* [ANN] perl5.5.660 @ 2000-02-26 19:29 Charles S. Wilson 2000-02-27 1:25 ` Corinna Vinschen 2000-03-01 22:17 ` Charles Wilson 0 siblings, 2 replies; 14+ messages in thread From: Charles S. Wilson @ 2000-02-26 19:29 UTC (permalink / raw) To: cygwin; +Cc: EFifer perl5.5.660 is now available at http://cygutils.netpedia.net/V1.1/perl-5.5.660/ You will also find a few precompiled modules at that location. This might be the last beta before perl5.6 is released, so please bang on it if you've the inclination. I've found one problem already that I couldn't figure out: attempting to execute any routine from the libnet module causes perl to dump core immediately. If anybody's got any ideas about how to fix that, please let me know. perl5.5.660 was built using the following external libraries from the cygutils archive: libcrypt cygipc-1.05 db-2.7.7 gdbm-1.8.0-1 and with the following environment settings. ntsec was not enabled. CYGWIN=binmode tty ntea built and installed by a non-administrator user Anyway, here are the perl harness results: Failed Test Status Wstat Total Fail Failed List of failed ------------------------------------------------------------------------------- io/pipe.t 1 256 15 12 80.00% 7-15 lib/glob-basic. 9 7 77.78% 3-9 lib/io_pipe.t 10 2 20.00% 9-10 lib/io_poll.t 8 1 12.50% 4 lib/odbm.t 64 1 1.56% 2 lib/posix.t 27 3 11.11% 8-10 op/grent.t 1 1 100.00% 1 op/magic.t 35 2 5.71% 3, 5 op/taint.t 149 11 7.38% 1, 3, 31, 142-149 pragma/warnings 348 1 0.29% 117 4 tests and 57 subtests skipped. Failed 10/234 test scripts, 95.73% okay. 25/11679 subtests failed, 99.79% okay. --Chuck ---------------------------------------------------------- ---------------------------------------------------------- FWIW, I also built and tested 5.5.660 with ntsec ON. Here are those results Failed Test Status Wstat Total Fail Failed List of failed ------------------------------------------------------------------------------- base/rs.t 13 3328 14 14 100.00% 1-14 io/pipe.t 1 256 15 12 80.00% 7-15 lib/dirhand.t 5 1 20.00% 3 lib/filecache.t 255 65280 1 1 100.00% 1 lib/filepath.t 255 65280 4 4 100.00% 1-4 lib/glob-basic. 9 1 11.11% 2 lib/io_dir.t 10 1 10.00% 3 lib/io_pipe.t 10 2 20.00% 9-10 lib/io_poll.t 8 1 12.50% 4 lib/io_unix.t 0 15 5 4 80.00% 2-5 lib/ipc_sysv.t 16 16 100.00% 1-16 lib/ndbm.t 13 3328 64 50 78.12% 1-2, 4, 9, 19-64 lib/odbm.t 64 62 96.88% 3-64 lib/posix.t 27 3 11.11% 8-10 op/grent.t 1 1 100.00% 1 op/magic.t 35 2 5.71% 3, 5 op/stat.t 58 8 13.79% 18-21, 26-27, 57-58 op/taint.t 149 3 2.01% 1, 3, 31 pragma/warnings 348 1 0.29% 117 4 tests and 44 subtests skipped. Failed 19/234 test scripts, 91.88% okay. 171/11679 subtests failed, 98.54% okay. -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [ANN] perl5.5.660 2000-02-26 19:29 [ANN] perl5.5.660 Charles S. Wilson @ 2000-02-27 1:25 ` Corinna Vinschen 2000-02-27 16:07 ` Charles S. Wilson 2000-02-28 0:56 ` Corinna Vinschen 2000-03-01 22:17 ` Charles Wilson 1 sibling, 2 replies; 14+ messages in thread From: Corinna Vinschen @ 2000-02-27 1:25 UTC (permalink / raw) To: Charles S. Wilson; +Cc: cygwin "Charles S. Wilson" wrote: > > perl5.5.660 is now available at > http://cygutils.netpedia.net/V1.1/perl-5.5.660/ You will also find a few > precompiled modules at that location. > [...] > FWIW, I also built and tested 5.5.660 with ntsec ON. Here are those > results > [...] > Failed 19/234 test scripts, 91.88% okay. 171/11679 subtests failed, > 98.54% okay. Wow, I'm glad that it's only twice as much than with ntsec OFF. I'm not familar with perl but I will try to reproduce the test