <overseers@gcc.gnu.org>,
   <gcc-patches@gcc.gnu.org>
Subject: Re: c++/3099: cygwin: Many g++ EH failures
Date: Mon, 12 Nov 2001 09:50:24 +0100 (CET)

 On Sun, 11 Nov 2001, Joseph S. Myers wrote:
 > I suspect the following patch might fix the problem; OK to commit?
 
 Yes, I had a look, and it should work.
 
 > If it does fix the problem, I'll submit a PR concerning the need for
 > someone to go through all gcc-bugs messages from the start of June until
 > now that contain a PR number in the Subject line but were not initial PR
 > submissions or sent to gcc-gnats, and send their URLs to the relevant PRs,
 > and reopen any wrongly closed for lack of feedback.
 
 Please do so.
 
 On Sun, 11 Nov 2001, Jason Molenda wrote:
 > There is a new gnatsweb maintainer these days, and I'd wager he
 > may be ameanable to accepting this change.
 
 And if it does indeed fix the problem, Joseph, would you mind submitting
 this patch to gnats-devel@sources.redhat.com as well?
 
 It is only relevant for the gnatsweb 2.9x-branch, as gnatsweb 4.0 has
 changed quite a bit in this respect AFAICS.
 
 Gerald
 -- 
 Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/
 

Re: c++/3099: cygwin: Many g++ EH failures

[rth]

Old Synopsis: cygwin:  Many g++ EH failures
New Synopsis: cygwin: Many g++ EH failures

Responsible-Changed-From-To: unassigned->rth
Responsible-Changed-By: rth
Responsible-Changed-When: Tue Jun 12 23:20:59 2001
Responsible-Changed-Why:
    I'll take it.
State-Changed-From-To: analyzed->feedback
State-Changed-By: rth
State-Changed-When: Tue Jun 12 23:20:59 2001
State-Changed-Why:
    The problem is that Cygwin was changed from sjlj EH to call-frame EH,
    but the Cygwin runtime was not updated to match, and does not register
    the .eh_frame data, so we rather quickly abort.
    
    From my point of view, the preferable option is to update the Cygwin
    runtime rather than drop back to sjlj, but I don't know if that's
    really an option for you.
    
    Please advise.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=3099&database=gcc

Re: c++/3099: cygwin: Many g++ EH failures

[mmitchel]

Synopsis: cygwin:  Many g++ EH failures

State-Changed-From-To: closed->analyzed
State-Changed-By: mmitchel
State-Changed-When: Tue Jun 12 03:51:23 2001
State-Changed-Why:
    Not really fixed, according to submitter.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=3099&database=gcc

Re: c++/3099: cygwin: Many g++ EH failures

[rth]

Synopsis: cygwin:  Many g++ EH failures

State-Changed-From-To: open->closed
State-Changed-By: rth
State-Changed-When: Mon Jun 11 13:02:07 2001
State-Changed-Why:
    Duplicate of 3098.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=3099&database=gcc

c++/3099: cygwin: Many g++ EH failures

[David.Billinghurst]

>Number:         3099
>Category:       c++
>Synopsis:       cygwin:  Many g++ EH failures
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jun 09 12:26:00 PDT 2001
>Closed-Date:
>Last-Modified:
>Originator:     David Billinghurst
>Release:        gcc 3.0 (pre-release)
>Organization:
>Environment:
i686-pc-cygwin
>Description:
This should be fixed before the 3.0 release.

On cygwin, over 100 g++ tests fail with what seems like runtime EH problems.

These failures are only seen after applying patch http://gcc.gnu.org/ml/gcc-bugs/2001-05/msg00854.html 
 to work around PR 3098 (undefined reference to `___deregister_frame_info_bases)


See http://gcc.gnu.org/ml/gcc-testresults/2001-06/msg00216.html for list of tests that fail.
>How-To-Repeat:
Apply patch http://gcc.gnu.org/ml/gcc-bugs/2001-05/msg00854.html 

configure --enable-threads=posix
make check
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: