emacs-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] Override Windows default Win-* key combinations when using E


From: Eli Zaretskii
Subject: Re: [PATCH] Override Windows default Win-* key combinations when using Emacs
Date: Thu, 14 Jan 2016 20:15:56 +0200

> Cc: address@hidden
> From: Jussi Lahdenniemi <address@hidden>
> Date: Thu, 14 Jan 2016 14:49:31 +0200
> 
> On 13.1.2016 17.57, Eli Zaretskii wrote:
> > If you can see in GDB where does this come from and show the
> > backtrace, that would be great.  Failing that, let's leave Windows 9X
> > testing alone for a while (you can always test the fallback code by
> > forcefully failing one of the run-time tests for the newer versions)
> > and test only on W2K.  We can return to the 9X problem later.
> 
> I was able to get a GDB dump of this, after installing MinGW from 2009. 
> The backtrace is attached.

Looks like it's unable to create a buffer.  Hmm...

If you can spare some time to try debugging this with me, please file
a bug report about this problem with "M-x report-emacs-bug", and let's
take it up there.  The first thing I'd like to know is what buffer is
that (I'm guessing *scratch* or *Messages*), and what is the value of
'a' in this call frame:

  #4  0x01103082 in make_lisp_ptr (ptr=0x2a3da84, type=Lisp_Vectorlike)
      at ../../emacs/src/lisp.h:1110

One of the major changes since Emacs 24.5 is that we allocate memory
in a completely different manner, using VirtualAlloc directly for
buffer text and HeapAlloc for everything else (see w32heap.c).
Perhaps some of that doesn't work on Windows 9X?

> I investigated the problem and found out that it resulted from NT4's 
> MultiByteToWideChar function not supporting the MB_ERR_INVALID_CHARS 
> flags.  I made the flag conditional on the OS version, after which Emacs 
> runs just fine on NT4.  I'm attaching a patch that contains this fix, 
> and the conditional use of _resetstkoflw (which is also not supported on 
> NT4) as well.

Thanks, I will push these to the emacs-25 branch soon.

> Also, I noticed that addpm.exe fails to load on a out-of-the-box NT4SP6, 
> due to its dependency to a Shell32.dll version that ships with IE4. 
> After installing IE4, addpm works.  It's worth noting that directly 
> installing IE6 on NT4 does NOT make addpm functional, as IE6 does not 
> contain the shell update.

Thanks, I will mention this in some place, although addpm is
semi-obsolescent nowadays.



reply via email to

[Prev in Thread] Current Thread [Next in Thread]