results for getting more information why these additional tests are failing. Corinna -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [ANN] perl5.5.660 2000-02-27 1:25 ` Corinna Vinschen @ 2000-02-27 16:07 ` Charles S. Wilson 2000-02-28 0:56 ` Corinna Vinschen 1 sibling, 0 replies; 14+ messages in thread From: Charles S. Wilson @ 2000-02-27 16:07 UTC (permalink / raw) To: Corinna Vinschen; +Cc: cygwin Corinna Vinschen wrote: > > "Charles S. Wilson" wrote: > > > > perl5.5.660 is now available at > > http://cygutils.netpedia.net/V1.1/perl-5.5.660/ You will also find a few > > precompiled modules at that location. > > [...] > > FWIW, I also built and tested 5.5.660 with ntsec ON. Here are those > > results > > [...] > > Failed 19/234 test scripts, 91.88% okay. 171/11679 subtests failed, > > 98.54% okay. > > Wow, I'm glad that it's only twice as much than with ntsec OFF. > > I'm not familar with perl but I will try to reproduce the test > results for getting more information why these additional tests > are failing. Even with my recent fixes to gdbm and cygipc, I still think that many of these errors are in the external libraries and not in perl itself. For instance, the complete failure of all ipc-sysv regression tests, and the failures in the lib/odbm and lib/ndbm tests. OTOH, there's plenty of ntsec-triggered failures in parts of perl that do not rely on external libraries -- the base/rs tests, lib/dirhand.t, lib/filecache.t, lib/filepath.t, etc. OTO-O-H, there's a few tests that failed *without* ntsec, and worked okay *with* ntsec -- lib/glob-basic. I'm going to try a let my hair grow back before looking at this again, though. I pulled out way too much of it... --Chuck -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [ANN] perl5.5.660 2000-02-27 1:25 ` Corinna Vinschen 2000-02-27 16:07 ` Charles S. Wilson @ 2000-02-28 0:56 ` Corinna Vinschen 1 sibling, 0 replies; 14+ messages in thread From: Corinna Vinschen @ 2000-02-28 0:56 UTC (permalink / raw) To: cygwin > "Charles S. Wilson" wrote: > > perl5.5.660 is now available at > http://cygutils.netpedia.net/V1.1/perl-5.5.660/ You will also find a few > precompiled modules at that location. > [...] > FWIW, I also built and tested 5.5.660 with ntsec ON. Here are those > results > [...] > Failed 19/234 test scripts, 91.88% okay. 171/11679 subtests failed, > 98.54% okay. Hi Charles, I can't reproduce your results exactly. In my environment CYGWIN_NT-5.0 CORINNA 1.1.0(0.16/3/2) 2000-02-24 09:43:09 i686 unknown CYGWIN=binmode nontea ntsec tty I got the following _differing_ results as non-admin user: Failed Test Status Wstat Total Fail Failed List of failed --------------------------------------------------------------------- base/rs.t OK io/pipe.t 15 2 13.33% 11-12 lib/dirhand.t OK lib/filecache.t OK lib/filepath.t 255 65280 4 3 75.00% 1, 3-4 lib/io_pipe.t OK lib/io_sock.t 22 5632 14 9 64.29% 6-14 lib/io_unix.t 2 512 5 5 100.00% 1-5 lib/ndbm.t OK op/magic.t 13 3328 35 17 48.57% 19-35 4 tests and 44 subtests skipped. Failed 14/234 test scripts, 94.02% okay. 130/11679 subtests failed, 98.89% okay. I found that op/stat.t must fail, because normal users doesn't have the permission to chown. This doesn't fail as admin user. So this one can be ignored. I think we could get a better result but as I mentioned before I'm not familar with perl. Nevertheless, as far as ntsec is concerned I would like to help. Please contact me when you have found more information. Corinna -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [ANN] perl5.5.660 2000-02-26 19:29 [ANN] perl5.5.660 