Re: xterm.c:handle_one_xevent question

[storm@cua.dk (Kim F. Storm)]

Miles Bader <miles@lsi.nec.co.jp> writes:

> I got an emacs abort at the following place in handle_one_xevent
> (xterm.c):
> 

It seems that my recent change to read_avail_input (which
significantly reduced the size of the event buffer) has triggered
some errors in the "read_socket_hook" functions on W32 and X (and
probably MAC as well).

The problem is that the numchars parameter indicating the buffer size
is not checked to see how much can be stuffed into the buffer (and
return when full), but rather [in an inconsistent way] to check that
there is always sufficient space in the buffer or abort otherwise.

I already posted a quick patch for W32 to this list, but haven't
committed it yet, as the rest of the code should be cleaned up.
If someone will commit it, you're welcome.

I will go through the X and MAC code to see what's needed to fix this,
but it will have to wait a few days, so if somebody else want to take
up the challange, feel free to do so.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk