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

Re: Questions about migrating svn_wc_schedule_t usage to wc-ng

From: Dave Brown <dave.brown_at_wandisco.com>
Date: Tue, 22 Sep 2009 21:21:46 -0700

Greg Stein wrote:
> Hey Dave,
>
> We'd like to move away from the old "schedule" concept because it
> doesn't do a very good job of detailing what is really going on for a
> given path.
>
> That said, the "node" functions are present to help with the
> transition away from entry_t (and access batons and whatever). So if
> you want to make a node function to return the old-school schedule
> value, for expediency, then go ahead. Eventually, we'll track down all
> callers and figure out the Right Change.
>
> Note that questions.c has a function called
> svn_wc__internal_is_replaced(). That is a Correct implementation for
> determining the old svn_wc_schedule_replace state. You could use that,
> expand on it, or wrap it and expose it to libsvn_client, if any of
> that is helpful.
>

Ah, svn_wc__internal_is_replaced() is helpful, as is the mini roadmap.
I'll try replacing uses of entry->schedule in libsvn_client/ along the
same lines. Or if that's too daunting, fall back to using the
old-school values (as I had started), since this seems transitional anyway.

-dave

> Cheers,
> -g
>
> On Tue, Sep 22, 2009 at 20:03, Dave Brown <dave.brown_at_wandisco.com> wrote:
>> Hi all,
>>
>> I've been working on removing svn_wc_entry_t usage in
>> libsvn_client/export.c. Most of the fields from wc_entry_t needed by
>> export.c have fetch api's already (svn_wc_private.h / node.c), except
>> wc_entry_t->schedule. I see that it =~ wc-ng's svn_wc__db_status_t, so
>> I've added a fetch function for a node's svn_wc_schedule_t, akin to
>> svn_wc__node_get_kind(), to translate wc-ng values -> wc_schedule_t. So
>> I'd like to check - does this sound like the right track? If so, I have
>> follow up questions on translating some of the svn_wc__db_status_t
>> values (entries.c is doing this now).
>>
>> It also occurred to me that perhaps the plan is to promote the wc_db.h
>> api's to be more public at some point, which might change how this sort
>> of thing would be handled.
>>
>> If this all sounds close to "right" I'll send along a patch.
>>
>> -dave
>>
>> ------------------------------------------------------
>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2398693
>>

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2398725
Received on 2009-09-23 06:23:20 CEST

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