Charles S. Wilson 2000-02-27 1:25 ` Corinna Vinschen @ 2000-03-01 22:17 ` Charles Wilson 2000-03-01 22:53 ` Charles Wilson 2000-03-02 10:14 ` Still 100% usage and weird problems with recent snapshots? Chris Faylor 1 sibling, 2 replies; 14+ messages in thread From: Charles Wilson @ 2000-03-01 22:17 UTC (permalink / raw) To: cygwin, EFifer FWIW, the problems described below were fixed with the Feb 28 snapshot. That includes the 100% usage problem and the weird configure failures when building. --Chuck "Charles S. Wilson" wrote: > > perl5.5.660 is now available at > http://cygutils.netpedia.net/V1.1/perl-5.5.660/ You will also find a few > precompiled modules at that location. > > This might be the last beta before perl5.6 is released, so please bang > on it if you've the inclination. I've found one problem already that I > couldn't figure out: attempting to execute any routine from the libnet > module causes perl to dump core immediately. If anybody's got any ideas > about how to fix that, please let me know. > > perl5.5.660 was built using the following external libraries from the > cygutils archive: > libcrypt > cygipc-1.05 > db-2.7.7 > gdbm-1.8.0-1 > and with the following environment settings. ntsec was not enabled. > CYGWIN=binmode tty ntea > built and installed by a non-administrator user > > Anyway, here are the perl harness results: > > Failed Test Status Wstat Total Fail Failed List of failed > ------------------------------------------------------------------------------- > io/pipe.t 1 256 15 12 80.00% 7-15 > lib/glob-basic. 9 7 77.78% 3-9 > lib/io_pipe.t 10 2 20.00% 9-10 > lib/io_poll.t 8 1 12.50% 4 > lib/odbm.t 64 1 1.56% 2 > lib/posix.t 27 3 11.11% 8-10 > op/grent.t 1 1 100.00% 1 > op/magic.t 35 2 5.71% 3, 5 > op/taint.t 149 11 7.38% 1, 3, 31, 142-149 > pragma/warnings 348 1 0.29% 117 > 4 tests and 57 subtests skipped. > Failed 10/234 test scripts, 95.73% okay. 25/11679 subtests failed, > 99.79% okay. > > --Chuck > > ---------------------------------------------------------- > ---------------------------------------------------------- > FWIW, I also built and tested 5.5.660 with ntsec ON. Here are those > results > > Failed Test Status Wstat Total Fail Failed List of failed > ------------------------------------------------------------------------------- > base/rs.t 13 3328 14 14 100.00% 1-14 > io/pipe.t 1 256 15 12 80.00% 7-15 > lib/dirhand.t 5 1 20.00% 3 > lib/filecache.t 255 65280 1 1 100.00% 1 > lib/filepath.t 255 65280 4 4 100.00% 1-4 > lib/glob-basic. 9 1 11.11% 2 > lib/io_dir.t 10 1 10.00% 3 > lib/io_pipe.t 10 2 20.00% 9-10 > lib/io_poll.t 8 1 12.50% 4 > lib/io_unix.t 0 15 5 4 80.00% 2-5 > lib/ipc_sysv.t 16 16 100.00% 1-16 > lib/ndbm.t 13 3328 64 50 78.12% 1-2, 4, 9, 19-64 > lib/odbm.t 64 62 96.88% 3-64 > lib/posix.t 27 3 11.11% 8-10 > op/grent.t 1 1 100.00% 1 > op/magic.t 35 2 5.71% 3, 5 > op/stat.t 58 8 13.79% 18-21, 26-27, 57-58 > op/taint.t 149 3 2.01% 1, 3, 31 > pragma/warnings 348 1 0.29% 117 > 4 tests and 44 subtests skipped. > Failed 19/234 test scripts, 91.88% okay. 171/11679 subtests failed, > 98.54% okay. -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [ANN] perl5.5.660 2000-03-01 22:17 ` Charles Wilson @ 2000-03-01 22:53 ` Charles Wilson 2000-03-02 10:14 ` Chris Faylor 2000-03-02 10:14 ` Still 100% usage and weird problems with recent snapshots? Chris Faylor 1 sibling, 1 reply; 14+ messages in thread From: Charles Wilson @ 2000-03-01 22:53 UTC (permalink / raw) To: cygwin, EFifer Geez. let's try this again. I wanted to send replies to two different messages on this list. Naturally, I attached message B to the reply meant for message A. Here's the REAL reply that should go in the perl5.5.660 thread. I installed the 20000227 cygwin-inst and the 20000228 cygwin1.dll, and re-ran the perl harness test. The new results are much improved: Same binary as before, tested under 'CYGWIN=binmode tty ntea' Failed Test Status Wstat Total Fail Failed List of failed ------------------------------------------------------------------------------- lib/io_poll.t 8 1 12.50% 4 lib/odbm.t 64 1 1.56% 2 op/grent.t 1 1 100.00% 1 op/taint.t 149 3 2.01% 1, 3, 31 pragma/warnings 348 1 0.29% 117 4 tests and 58 subtests skipped. Failed 5/234 test scripts, 97.86% okay. 7/11679 subtests failed, 99.94% okay. Same binary, but this time with CYGWIN='binmode tty ntea ntsec' Failed Test Status Wstat Total Fail Failed List of failed ------------------------------------------------------------------------------- lib/filepath.t 255 65280 4 3 75.00% 1, 3-4 lib/glob-basic. 9 1 11.11% 2 lib/io_dir.t 10 1 10.00% 3 lib/io_poll.t 8 1 12.50% 4 lib/io_unix.t 0 15 5 4 80.00% 2-5 lib/odbm.t 13 3328 64 62 96.88% 2, 4-64 op/grent.t 1 1 100.00% 1 op/stat.t 58 8 13.79% 18-21, 26-27, 57-58 op/taint.t 149 3 2.01% 1, 3, 31 pragma/warnings 348 1 0.29% 117 4 tests and 58 subtests skipped. Failed 10/234 test scripts, 95.73% okay. 85/11679 subtests failed, 99.27% okay. --Chuck On Thu, 2 Mar 2000, Charles Wilson wrote: > FWIW, the problems described below were fixed with the Feb 28 snapshot. > That includes the 100% usage problem and the weird configure failures > when building. > > --Chuck > > "Charles S. Wilson" wrote: > > > > perl5.5.660 is now available at > > http://cygutils.netpedia.net/V1.1/perl-5.5.660/ You will also find a few > > precompiled modules at that location. > > > > This might be the last beta before perl5.6 is released, so please bang > > on it if you've the inclination. I've found one problem already that I > > couldn't figure out: attempting to execute any routine from the libnet > > module causes perl to dump core immediately. If anybody's got any ideas > > about how to fix that, please let me know. > > > > perl5.5.660 was built using the following external libraries from the > > cygutils archive: > > libcrypt > > cygipc-1.05 > > db-2.7.7 > > gdbm-1.8.0-1 > > and with the following environment settings. ntsec was not enabled. > > CYGWIN=binmode tty ntea > > built and installed by a non-administrator user > > > > Anyway, here are the perl harness results: > > > > Failed Test Status Wstat Total Fail Failed List of failed > > ------------------------------------------------------------------------------- > > io/pipe.t 1 256 15 12 80.00% 7-15 > > lib/glob-basic. 9 7 77.78% 3-9 > > lib/io_pipe.t 10 2 20.00% 9-10 > > lib/io_poll.t 8 1 12.50% 4 > > lib/odbm.t 64 1 1.56% 2 > > lib/posix.t 27 3 11.11% 8-10 > > op/grent.t 1 1 100.00% 1 > > op/magic.t 35 2 5.71% 3, 5 > > op/taint.t 149 11 7.38% 1, 3, 31, 142-149 > > pragma/warnings 348 1 0.29% 117 > > 4 tests and 57 subtests skipped. > > Failed 10/234 test scripts, 95.73% okay. 25/11679 subtests failed, > > 99.79% okay. > > > > --Chuck > > > > ---------------------------------------------------------- > > ---------------------------------------------------------- > > FWIW, I also built and tested 5.5.660 with ntsec ON. Here are those > > results > > > > Failed Test Status Wstat Total Fail Failed List of failed > > ------------------------------------------------------------------------------- > > base/rs.t 13 3328 14 14 100.00% 1-14 > > io/pipe.t 1 256 15 12 80.00% 7-15 > > lib/dirhand.t 5 1 20.00% 3 > > lib/filecache.t 255 65280 1 1 100.00% 1 > > lib/filepath.t 255 65280 4 4 100.00% 1-4 > > lib/glob-basic. 