These changes make all the gcc warnings go away.
There are some bug fixes buried in here as well ... 8^)>
With these changes and the pcpqa changes I just posted, QA now passes
all local tests.
Changes committed to git://oss.sgi.com/kenj/pcp.git dev
src/cpan/PMDA/local.c | 7 +++-
src/dbpmda/src/dso.c | 4 +-
src/dbpmda/src/lex.l | 5 +++
src/dbpmda/src/pmda.c | 4 +-
src/dbpmda/src/util.c | 5 +--
src/libpcp/src/connect.c | 6 +--
src/libpcp/src/interp.c | 3 +
src/libpcp/src/logutil.c | 7 ++--
src/libpcp/src/optfetch.c | 2 -
src/libpcp/src/p_pmns.c | 6 +--
src/libpcp/src/p_result.c | 2 -
src/libpcp/src/rtime.c | 2 -
src/libpcp/src/spec.c | 6 +--
src/libpcp/src/units.c | 6 +--
src/libpcp/src/util.c | 29 +++++++++++++----
src/libpcp_gui/src/record.c | 6 ++-
src/libpcp_pmcd/src/trace.c | 2 -
src/libpcp_pmda/src/callback.c | 2 -
src/pcp/pcp.sh | 2 -
src/pmcd/src/config.c | 5 +--
src/pmcd/src/dopdus.c | 10 +++---
src/pmcd/src/pmcd.c | 4 +-
src/pmclient/pmclient.c | 2 -
src/pmdas/cisco/telnet.c | 2 -
src/pmdas/hotproc/src/lex.l | 5 +++
src/pmdas/linux/proc_net_tcp.c | 6 +++
src/pmdas/linux/proc_slabinfo.c | 6 +++
src/pmdas/linux/proc_sys_fs.c | 20 ++++++++----
src/pmdas/lustrecomm/file_single.c | 4 +-
src/pmdas/mailq/mailq.c | 6 +++
src/pmdas/mmv/mmv.c | 26 ++++++++++------
src/pmdas/mounts/mounts.c | 5 ++-
src/pmdas/pmcd/src/pmcd.c | 8 ++--
src/pmdas/process/process.c | 9 ++++-
src/pmdas/roomtemp/mlan/linuxlnk.c | 2 -
src/pmdas/shping/shping.c | 4 +-
src/pmdas/txmon/txrecord.c | 2 -
src/pmdas/weblog/weblog.c | 4 +-
src/pmdumplog/pmdumplog.c | 6 +--
src/pmie/src/README.DEBUG | 19 +++++++++++
src/pmie/src/pmie.c | 9 +++--
src/pmie/src/show.c | 4 --
src/pmie/src/stomp.c | 2 -
src/pmie/src/symbol.c | 2 -
src/pminfo/pminfo.c | 2 -
src/pmlc/actions.c | 12 ++++++-
src/pmlc/lex.l | 5 +++
src/pmlc/pmlc.c | 2 -
src/pmlogextract/lex.l | 5 +++
src/pmlogextract/metriclist.c | 2 -
src/pmlogger/callback.c | 4 +-
src/pmlogger/gram.y | 2 -
src/pmlogger/lex.l | 5 +++
src/pmlogger/pmlogger.c | 60 ++++++++++++++++++++++---------------
src/pmlogsummary/pmlogsummary.c | 5 +--
src/pmns/pmnsdel.c | 8 +++-
src/pmnscomp/pmnscomp.c | 2 -
src/pmprobe/pmprobe.c | 3 +
src/pmproxy/pmproxy.c | 4 +-
src/pmstat/pmstat.c | 4 +-
src/pmstore/pmstore.c | 2 -
src/pmval/pmval.c | 4 +-
62 files changed, 276 insertions(+), 133 deletions(-)
commit d7f66aee0444a4e9fcf1fd96f4cdbfac17532702
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:46:34 2009 +1000
Remove gcc warnings.
Most of these changes are totally benign and not necessary because the
variables are _really_ initialized before being used ... but until
the gcc pixies are able to only warn in situations where there really
is a problem, we have little choice.
There are a couple of very very low probability bugs corrected here also.
commit 0943cb5c9951f0c528f161d719c1d0905bb52d31
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:45:22 2009 +1000
Check return value from dup().
commit 8d88452a93fc5a7c5c48aa85e5261ebffa248240
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:44:29 2009 +1000
Check return value from chown().
commit bb90664441edb892af0c37cd64d24597328984ab
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:41:55 2009 +1000
Assorted minor changes.
Add initialization to pander to gcc warnings.
Use mkstemp() if available in preference to tmpnam().
Fix formatting of messages via pmconfirm when using record mode from
gui tools.
commit 07c987936049b84ae5d0c18af396962d9fb1da97
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:40:08 2009 +1000
Fix unlikely off-by-one error in strncat() use.
commit 7d85c046adf8adfba4d97a464b6a3780c48e5a03
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:38:50 2009 +1000
Check write() return value and minor initialization to remove gcc warning.
commit 007e35f797d7f16be67a0997f5b05593bc863ba1
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:37:42 2009 +1000
Check return value from dup().
commit c70d2a9fe1e3c1b4d7faefa44389f288c9be483b
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:36:54 2009 +1000
Check return value from fgets().
commit 44e6aeb91d39f94c1ec8c912fae4576e74cdec0e
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:36:00 2009 +1000
Check return value from fgets().
commit 4430977a324238658355a287a579df2b852718dc
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:34:58 2009 +1000
Use mkstemp() if available.
Safer replacement for tempnam().
commit c70c05f052e6025d72833ea6d120ff8faaf9554f
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:33:12 2009 +1000
Check getcwd() return value.
commit d790ad583eacad07f704a64250f80613257de588
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:32:07 2009 +1000
Cleanup fgets() and fscanf() use.
Check return values to make sure we're getting what we expect as we
parse these ascii files.
commit a9eced25ce3bd752f4ef057f7767d98c4a850e36
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:28:43 2009 +1000
Check return value from dup().
commit 3c5a462c10f277e3e088304d9121311eea52f884
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:27:31 2009 +1000
Check getcwd() return value.
commit 5716ca025d9a19ccbe0d02c00d2b7702b1577516
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:26:09 2009 +1000
gcc cleanup and better system call checking.
Check the return status from read(), write() and pipe().
commit 39dfce0c0cd770a76e0e14bdd6e06f57f0721c4b
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:22:53 2009 +1000
Use xterm instead of xwsh.
The watch command was broken, this makes it work again. Also cleaned
up the use of system(), specifically check the return value.
commit 2b16263f80d6c63ff25f189bbd780a5b82c44078
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 20:18:57 2009 +1000
Hack to workaround gcc warnings from flex generated code.
Don't even ask about this ... it is obscene in the extreme, but until
the flex pixies are generating C code that does not cause gcc to babble
on with warnings, this is what you have to do.
commit ec0767c332236fea2fc4369306305924c9255fc1
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 07:18:51 2009 +1000
Reinstate optional pmFetch warnings.
When the warnings were changed to be off by default, the plumbing to allow
the previous default to be enabled (-v) was incomplete ... this was causing
QA/363 to fail, besides being just wrong.
commit 8b03b14be1369d42ae964b5bfd3a7f75fd3372d2
Author: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Sun May 17 07:10:53 2009 +1000
Cosmetic change.
Removal of simabi introduced a very minor white space issue that was
causing QA/507 to fail.
|