Re: [h-e-w] gnuserv maintenance

[Jason Rumney]

Matthew X. Economou wrote:

> The sockets version of the gnuserv port to NT currently has no maintainer.
I don't think that is true. There is just no maintenance to do as far as 
the author is aware.

Guy Pascoigne-Piggford is the author of the version on wyrdrune.com. He 
has recently posted to emacs-devel with respect to porting emacsclient 
to Windows, so that the functionality of gnuserv/gnuclient is available 
by default.

Your ideas might be useful for him to consider for that port.

> (1) gnuclientw exits immediately after sending the file to Emacs, i.e. "-q" is 
> always set.
>  
Many people consider this a feature, but I agree with you. Since there 
is no Window associated with gnuclientw, users will not notice that it 
hangs around until the buffer is killed, so I don't know what the 
rationale behind the current behaviour is

> (2) The connection between gnuclient/gnuclientw and gnuserv closes after 300 
> seconds.
>  
This sounds like a standard TCP timeout. This is a disadvantage of using 
TCP sockets, and will have to be considered for the emacsclient por

> I plan on making the following changes:
>
> (1) Update the sources to build on Visual Studio .NET.
>  
I think this is counter-productive. Microsoft's compilers seem to be 
getting more and more proprietary, pushing you towards .NET rather than 
Posix C.  A better idea would be to update it to build on the GNU 
compilers from MingW32 and Cygwin.

> (2) Merge the mailslot version.
>  
I don't think mailslots offer any more security than TCP sockets. They 
can be used remotely and reportedly are not blocked by some personal 
firewalls. Stefan Monnier has a simple idea for authenticating TCP 
sockets (similar to how many versions of X work), but for 21.4, we will 
probably just restrict them to localhost.