9 1 11.11% 2 > > lib/io_dir.t 10 1 10.00% 3 > > lib/io_pipe.t 10 2 20.00% 9-10 > > lib/io_poll.t 8 1 12.50% 4 > > lib/io_unix.t 0 15 5 4 80.00% 2-5 > > lib/ipc_sysv.t 16 16 100.00% 1-16 > > lib/ndbm.t 13 3328 64 50 78.12% 1-2, 4, 9, 19-64 > > lib/odbm.t 64 62 96.88% 3-64 > > lib/posix.t 27 3 11.11% 8-10 > > op/grent.t 1 1 100.00% 1 > > op/magic.t 35 2 5.71% 3, 5 > > op/stat.t 58 8 13.79% 18-21, 26-27, 57-58 > > op/taint.t 149 3 2.01% 1, 3, 31 > > pragma/warnings 348 1 0.29% 117 > > 4 tests and 44 subtests skipped. > > Failed 19/234 test scripts, 91.88% okay. 171/11679 subtests failed, > > 98.54% okay. > > -- > Want to unsubscribe from this list? > Send a message to cygwin-unsubscribe@sourceware.cygnus.com > > -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [ANN] perl5.5.660 2000-03-01 22:53 ` Charles Wilson @ 2000-03-02 10:14 ` Chris Faylor 2000-03-02 13:26 ` Corinna Vinschen 0 siblings, 1 reply; 14+ messages in thread From: Chris Faylor @ 2000-03-02 10:14 UTC (permalink / raw) To: cygwin; +Cc: Corinna Vinschen On Thu, Mar 02, 2000 at 01:53:31AM -0500, Charles Wilson wrote: >Same binary, but this time with CYGWIN='binmode tty ntea ntsec' Corinna, shouldn't ntea and ntsec be mutually exclusive? Should we change the DLL so that only one or the other can be active? cgf -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [ANN] perl5.5.660 2000-03-02 10:14 ` Chris Faylor @ 2000-03-02 13:26 ` Corinna Vinschen 0 siblings, 0 replies; 14+ messages in thread From: Corinna Vinschen @ 2000-03-02 13:26 UTC (permalink / raw) To: cygwin Chris Faylor wrote: > > On Thu, Mar 02, 2000 at 01:53:31AM -0500, Charles Wilson wrote: > >Same binary, but this time with CYGWIN='binmode tty ntea ntsec' > > Corinna, shouldn't ntea and ntsec be mutually exclusive? Should we > change the DLL so that only one or the other can be active? Maybe that isn't a good idea. Some people are using FAT and NTFS and if you have to use ntsec without ntea you don't have a chance to simulate permissions on the FAT partitions. Personally I would only change the set/get_file_attributes() so that ntea is deactivated on NTFS partitions if ntsec is set. Corinna -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Still 100% usage and weird problems with recent snapshots? 2000-03-01 22:17 ` Charles Wilson 2000-03-01 22:53 ` Charles Wilson @ 2000-03-02 10:14 ` Chris Faylor 2000-03-02 12:05 ` Dr. Volker Zell 2000-03-02 13:00 ` Glenn Spell 1 sibling, 2 replies; 14+ messages in thread From: Chris Faylor @ 2000-03-02 10:14 UTC (permalink / raw) To: cygwin On Thu, Mar 02, 2000 at 01:17:39AM -0500, Charles Wilson wrote: >FWIW, the problems described below were fixed with the Feb 28 snapshot. >That includes the 100% usage problem and the weird configure failures >when building. I still have one person reporting problems with cygwin but I am having no luck duplicating them and the strace output that he's provided does not make any sense. Has the recent snapshot solved the problem for everyone else or are there still problems? I would like to get this resolved since one of these snapshots is going to become the official Cygwin 1.1.1 beta that will show up in the pending net release which we hope to have available in the next couple of weeks. cgf -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: Still 100% usage and weird problems with recent snapshots? 