This is the mail archive of the gcc-prs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

c++/413: exception handling with -fsjlj-exceptions is broken



>Number:         413
>Category:       c++
>Synopsis:       exception handling with -fsjlj-exceptions is broken
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          wrong-code
>Submitter-Id:   net
>Arrival-Date:   Fri Jul 14 01:56:00 PDT 2000
>Closed-Date:
>Last-Modified:
>Originator:     fjoe@iclub.nsu.ru
>Release:        g++ 2.95.2
>Organization:
>Environment:
lark:/$g++ -v
Using builtin specs.
gcc version 2.95.2 19991024 (release)
lark:/$uname -a
FreeBSD lark.nsk.bsgdesign.com 4.0-STABLE FreeBSD 4.0-STABLE #0: Mon Jul 10 17:21:18 NOVST 2000     fjoe@lark.nsk.bsgdesign.com:/usr/obj/usr/src/sys/lark  i386
lark:/$
>Description:
emitted setjmp got optimized after flow analysis in such way
that after exceptions is raised execution continues in
try {} block instead of catch {}
this is very similar to #258 but it can be 100% reproduced
without any optimization options turned on
>How-To-Repeat:
unfortunately I could find only one piece of code which
can show described behaviour:
Reactor_Exceptions_Test from The ACE (Adaptive Communications
Environment) 5.1.1 under FreeBSD 4.x using native g++ 2.95.2 compiler
(please note that native compiler should be used, compiler installed from ports uses
dwarf 2 unwinding info mechanism and is not affected)
>Fix:
none
>Release-Note:
>Audit-Trail:
>Unformatted:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]