You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Stefan Küng <to...@gmail.com> on 2005/10/18 16:15:00 UTC
Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Hi,
Due to some crashreports with TSVN, I could reproduce a crash also with
the CL client, compiled from the 1.3.x branch (compiled yesterday).
The problem is though that the crash only happens with a specific
repository - I couldn't find a way to get it to crash with my
repositories. But luckily, the user reporting the crash made the repo
available to me.
Here's the recipe:
svn co https://194.208.64.49/svn/DUnit/ crashwc
svn cd crashwc
svn st -u
--> crash
Those of you who are able to fix this or want to reproduce the crash:
username: svn, password: SVN
(please only use that ip for this, I don't have that much
bandwith/datatransfer available).
This doesn't crash with the latest release (1.2.4) but only with the
upcoming 1.3 (1.3.x branch).
It also crashes if you access that repository via file:/// or svn://, so
it's not a problem with the 0.25.3 neon I built svn and TSVN with.
And I don't know if it crashes on Linux too, but it definitely crashes
on Win2k and XP.
Stefan
P.S. Please let me know as soon as you don't need the repository anymore
so I can close down the access.
--
___
oo // \\ "De Chelonian Mobile"
(_,\/ \_/ \ TortoiseSVN
\ \_/_\_/> The coolest Interface to (Sub)Version Control
/_/ \_\ http://tortoisesvn.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by "Peter N. Lundblad" <pe...@famlundblad.se>.
On Tue, 18 Oct 2005, Philip Martin wrote:
> Stefan Küng <to...@gmail.com> writes:
>
> > svn co https://194.208.64.49/svn/DUnit/ crashwc
> > svn cd crashwc
> > svn st -u
> > --> crash
> >
> > Those of you who are able to fix this or want to reproduce the crash:
> > username: svn, password: SVN
> > (please only use that ip for this, I don't have that much
> > bandwith/datatransfer available).
> >
> > This doesn't crash with the latest release (1.2.4) but only with the
> > upcoming 1.3 (1.3.x branch).
> > It also crashes if you access that repository via file:/// or svn://,
> > so it's not a problem with the 0.25.3 neon I built svn and TSVN with.
> > And I don't know if it crashes on Linux too, but it definitely crashes
> > on Win2k and XP.
> >
> > Stefan
> >
> > P.S. Please let me know as soon as you don't need the repository
> > anymore so I can close down the access.
>
> You don't need to keep it, the crash is caused by a HEAD revison
> without an svn:author:
>
It is the ood stuff added to status -u in r16344. It doesn't handle the
case where an entryprop is set with no value (which seems weird in itself,
but anyway...).
Regards,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Daniel Rall <dl...@finemaltcoding.com>.
On Tue, 18 Oct 2005, Paul Burba wrote:
...
> Daniel,
>
> I just confirmed that this problem can also occur in change_file_prop.
> This variation of your previous patch fixes that as well.
Thanks Paul. Committed in r16796, and nominated for backport.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Daniel Rall <dl...@finemaltcoding.com>.
On Tue, 18 Oct 2005, John Peacock wrote:
> Daniel Rall wrote:
> >To fix this problem, checking for a NULL value for "svn:author" is really
> >all that's needed, but perhaps the following patch would be better (given
> >that it's more defensive against other unexpected NULL values)?
>
> Yeah, being proactive is better. +1
r16784, and nominated for backport to 1.3.x.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by John Peacock <jp...@rowman.com>.
Daniel Rall wrote:
> To fix this problem, checking for a NULL value for "svn:author" is really
> all that's needed, but perhaps the following patch would be better (given
> that it's more defensive against other unexpected NULL values)?
Yeah, being proactive is better. +1
John
--
John Peacock
Director of Information Research and Technology
Rowman & Littlefield Publishing Group
4501 Forbes Boulevard
Suite H
Lanham, MD 20706
301-459-3366 x.5010
fax 301-429-5748
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Daniel Rall <dl...@finemaltcoding.com>.
On Tue, 18 Oct 2005, Philip Martin wrote:
> Daniel Rall <dl...@finemaltcoding.com> writes:
>
> >> Does your patch change the behaviour, i.e. it's now NULL for some
> >> other reason?
> >
> > No, the behavior should be identical; the field is initialized to NULL,
> > indicating that it is *not* out of date.
>
> > - /** Set to the user name of the youngest commit, or @c NULL if out
> > - * of date.
> > + /** Set to the user name of the youngest commit, or @c NULL if not
> > + * out of date.
> > * @since New in 1.3
> > */
> > const char *ood_last_cmt_author;
>
> That seems to imply that NULL is a reliable indication of the item not
> being out of date, but it may also happen if svn:author is missing and
> the item is out of date. I think ood_last_cmt_rev won't have this
> problem, callers should be using that to determine whether items are
> out of date.
Philip, you are right again -- r16798. Thanks for the great review.
- Dan
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Philip Martin <ph...@codematters.co.uk>.
