You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Greg Hudson <gh...@MIT.EDU> on 2002/12/02 03:26:35 UTC
ra_svn and Windows
I think I'm ready to check in ra_svn. It's only minimally useful at
the moment (because it only supports anonymous authentication), but
it's a good starting point.
I'd prefer not to break the windows build. The client code might
conceivably build and work, but I'm clueless on how to write the
equivalent of a daemon process on Windows. (The socket stuff seems
pretty simple, but the other trappings of a daemon process seem
difficult. on Unix, at startup you background, dissociate from the
controlling tty, change dirs to "/", and ditch stdin/stdout/stderr; I
also need to fork a process to handle each connection. Alternatively,
I could run out of inetd, but that's equally Windows-unfriendly.)
I'm guessing that without the dsp generator in place, checking in new
code won't hurt the Windows build because the .dsp files won't have
changed. But it looks like we're pretty close to have a dsp
generator, at which point the new directives in build.conf are likely
to get into people's way.
What should I do here? In the long term, of course, both the client
and server should operate on Windows, but I'll need some help on that
since I lack a Windows development environment.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra_svn and Windows
Posted by Branko Čibej <br...@xbc.nu>.
Greg Hudson wrote:
>I'm guessing that without the dsp generator in place, checking in new
>code won't hurt the Windows build because the .dsp files won't have
>changed. But it looks like we're pretty close to have a dsp
>generator, at which point the new directives in build.conf are likely
>to get into people's way.
>
>What should I do here? In the long term, of course, both the client
>and server should operate on Windows, but I'll need some help on that
>since I lack a Windows development environment.
>
>
I'd say just go ahead and commit, if necessary we can always
special-case the build.conf directives in the .dsp generator. I haven't
had fun writing Windows services in ages, but I do have some skeleton
code lying around, and then there'a Apache I can plagiarize from. I'll
get it working once the Windows build breaks, if not before. :-)
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra_svn and Windows
Posted by Greg Hudson <gh...@MIT.EDU>.
On Sun, 2002-12-01 at 22:28, Garrett Rooney wrote:
> can't APR handle this kind of stuff portably? what does apache do?
Hm, there's apr_proc_detach(), which does all the initial daemonization
stuff. I'm not sure what it does on Windows. And of course there are
functions to create processes, although the only portable ones execute
another program.
Apache does a lot of stuff we don't need to (pre-forking a process pool,
I think), so it doesn't seem like a great example of an apr-using
forking daemon.
At any rate, no matter how apr-PC I make my code, I can't test it on
Windows. My question is what I can do to avoid burdening the
Windows-enabled developers with having to fix my code on a short
schedule.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra_svn and Windows
Posted by Garrett Rooney <ro...@electricjellyfish.net>.
On Sunday, December 1, 2002, at 10:26 PM, Greg Hudson wrote:
> I think I'm ready to check in ra_svn. It's only minimally useful at
> the moment (because it only supports anonymous authentication), but
> it's a good starting point.
>
> I'd prefer not to break the windows build. The client code might
> conceivably build and work, but I'm clueless on how to write the
> equivalent of a daemon process on Windows. (The socket stuff seems
> pretty simple, but the other trappings of a daemon process seem
> difficult. on Unix, at startup you background, dissociate from the
> controlling tty, change dirs to "/", and ditch stdin/stdout/stderr; I
> also need to fork a process to handle each connection. Alternatively,
> I could run out of inetd, but that's equally Windows-unfriendly.)
can't APR handle this kind of stuff portably? what does apache do?
-garrett
--
garrett rooney Remember, any design flaw you're
rooneg@electricjellyfish.net sufficiently snide about becomes
http://electricjellyfish.net/ a feature. -- Dan Sugalski
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org