Mailing-List: contact dev-help@subversion.apache.org; run by ezmlm
Precedence: bulk
Received-SPF: neutral (nike.apache.org: local policy)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=btopenworld.com;
  h=X-YMail-OSG:Received:X-Mailer:References:Message-ID:Date:From:Reply-To:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding;
  b=bE0NMJdEa9Nr+CZFwbhXfElBmrP3nqBmpCKfve51lTg8ngraDNYlW3ecfYse8XO9t0MGJOjehzZ6iF8LJszQS/P8Hexvw6vN0OdSgrUlazjmRWTazsQ8iHzIXS6i9Mt6nuimd/qgirO+GQWeBPMQBBhlcHloG2MzM5NCoz/6rsk=;
References: <4EFABEBF.2070408@gmail.com>
 <CAJjMeYOU==go58QLvRe30tk182m+KY1Ffd8DU4Baas6VaEQzWg@mail.gmail.com>
 <4EFC8DFA.6060008@gmail.com>
 <CAH=6jSZFntL-OmSv0AuR65UQ--KFptWSv5y9NcBj82d2mkPn2A@mail.gmail.com>
Message-ID: <1325586209.19073.YahooMailNeo@web86702.mail.ird.yahoo.com>
Date: Tue, 3 Jan 2012 10:23:29 +0000 (GMT)
From: Julian Foad <julianfoad@btopenworld.com>
Reply-To: Julian Foad <julianfoad@btopenworld.com>
Subject: Re: crash fetching status
To: =?iso-8859-1?Q?Stefan_K=FCng?= <tortoisesvn@gmail.com>,
  Hyrum K Wright <hyrum.wright@wandisco.com>
Cc: Subversion Development <dev@subversion.apache.org>
In-Reply-To: 
 <CAH=6jSZFntL-OmSv0AuR65UQ--KFptWSv5y9NcBj82d2mkPn2A@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

Stefan K=FCng wrote:=0A=0A> On Thu, Dec 29, 2011 at 16:57, Stefan K=FCng wr=
ote:=0A>> On 29.12.2011 16:43, Hyrum K Wright wrote:=0A>>> On Wed, Dec 28, =
2011 at 1:01 AM, Stefan K=FCng wrote:=0A=0A[...]=0A=0A>>> I looked at simil=
ar places elsewhere, and they seemed to follow a=0A>>> pattern. =A0The foll=
owing patch is in a similar vein:=0A>>>=0A>>> [[[=0A>>> Index: subversion/l=
ibsvn_wc/status.c=0A>>> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=0A>>> - =A0f->repos_relpath =3D svn_relpath_join(find_dir_repos_relpath(pb=
, pool),=0A>>> - =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =
=A0 =A0 =A0 =A0 =A0f->name, pool);=0A>>> +=0A>>> + =A0dir_repos_relpath =3D=
 find_dir_repos_relpath(pb, pool);=0A>>> + =A0if (dir_repos_relpath)=0A>>> =
+ =A0 =A0f->repos_relpath =3D svn_relpath_join(dir_repos_relpath, f->name,=
=0A>>> pool);=0A>>> + =A0else=0A>>> + =A0 =A0f->repos_relpath =3D apr_pstrd=
up(pool, f->name);=0A>>> +=0A>>>=0A>>> All the tests pass, though I'm not s=
ure if there are other correctness=0A>>> issues here or not, so I haven't c=
ommitted this.=0A>>=0A>> Just a thought:=0A>> svn_relpath_join() joins two =
path/url parts. If one is an empty string, it=0A>> still works. But it cras=
hes when one of the parts is NULL instead of an=0A>> empty string.=0A=0AYes=
, because NULL is not a valid relpath.=A0 The 'relpath' concept is well def=
ined (in svn_dirent_uri.h).=A0 The representation of an empty relpath is th=
e empty string.=0A=0AThe real problem here is that find_dir_repos_relpath()=
 can return NULL whereas its callers (including but not limited to this one=
) expect a non-null value.=0A=0ASo the real solution is either to change fi=
nd_dir_repos_relpath() so it always returns a valid 'relpath', or to change=
 its callers to code around the fact that it can return NULL.=0A=0ADoes it =
return NULL to mean the empty relpath or for some other reason?=0A=0A=0A>> =
Maybe just rewrite svn_relpath_join() so that NULL is treated as an empty=
=0A>> string?=0A=0ANo.=A0 Don't make an exception.=0A=0A- Julian=0A