[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: [Subclipse-users] Recursive status problems

From: Mark Phippard <markphip_at_gmail.com>
Date: Wed, 1 Jul 2009 08:45:24 -0400

I fear that anything you try it still going to have some problems just
because there are still 22,000 new files that some degree of
processing has to be done on to decide that they should be ignored.
But here are some ideas:

1) Use the Eclipse Linked Resources feature to link a folder in your
project to the tree. Subclipse would probably be able to ignore this
the quickest. It likely does not work with your makefiles though.

2) Root the Linux kernel in a folder name that you can add to the
global ignored resources list. In preferences go to Team > Ignored
Resources.

3) Similar to #2 use the svn:ignore property to ignore the tree. You
might already be doing this.

#3 is unlikely to offer much of a speedup, because svn status still
has to be called. It just might make svn status a little quicker. So
#2 would be the first thing to try. Even if you do #2, however, you
should still absolutely do option #3 too. There will be times where
we will run svn status recursively on your project. This is generally
pretty fast, even on a tree this big. However, if you use svn:ignore
it will be even faster as the SVN code will not descend into the tree.

Mark

On Wed, Jul 1, 2009 at 7:17 AM, Chris Smith<chris.smith_at_st.com> wrote:
> Hi everyone,
>
> I think I'm seeing the issue reported by Ben Hutchinson in the thread "Extremely slow performance in Subclipse method StatusCacheManager.basicGetStatus()" back in March and I wondered if anyone could offer any advice, or better still a workaround.
>
> The situation is this. I and users of the embedded Linux plugins I work on will often create a new Makefile Project, not in the default, empty directory in the workspace but in a directory that already contains a copy of the Linux kernel - all 22,000 files of it. When Subclipse isn't present, this is quite a neat way to quickly wrap a project around the kernel tree I'm working on, but when I install Subclipse I find that Eclipse often hangs after the project is created, or that most operations refuse to proceed because some other non-visible operation is still in progress. Looking at the backtrace it's busily trying to "svn status" files in my new project. The kernel trees I'm talking about are not in Subversion, or indeed any other version control at this time.
>
> I've read in other threads that there might be a problem with a flood of "resource changed" events or something like that? Would I be provoking that in this case? Can I prevent this? Note that I package and supply Eclipse and some of our own plug-ins to our customers, so patching Eclipse or CDT is an option. Can I programatically mark the project as "definitely not in Subversion - leave me alone" in some way when it's created?
>
> Context: I am using Eclipse Ganymede and CDT5 on a Fedora 10 host with Subclipse 1.4.8 and Subversion 1.5.4. Both JavaHL and SVNKit interfaces seem to be affected.
>
> I'd be very grateful for any light anyone could shed on this issue, as it would be great for to be able to continue creating projects in this way and continue use Subclipse.
>
> Many thanks,
>
>
> Chris
>
> ------------------------------------------------------
> http://subclipse.tigris.org/ds/viewMessage.do?dsForumId=1047&dsMessageId=2367007
>
> To unsubscribe from this discussion, e-mail: [users-unsubscribe_at_subclipse.tigris.org].
>

-- 
Thanks
Mark Phippard
http://markphip.blogspot.com/
------------------------------------------------------
http://subclipse.tigris.org/ds/viewMessage.do?dsForumId=1047&dsMessageId=2367027
To unsubscribe from this discussion, e-mail: [users-unsubscribe_at_subclipse.tigris.org].
Received on 2009-07-01 14:45:37 CEST

This is an archived mail posted to the Subclipse Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.