Re: "Forked child base mismatch" using MKS Toolkit shell

Mike> --On Monday, June 17, 2002 8:03 PM -0700 Andy Piper <andy(a)xemacs.org&gt; Mike> wrote: Mike> When the native Windows version of XEmacs creates a process Mike> it injects some code into a thread in the new process to Mike> call SetConsoleCtrlHandler in that process. As it currently Mike> stands it does this before starting the main thread in the Mike> new process. This causes some minor change in the memory Mike> layout of the process which apparently confuses the MKS Mike> shell. My change was to defer this forced call to Mike> SetConsoleCtrlHandler until the main thread had a chance to Mike> get started. This solves the problem with the MKS shell but Mike> made it impossible to interrupt some programs. Mike> In my workspace I have a version which makes the behavior Mike> conditional on the setting of a Lisp variable Mike> mswindows-mks-compatible-start-process, but I've never Mike> proposed checking that in since my general feeling is that Mike> there must be a better way to handle this whole problem. Mike> I've discovered that, even with the original code, it's not Mike> possible to reliably interrupt some subprocesses. Other Hi Mike, I have put MKS on my personal list of offenders when I found they had broken RCS compatibility by changing the RCS file format in their product called "Source Integrity". Therefor I am not in favor of making any compromise to support MKS. Adrian Mike> programs that start subprocesses are able to interrupt them Mike> so there must be a way to do this properly. I just haven't Mike> had time to look into it in detail. In the maintime a note Mike> in PROBLEMS would be a good idea. Mike> Mike -- Adrian Aichner mailto:adrian＠xemacs.org http://www.xemacs.org/