At 04:52 PM 11/29/2004, you wrote: >On Mon, Nov 29, 2004 at 03:18:05PM -0500, Larry Hall wrote: >> At 02:59 PM 11/29/2004, you wrote: >> >The postinstall scripts are all cygwin shell scripts and thus require >> >a functional 'sh' to do anything. And, as with all other cygwin >> >programs, 'sh' just exits without doing anything. >> >> >> Yep, so I'd recommend starting there. Check the Windows properties here >> to see who owns the file and who has permissions to execute it, etc. Also, >> you could do a strace on 'sh' to see if anything interesting shows up from >> that. Hopefully you'll see something interesting as a result. > >File permissions appear ok as far as I can tell. The 'strace' output >seems consistent with no permission problems; I'm able to execute all >commands, cygwin1.dll is found, and things run for a short while >before dying.
But with no output, right? No dump file, nothing? >'strace sh.exe' looks like any other command I try: > > ... > 564 105230 [main] bash 3892 cygheap_user::ontherange: what 2, pw 0xA0510C8 > 224 105454 [main] bash 3892 cygheap_user::ontherange: HOME is already in > the environment /cygdrive/h > 756 106210 [main] bash 3892 sigproc_init: process/signal handling > enabled(1001) > 365 106575 [main] bash 3892 __cygwin_lock_lock: threadcount 1. not locking > 304 106879 [main] bash 3892 __cygwin_lock_unlock: threadcount 1. not > unlocking This isn't the strace of 'sh.exe'. It's the strace of 'bash.exe'. Can you try 'sh.exe' and see if it's different. FWIW, the lines you provided above I can find in my strace of 'bash.exe'. Mine goes on however. >and I'm back to the Windows command prompt at this point. I thought >maybe there could be something odd in my environment, but deleting all >entries has no effect: I still see the exact same behavior with an >environment that consists, in its entirety, of "PATH=.". > >By hunting around on the internet, I was able to find some old cygwin >versions to test. With an old version of 'mount', I get normal output >with the 1.5.5-1 version of cygwin1.dll, but 1.5.6-1 behaves the same >as 1.5.11-1 and 1.5.12-1 (no output, exits following sigproc_init). > >I see that the 1.5.6-1 release changed signal processing which seems >suggestive given that the last substantive output of 'strace' comes >when signal handling is enabled. If you have 1.5.5 now that works for you, you can use that to bootstrap your build and get a debug version of the current DLL to look at and debug into. That might be your best bet. You can definitely get more context that way, if you're so inclined. -- Larry Hall http://www.rfk.com RFK Partners, Inc. (508) 893-9779 - RFK Office 838 Washington Street (508) 893-9889 - FAX Holliston, MA 01746 -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
