Bug 15476 - g++: internal compiler error in scan_region (compiling hylafax)
Summary: g++: internal compiler error in scan_region (compiling hylafax)
Keywords:
Status: CLOSED RAWHIDE
Alias: None
Product: Red Hat Linux
Classification: Retired
Component: gcc
Version: 7.1
Hardware: i386
OS: Linux
medium
high
Target Milestone: ---
Assignee: Jakub Jelinek
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2000-08-05 14:32 UTC by giulioo
Modified: 2008-05-01 15:37 UTC (History)
0 users

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2000-08-21 10:45:20 UTC
Embargoed:


Attachments (Terms of Use)
Files needed to reproduce the compiler error (cd test-dir/faxmail;make) (64.31 KB, application/octet-stream)
2000-08-05 14:36 UTC, giulioo
no flags Details

Description giulioo 2000-08-05 14:32:09 UTC
There is an utility in hylafax that cannot be built on pinstripe, while it 
builds ok on rh61.

I assume it's a compiler bug since the compiler seems to say so; I don't 
understand anything about compilers.

/usr/bin/g++        -D__ANSI_CPP__ -I. -I.. -I.././regex -I.././faxmail -
I.././u
til -I/usr/include -g -O  -c MsgFmt.c++
MsgFmt.c++: In method `fxStr MsgFmt::mapHeader (const fxStr &)':
MsgFmt.c++:68: Internal compiler error in scan_region, at except.c:2808
Please submit a full bug report.
See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.
make[1]: *** [MsgFmt.o] Error 1
make[1]: Leaving directory `/usr/src/redhat/BUILD/hylafax-cvs-
20000604/faxmail'
make: *** [default] Error 2

I'm using the hylafax source that comes with the latest srpm from 
ftp.hylafax.org: hylafax-4.1beta2-5rh6.src.rpm

For your convenience I'm  attaching a tar.gz with just the files needed to 
reproduce the error.

tar zxvf test-dir.tar.gz
cd test-dir/faxmail
make

Comment 1 giulioo 2000-08-05 14:36:26 UTC
Created attachment 2096 [details]
Files needed to reproduce the compiler error (cd test-dir/faxmail;make)

Comment 2 giulioo 2000-08-15 16:25:49 UTC
I see now that I can compile it ok if I take "-O" out.
However, since rh61 can compile it with "-O", I still thinks it may be 
considered a bug.


Comment 3 Jakub Jelinek 2000-08-21 10:44:39 UTC
Yes, it is considered a bug.
BTW: -O2 and above works as well, it is just -O which ICEs.
I have distilled a small testcase from it which I'm submitting to gcc-patches,
and will try to debug it.

Comment 4 Jakub Jelinek 2000-08-25 13:16:48 UTC
Fixed in gcc-c++-2.96-54


Note You need to log in before you can comment on or make changes to this bug.