2000-03-02 10:14 ` Still 100% usage and weird problems with recent snapshots? Chris Faylor @ 2000-03-02 12:05 ` Dr. Volker Zell 2000-03-02 13:00 ` Glenn Spell 1 sibling, 0 replies; 14+ messages in thread From: Dr. Volker Zell @ 2000-03-02 12:05 UTC (permalink / raw) To: cygwin >>>>> "Chris" == Chris Faylor <cgf@cygnus.com> writes: Chris> Has the recent snapshot solved the problem for everyone else or are Chris> there still problems? Chris> I would like to get this resolved since one of these snapshots is going Chris> to become the official Cygwin 1.1.1 beta that will show up in the Chris> pending net release which we hope to have available in the next couple Chris> of weeks. Here my setup and testcases: System: WinNT 4.0/SP3 Cygwin B20.1 with egcs-2.91.66 - cygwin1.dll - cygwin1-20000228.dll - everything mounted binary, - CYGWIN = tty title binmode ntea nontsec Software: Xemacs-21.1.2 - http://www.xemacs.org/ inetutils-1.3.2 - ftp://ftp.gnu.org/pub/gnu (with patch from Sergey) ssh-1.2.26 - ftp://ftp.cs.hut.fi/pub/ssh (with patch from Sergey) sendmail-8.9.3 - http://www.sendmail.org/ apache-1.3.6 - http://www.apache.org/ o No CPU problems. I'm running inetd, sshd, apache and sendmail as NT services via invoker-1.4 - http://innovation.idetix.com/invoker/ and Xemacs on top and my CPU time is about 14% and almost consumed by taskmgr.exe o Two things I noticed (Chris you alreday fixed this in a snapshot before) 1) When I call ftp and type in any commands the commands don't echo on the screen only after typing <RETURN>. The same thing happens if I try sending me an email via sendmail like this: vzell:/tmp> sendmail vzell This is a test string <---- I only see what I have typed after pressing RETURN here <---- This is CTRL-d vzell:/tmp> 2) When invoking a bash shell inside from Xemacs and then trying to telnet to my own machine from this session I get the following behaviour. vzell:~> vzell:~> telnet vzell Trying 140.84.39.166... Connected to vzell.de.oracle.com. Escape character is '^]'. CYGWIN_NT-4.0 1.1.0(0.17/3/2) (vzell) (tty6) login: vzell vzell <----- echoed Password: vzell:~> pwd pwd <----- echoed /users/vzell vzell:~> vzell:~> So in this case everything I type in is echoed to thze sreen twice, as soone as I type it in and AFTER pressing RETURN. Unitl now I used the cygwin-inst-19990929 snapshot which was perfectly stable for me. Also it doesn't show the above anomalies. Ciao Volker -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: Still 100% usage and weird problems with recent snapshots? 2000-03-02 10:14 ` Still 100% usage and weird problems with recent snapshots? Chris Faylor 2000-03-02 12:05 ` Dr. Volker Zell @ 2000-03-02 13:00 ` Glenn Spell 2000-03-02 13:06 ` Chris Faylor 1 sibling, 1 reply; 14+ messages in thread From: Glenn Spell @ 2000-03-02 13:00 UTC (permalink / raw) To: cygwin On 2 Mar 2000 around 1:13PM (-0500) Chris Faylor wrote: > Has the recent snapshot solved the problem for everyone else or are > there still problems? As I reported earlier, the bash I was using up until recently will not run with any snapshots after the 99-12-04. The only thing I am aware of about this bash-2.03 that won't run is that I compiled it with gcc-2.95 (I have not yet updated gcc from 2.95.) My work-around in order to use the latest snapshots was to revert to a bash-2.03 that I compiled with egcs-1.1.1. I'm not suggesting anything... just reporting observations. I offer this just in case it may be related to anything else that's going on with the latest snapshots. Following is all the info I can come up with since this bash does not run far enough to get any strace output. When starting this