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

Re: [Issue 3596] 'hotcopy' of packed fsfs repos may corrupt target revprops.db

From: Daniel Shahaf <d.s_at_daniel.shahaf.name>
Date: Fri, 16 Apr 2010 19:41:39 +0300 (Jerusalem Daylight Time)

Daniel Shahaf wrote on Fri, 16 Apr 2010 at 02:01 +0300:
> Daniel Shahaf wrote on Fri, 16 Apr 2010 at 00:52 +0300:
> > (probably need some coffee)
>
> Indeed. The other two svn_fs_fs__* functions in rep-cache.c still check
> rep_cache_db (not rep_cache_db_opened). Which makes me wonder if there's
> still a race condition there or not.

Apparently, svn_atomic__init_once()'s postcondition is that the
initialization function has been called *and the call has finished*. So
rep_cache_db is guaranteed to be non-NULL (in all threads) after
svn_fs_fs__open_rep_cache() is called, and the problem I feared last
night is non-existent.

/me -> edit STATUS

Daniel
Received on 2010-04-16 18:41:26 CEST

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