Greg Hudson wrote:
>On Fri, 2005-05-06 at 03:29 +0200, Branko Čibej wrote:
>
>
>>Alternatively, the solution might be to allow "permanently missing"
>>files in the working copy (that is, files that appear in .svn/entries,
>>but are marked as "don't create"). Their status would have to be checked
>>at every update to see if the name had changed, and would add even more
>>spaghetti to our WC code.
>>
>>
>
>Isn't this pretty much exactly what we need for non-recursive checkouts?
>
>
Not exactly. In a non-recursive checkout, you'd only change the "don't
create" flag if explicitly requested to do so. In this case, it'd be
based on whether the filename is compatible with the underlying
filesystem, That's what I don't want to do.
>>http://subversion.tigris.org/issues/show_bug.cgi?id=1559
>>
>>Should close as WONTFIX. There's srvany from Microsoft, there's
>>FireDaemon, and last but not least, there's svnservice
>>(http://dark.clansoft.dk/~mbn/svnservice/).
>>
>>
>
>I think svnserve should be able to act as a Windows service; I just want
>apr to do all the heavy lifting.
>
>
I don't think that will ever happen. It would be equivalent to APR
generating /etc/init.d scripts for Unix. Windows service startup is a
lot more complex than daemonizing a Unix process.
-- Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri May 6 17:48:01 2005