Re: Question about executable startup failure

[Brian Inglis <Brian dot Inglis at SystematicSw dot ab dot ca>]

On 2019-11-13 23:48, Lavrentiev, Anton (NIH/NLM/NCBI) [C] via cygwin wrote:
> Here's a situation: I have a binary that was built (from a C source code)
> with Cygwin 3.0.7 but "accidentally" used with much older Cygwin 2.11.2.
> 
> The binary won't actually launch.  Instead, it most uneventfully (silently, 
> no crash - no drama) exits with an exit code of 127, meaning "command not 
> found". The same exit code would have resulted from any mistyped command,
> which shell couldn't recognize / locate a matching file for (yet it won't be
> shy of saying so).
> 
> When I start the binary under debugger, I see the following:
> 
> [New Thread 12576.0x4d70]
> [New Thread 12576.0x4860]
> [New Thread 12576.0x44b8]
> [New Thread 12576.0x4528]
> [Thread 12576.0x44b8 exited with code 3221225785]
> [Thread 12576.0x4860 exited with code 3221225785]
> During startup program exited with code 0xc0000139.
> 
> So there's actually an explanation:  0xc0000139(=3221225785) is defined to be
> 
> ntstatus.h:#define STATUS_ENTRYPOINT_NOT_FOUND ((NTSTATUS)0xC0000139)
> 
> Now, the question is:  is it by design that the exit code is posted so 
> timidly?  Shouldn't there be a message that some dynamic linkage wasn't
> satisfied?
> Wouldn't 126 (which is "command cannot execute") be a better choice for the 
> exit code in this case?  Especially if the message cannot be produced at
> all?
> 
> I mean, it is actually a good thing that the executable won't start if it 
> happens to be incompatible with some random Cygwin version;  it just needs to
> be more vocal about the problem and post a more appropriate exit code, too.
> 
> Thanks,
> Anton
> 
> P.S. I didn't get to know what exactly the entry point was missing that the 
> binary required (obviously from the newer Cygwin ABI).
> 
> Reference:
> https://pubs.opengroup.org/onlinepubs/9699919799//utilities/V3_chap02.html#tag_18_08_02
> If a command is not found, the exit status shall be 127. If the command name
> is found, but it is not an executable utility, the exit status shall be
> 126. Applications that invoke utilities without using the shell should use
> these exit status values to report similar errors.

That assumes a POSIX layer is in control: with incompatible Cygwin ABI versions,
calling a missing entrypoint, Cygwin did not get far enough through startup to
detect and report the mismatch; gdb says so here:

> During startup program exited with code 0xc0000139.

so Windows handles termination.

$ fgrep -R 'During startup program exited with code' src/cygwin/newlib-cygwin/
$ fgrep -R 'During startup program exited with code' /bin/
Binary file /bin/gdb.exe matches

-- 
Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada

This email may be disturbing to some readers as it contains
too much technical detail. Reader discretion is advised.

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