> On 22 May 2025, at 14:30, Jon Turney <jon.tur...@dronecode.org.uk> wrote:
> 
> On 22/05/2025 14:15, Gary Liddle via Cygwin wrote:
>> Yesterday I updated Cygwin using  setup-x86_64 to the latest non test 
>> release, and bash —login would no longer start.
>> It turns out bash won’t start for me when a .profile runs a sub-command.
>> This is highlighted by:-
>>      gdb bash
>>      dll cygwin1.dll
>>      b timegm
>>      run
>>      bash-5.2$ echo foo
>>      foo
>>      bash-5.2$ a=`echo foo`
>>      Program received signal SIGSEGV, Segmentation fault.
>>      0x00007ffa5cb3ce31 in timegm () from D:\cygwin64\bin\cygwin1.dll
>>      (gdb)
>>      ...
>>      bash-5.2$ a=`xxx`
>>      bash: xxx: command not found
>>      
>>      Program received signal SIGSEGV, Segmentation fault.
>>      0x00007ffa5cb3ce31 in timegm () from D:\cygwin64\bin\cygwin1.dll
>>      (gdb)
>>      uname -a CYGWIN_NT-10.0-26100 garyl 3.7.0-0.98.gb39b510c1ce6.x86_64 
>> 2025-05-18 15:52 UTC x86_64 Cygwin
> [...]
> 
> Thanks for reporting this.
> 
> 
>> I am at a loss as to how to dig further because gdb doesn’t seem to want to 
>> go near timegm and I can’t see anything in the strace output that seems to 
>> point to anything useful.
> 
> The information in [1] may be helpful in investigating further with gdb.
> 
> [1] https://cygwin.com/faq.html#faq.programming.debugging-cygwin
> 



Thanks for the link, but all the .dbg programmes segfault in:- 
        ie bash.exe.dbg
        mainCRTStartup () at 
/usr/src/debug/cygwin-3.4.10-1/winsup/cygwin/crt0.c:16
for all versions (which might be a clue?).

, but the "set cygwin-exceptions on” produces a bit more.

        gdb bash
        Reading symbols from /usr/lib/debug//usr/bin/bash.exe.dbg...
        (gdb) set cygwin-exceptions on
        (gdb) run
        Starting program: /usr/bin/bash
        bash-5.2$ a=`ls`

        Thread 1 "bash" received signal SIGSEGV, Segmentation fault.
        sigdelayed () at sigfe.s:124
        124             movl    %ebx,0x24(%rsp)
        (gdb)
        (gdb) where
        #0  sigdelayed () at sigfe.s:124
        #1  0x000000010042db0b in wait_for (pid=<optimized out>,
            flags=<error reading variable: Cannot access memory at address 
0xfffffffdffffbf08>)
                    at /usr/src/debug/bash-5.2.21-1/jobs.c:3208
        Backtrace stopped: Cannot access memory at address 0xfffffffdffffbef8
…

Which looks very not helpful.

Regards

Gary

-- 
Problem reports:      https://cygwin.com/problems.html
FAQ:                  https://cygwin.com/faq/
Documentation:        https://cygwin.com/docs.html
Unsubscribe info:     https://cygwin.com/ml/#unsubscribe-simple

Reply via email to