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

RE: [PATCH] WC DB verification 1

From: Bert Huijben <bert_at_qqmail.nl>
Date: Tue, 10 May 2011 13:28:08 +0200

> -----Original Message-----
> From: Greg Stein [mailto:gstein_at_gmail.com]
> Sent: dinsdag 10 mei 2011 11:18
> To: dev_at_subversion.apache.org
> Subject: Re: [PATCH] WC DB verification 1
>
> >...
> > I would suggest adding a new function for this purpose; and maybe a
> separate
> > program to call it for our beta cycle.
>
> Please... not another program. There was also a recent call for 'svn
> upgrade' to be separate. It really sucks to have an installation smear
> a dozen programs into your system ("which one do I use? for what? hmm.
> for this? ... um..."). We already have a notion of subcommand. That is
> normal, expected, and understood. We can continue to use that without
> fear of user misunderstanding. In fact, it is probably best to tell
> users "just use 'svn'. you don't need anything else". (administrators
> are another thing, but we don't have to complicate their lives, too!
> ... we have too many executables)

I'm still not sure if this tool should survive outside the subversion
developer world.

And I didn't hear you suggest that we should include all the
svnraisetreeconflict and entries-dump and like programs in 'svn'.

I would suggest creating the API inside a library so other tools may use it,
and *for now* enable its use via a separate program.

If it becomes useful in detecting problems (and maybe even fixing) we might
make it better accessible.

(But until the primary check is more than do we set parent_relpath correctly
I don't see its use. But that might change in the next commit)

> > This code should never be necessary once we call the db stable. And
> making
> > the normal cleanup very slow makes it less useful for its normal task.
>
> Are you kidding? There is no such thing as "normal" when you're
> talking about 'svn cleanup'. That command should take as long as it
> needs to ensure that everything is good. Let me repeat: AS LONG AS IT
> NEEDS. If people need to run 'svn cleanup' more than once a year, then
> we've failed. But if they only need to run it once a year, then it can
> goddamned well take as much as it needs to ensure that the user has
> not lost any of their changes.
>
> I find it absurd to place time constraints on 'svn cleanup'. The
> better approach is to ensure it is never needed.

I agree that there should be no 'normal use' for svn cleanup. (And
personally I only run svn cleanup in testing cases)

But there are *a lot* of users who think of svn cleanup as a tool to update
their recorded_size and recorded_mod time after they unleashed some tool on
their working copy.
I even heard users suggesting to run svn cleanup before *every* command
invocation.

Yes, I tell all these users not to do this...
But there are more than a few users who currently use svn cleanup this way.

        Bert
Received on 2011-05-10 13:28:46 CEST

This is an archived mail posted to the Subversion Dev mailing list.