bash via Insight, the Source Window moves to the file "shell.c" function "main", at the following line (*): 305 /* Catch early SIGINTs. */ * 306 code = setjmp (top_level); - 307 if (code) - 308 exit (2); - 0x401208 <main>: push %ebp - 0x401209 <main+1>: mov %ebp,%esp - 0x40120b <main+3>: sub %esp,28 - 0x40120e <main+6>: push %edi - 0x40120f <main+7>: push %esi - 0x401210 <main+8>: push %ebx - 0x401211 <main+9>: call 0x462360 <__main> * 0x401216 <main+14>: mov %ds:0x46a8b0,0x1 - 0x401220 <main+24>: add %esp,-4 - 0x401223 <main+27>: push 0x46a8b4 - 0x401228 <main+32>: push 0 - 0x40122a <main+34>: push 0 - 0x40122c <main+36>: call 0x462450 <sigprocmask> - 0x401231 <main+41>: add %esp,-12 - 0x401234 <main+44>: push 0x46a7e0 - 0x401239 <main+49>: call 0x462448 <setjmp> - 0x40123e <main+54>: add %esp,32 - 0x401241 <main+57>: test %eax,%eax When I start the program running I immediately get a segfault and the line number does not change in the Source Window. Then, the source code disappears and Insight shows the current function as "size_of_stack_frame" and prints this lone message in the Insight Source Window: "Unable to Read Instructions at 0x61041a16" The Insight Memory Window displays the following: 0 4 8 C 0x61041a16 {0x74107339} {0x0cec8358} {0x5008438d}{0x00bae1e8} 0x74107339 N/A N/A N/A N/A View Thread List (Processes) lists the following: 2 thread -177397.0xfffc280f 0xbff9638a in _size_of_stack_reserve__ () 1 thread -177397.0xfffdcfc3 0x61041a16 in _size_of_stack_reserve__ () At another invocation, Windows reports the following: BASH caused an invalid page fault in module CYGWIN1.DLL at 0157:61041a16. Registers: EAX=fffdffab CS=0157 EIP=61041a16 EFLGS=00010282 EBX=00000000 SS=015f ESP=0259fba4 EBP=0259fbbc ECX=ffffffff DS=015f ESI=fffdffab FS=2007 EDX=ffffffff ES=015f EDI=ffffffff GS=0000 Bytes at CS:EIP: 39 73 10 74 58 83 ec 0c 8d 43 08 50 e8 e1 ba 00 I hope this helps 'cause that's all I know. -glenn -- ________________________________________ _ _____ ) )_ _ (__\____o /_/_ | ) Glenn Spell <glenn@gs.fay.nc.us> ) >-----._/_/__]> )________________________________________) `0 | -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: Still 100% usage and weird problems with recent snapshots? 2000-03-02 13:00 ` Glenn Spell @ 2000-03-02 13:06 ` Chris Faylor 2000-03-02 13:52 ` Glenn Spell 0 siblings, 1 reply; 14+ messages in thread From: Chris Faylor @ 2000-03-02 13:06 UTC (permalink / raw) To: cygwin On Thu, Mar 02, 2000 at 04:00:10PM -0500, Glenn Spell wrote: >On 2 Mar 2000 around 1:13PM (-0500) Chris Faylor wrote: >> Has the recent snapshot solved the problem for everyone else or are >> there still problems? > >As I reported earlier, the bash I was using up until recently will >not run with any snapshots after the 99-12-04. The only thing I am >aware of about this bash-2.03 that won't run is that I compiled it >with gcc-2.95 (I have not yet updated gcc from 2.95.) You reported that you were having problems but I don't recall anything that indicated you still had problems with the newest snapshot. It is not conceivable that gdb would report that you are stopped at the code that you show below and that no strace output shows up. How are you running the strace program? You should do something like this: strace -osomefile bash cgf -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: Still 100% usage and weird problems with recent snapshots? 2000-03-02 13:06 ` Chris Faylor @ 2000-03-02 13:52 ` Glenn Spell 2000-03-02 14:23 ` Chris Faylor 0 siblings, 1 reply; 14+ messages in thread From: Glenn Spell @ 2000-03-02 13:52 UTC (permalink / raw) To: cygwin On 2 Mar 2000 around 4:06PM (-0500) Chris Faylor wrote: > On Thu, Mar 02, 2000 at 04:00:10PM -0500, Glenn Spell wrote: > > > >As I reported earlier, the bash I was using up until recently will > >not run with any snapshots after the 99-12-04. The only thing I am > >aware of about this bash-2.03 that won't run is that I compiled it > >with gcc-2.95 (I have not yet updated gcc from 2.95.) > > You reported that you were having problems but I don't recall anything > that indicated you still had problems with the newest snapshot. I figure that I'll need to upgrade gcc, rebuild bash, and then see if there is any problem... and I doubt if there will be or others would be seeing it also. I don't have plans at the moment to upgrade gcc. I haven't filed any other reports because this is not really a problem that I need solved. I'm reporting now just in case it can help you trace something that may be affecting others in different ways. > It is not conceivable that gdb would report that you are stopped at > the code that you show below and that no strace output shows up. It is for me. ;-) > How are you running the strace program? I move to the directory where I'm keeping the bash that doesn't work, a directory that is not in my path, and I strace ./bash and nothing happens. "^C" gets me out. > You should do something like this: > > strace -osomefile bash strace -otestout ./bash hangs... nothing happens. The file "testout" is created. I CTRL-C and end the process. Checking "testout" with pfe shows an empty file. Checking it with Hex Edit shows only "00000000 " (without the quotes) (This is using the strace and DLL from DJ's latest preview. -glenn -- ________________________________________ _ _____ ) )_ _ (__\____o /_/_ | ) Glenn Spell <glenn@gs.fay.nc.us> ) >-----._/_/__]> )________________________________________) `0 | -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: Still 100% usage and weird problems with recent snapshots? 2000-03-02 13:52 ` Glenn Spell @ 2000-03-02 14:23 ` Chris Faylor 0 siblings, 0 replies; 14+ messages in thread From: Chris Faylor @ 2000-03-02 14:23 UTC (permalink / raw) To: cygwin On Thu, Mar 02, 2000 at 04:52:12PM -0500, Glenn Spell wrote: >I CTRL-C and end the process. Checking "testout" with pfe shows >an empty file. Checking it with Hex Edit shows only > > "00000000 " > >(without the quotes) >(This is using the strace and DLL from DJ's latest preview. Ok, then this definitely is a problem with the way bash is linked then. The program is obviously not starting in the correct place and cygwin's initialization code is not being processed. This is either a linker or a gcc issue. cgf -- Want to unsubscribe from this list? Send a message to cygwin-unsubscribe@sourceware.cygnus.com ^ permalink raw reply [flat|nested] 14+ messages in thread
end of thread, other threads:[~2000-03-02 14:23 UTC | newest] Thread overview: 14+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2000-02-26 19:29 [ANN] perl5.5.660 Charles S. Wilson 2000-02-27 1:25 ` Corinna Vinschen 2000-02-27 16:07 ` Charles S. Wilson 2000-02-28 0:56 ` Corinna Vinschen 2000-03-01 22:17 ` Charles Wilson 2000-03-01 22:53 ` Charles Wilson 2000-03-02 10:14 ` Chris Faylor 2000-03-02 13:26 ` Corinna Vinschen 2000-03-02 10:14 ` Still 100% usage and weird problems with recent snapshots? Chris Faylor 2000-03-02 12:05 ` Dr. Volker Zell 2000-03-02 13:00 ` Glenn Spell 2000-03-02 13:06 ` Chris Faylor 2000-03-02 13:52 ` Glenn Spell 2000-03-02 14:23 ` Chris Faylor
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).