This is the mail archive of the
gdb-prs@sources.redhat.com
mailing list for the GDB project.
testsuite/1463: structs.exp gives different results when run standalone
- From: mec dot gnu at mindspring dot com
- To: gdb-gnats at sources dot redhat dot com
- Date: 22 Nov 2003 22:26:29 -0000
- Subject: testsuite/1463: structs.exp gives different results when run standalone
- Reply-to: mec dot gnu at mindspring dot com
>Number: 1463
>Category: testsuite
>Synopsis: structs.exp gives different results when run standalone
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: unasigned
>State: open
>Class: test-bug
>Submitter-Id: net
>Arrival-Date: Sat Nov 22 22:28:01 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator: mec.gnu@mindspring.com
>Release: gdb HEAD 2003-11-21
>Organization:
>Environment:
target = native, host = i686-pc-linux-gnu, osversion = red-hat-8.0
gdb = HEAD 2003-11-21
gcc = 3.3.2
binutils = 2.14
glibc = 2.2.93-5-rh
gformat = dwarf-2
glevel = 2
>Description:
The new test script gdb.base/structs.exp gives different results when I run it standalone with "RUNTESTFLAGS=structs.exp" versus running the whole test suite.
Here are some counts:
# run the whole test suite
PASS 565
FAIL 0
KPASS 28
KFAIL 16
# run with RUNTESTFLAGS=structs.exp
PASS 593
FAIL 4
KPASS 0
KFAIL 12
This is sensitive to the debug format: it happens with dwarf-2 and does not happen with stabs+.
This is sensitive to compiler version: it happens with gcc 3.3.2 and gcc gcc-3_3-branch 2003-11-21. It does not happen with gcc 2.95.3, gcc 3.2-7-rh, and gcc HEAD 2003-11-21.
This is not sensitive to the binutils version.
I don't know if this is a bug in the test script (something not initialized properly?) or a bug in gdb (failing to initialize something in the inferior?) It really feels like an initialization failure.
This might be sensitive to target, host, osversion, and glibc as well.
>How-To-Repeat:
Get gcc 3.3.2 or some gcc 3.3.X version.
Run the whole test suite with dwarf-2 debug format.
Then run the single script structs.exp with "RUNTESTFLAGS=structs.exp". Compare the results.
>Fix:
It doesn't happen with gcc 2.95.3, and it doesn't happen with gcc HEAD (which will become gcc 3.4), so if we wait a year, the problem will go away on its own.
>Release-Note:
>Audit-Trail:
>Unformatted: