[Top][All Lists]

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

[Monotone-devel] Win32 TODO, was Re: Removing things from the database

From: Christof Petig
Subject: [Monotone-devel] Win32 TODO, was Re: Removing things from the database
Date: Fri, 18 Feb 2005 11:41:38 +0100
User-agent: Mozilla/5.0 (X11; U; Linux ppc; en-US; rv:1.7.5) Gecko/20050105 Debian/1.7.5-1

Nathaniel Smith schrieb:
On Thu, Feb 17, 2005 at 11:15:41PM +0100, Christof Petig wrote:

Nathaniel Smith schrieb:

Of course, it would be nicer if we had more convenient
database-to-database sync, which we will sometime but not yet.

It's already in the (not really up to date) ssh branch ;-)
monotone sync/push/pull file:/path/x.db collection opens a pipe to a
different monotone process which serves the other database.

Yep, that's what I was thinking of :-)

This is still blocked on not having a sensible solution for Windows
support, right?

It's blocked on
- me working on the cvssync branch
- me not needing ssh badly (since ssh port redirection and a normal
netsync server work very well and do not have the drawback of allowing
only one connection per database)
- no sensible solution on Win32
   (which is blocked on my unwillingness to compile boost for mingw)

Since I'm in the baroque situation to prefer developing windows programs
on an iBook (under Linux) [sadly I have to gdb on real hardware because
of missing emulation power] it's not all fun. I'll try a verbatim copy
of my linux boost headers into the cross environment combined with the
dll from the windows monotone package (if it's statically linked (or
compiled with a different compiler) I might fall back to concentrate on
compilation (without linking & testing))

Does anybody have boost 1.32.0 dlls compiled with g++ 3.4 for windows?
Or any other boost version with headers included?


PS: The only problematic parts are CreateProcess vs fork and CreatePipe
vs pipe which is needed for both ssh and cvs. [netxx covers sockets
which are needed by cvs pserver access] So I might resort to getting
support for this (process+pipe creation, netxx pipe support) into
monotone first and then use it within the other branches (which will get
_much_ less invasive then).

Anybody interested in volunteering to write and test a pipe+fork+execvp
replacement for windows? All I need is a file descriptor to/from a command.

Attachment: signature.asc
Description: OpenPGP digital signature

reply via email to

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