delorie.com/archives/browse.cgi   search  
Mail Archives: cygwin-developers/1998/11/26/17:19:46

From: cgf AT cygnus DOT com (Christopher Faylor)
Subject: Re: New snapshot
26 Nov 1998 17:19:46 -0800 :
Message-ID: <19981126195227.A18444.cygnus.cygwin32.developers@cygnus.com>
References: <01BE1930 DOT 774B49E0 AT sos>
Mime-Version: 1.0
To: Corinna Vinschen <corinna DOT vinschen AT cityweb DOT de>,
cygwin32-developers AT cygnus DOT com

On Fri, Nov 27, 1998 at 01:35:12AM +0100, Corinna Vinschen wrote:
>Hi again Chris,
>
>Before a few minutes, I started the pdksh Shell the first time with
>the new snapshot. This shows another interesting behaviour. _Every_
>time, I start a process from command line, it does it work, then there is
>a long pause as in tcsh, and then the following message appears, before
>the shell prompt reappears:
>
>****** exceptions.cc:401 __get_signal_mutex failed, res 258, error 2
>  0) function exceptions.cc, line 364
>
>This is the message from sigproc.cc(__get_signal_mutex),
>
>	...
>	else if (howlong >= 1000)
>		small_printf(...)
>
>which didn't exist in the last snapshot. I'm sure, the previously
>described problem in tcsh has the same reason.
>
>I hoped to find the solution myself, but for today, I didn't understand
>enough of the underlying code. I'm too tired.

That error is diagnostic output that I added when debugging the slow
configure problem.  It's indicating a race in the sigsuspend code.

I'll look into it.

-chris

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019