Daniel Rall <dl...@finemaltcoding.com> writes:
>> Does your patch change the behaviour, i.e. it's now NULL for some
>> other reason?
>
> No, the behavior should be identical; the field is initialized to NULL,
> indicating that it is *not* out of date.
> - /** Set to the user name of the youngest commit, or @c NULL if out
> - * of date.
> + /** Set to the user name of the youngest commit, or @c NULL if not
> + * out of date.
> * @since New in 1.3
> */
> const char *ood_last_cmt_author;
That seems to imply that NULL is a reliable indication of the item not
being out of date, but it may also happen if svn:author is missing and
the item is out of date. I think ood_last_cmt_rev won't have this
problem, callers should be using that to determine whether items are
out of date.
--
Philip Martin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Daniel Rall <dl...@finemaltcoding.com>.
I committed the svn_wc.h doc string tweaks in r16793, and have nominated
them for backport.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Daniel Rall <dl...@finemaltcoding.com>.
On Tue, 18 Oct 2005, Philip Martin wrote:
> Daniel Rall <dl...@finemaltcoding.com> writes:
...
> No comments on the patch, but I have difficulty understanding the doc
> string for svn_wc_status2_t:
>
> * When the working copy item is out of date compared to the
> * repository, the following fields represent the state of the
> * youngest revision of the item in the repository. If not out of
> * date, the fields are set as described below.
I've adjusted the second sentence in an attempt to clarify its subject
(see patch below). Does that help?
> That appears to state that the fields are valid when the wc is out of
> date, but then this:
>
> /** Set to the user name of the youngest commit, or @c NULL if out
> * of date.
> * @since New in 1.3
> */
> const char *ood_last_cmt_author;
>
> states that the value is NULL when out of date.
>
> Is this field NULL when the wc is out of date, or NULL when the wc is
> *not* out of date?
UGGHH, thanks Philip. You're correct; the doc string is wrong.
> Is the caller expected to check for NULL to determine whether the
> field is out of date?
Or possibly, *not* out of date. ;-P
> Does your patch change the behaviour, i.e. it's now NULL for some
> other reason?
No, the behavior should be identical; the field is initialized to NULL,
indicating that it is *not* out of date.
--- svn_wc.h (revision 16782)
+++ svn_wc.h (working copy)
@@ -1581,29 +1581,30 @@
*
* When the working copy item is out of date compared to the
* repository, the following fields represent the state of the
- * youngest revision of the item in the repository. If not out of
- * date, the fields are set as described below.
+ * youngest revision of the item in the repository. If the working
+ * copy is not out of date, the fields are initialized as described
+ * below.
*/
/** Set to the youngest committed revision, or @c SVN_INVALID_REVNUM
- * if out of date.
+ * if not out of date.
* @since New in 1.3
*/
svn_revnum_t ood_last_cmt_rev;
- /** Set to the most recent commit date, or @c 0 if out of date.
+ /** Set to the most recent commit date, or @c 0 if not out of date.
* @since New in 1.3
*/
apr_time_t ood_last_cmt_date;
/** Set to the node kind of the youngest commit, or @c
- * svn_wc_status_none if out of date.
+ * svn_wc_status_none if not out of date.
* @since New in 1.3
*/
svn_node_kind_t ood_kind;
- /** Set to the user name of the youngest commit, or @c NULL if out
- * of date.
+ /** Set to the user name of the youngest commit, or @c NULL if not
+ * out of date.
* @since New in 1.3
*/
const char *ood_last_cmt_author;
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Philip Martin <ph...@codematters.co.uk>.
Daniel Rall <dl...@finemaltcoding.com> writes:
> To fix this problem, checking for a NULL value for "svn:author" is really
> all that's needed, but perhaps the following patch would be better (given
> that it's more defensive against other unexpected NULL values)?
No comments on the patch, but I have difficulty understanding the doc
string for svn_wc_status2_t:
* When the working copy item is out of date compared to the
* repository, the following fields represent the state of the
* youngest revision of the item in the repository. If not out of
* date, the fields are set as described below.
That appears to state that the fields are valid when the wc is out of
date, but then this:
/** Set to the user name of the youngest commit, or @c NULL if out
* of date.
* @since New in 1.3
*/
const char *ood_last_cmt_author;
states that the value is NULL when out of date.
Is this field NULL when the wc is out of date, or NULL when the wc is
*not* out of date?
Is the caller expected to check for NULL to determine whether the
field is out of date?
Does your patch change the behaviour, i.e. it's now NULL for some
other reason?
--
Philip Martin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Paul Burba <pa...@softlanding.com>.
Daniel Rall <dl...@finemaltcoding.com> wrote on 10/18/2005 01:10:26 PM:
> Peter and I also tracked this to the same spot. I wrote a simple patch
a
> while ago, and am trying to get my httpd WC into a testable state.
>
> To fix this problem, checking for a NULL value for "svn:author" is
really
> all that's needed, but perhaps the following patch would be better
(given
> that it's more defensive against other unexpected NULL values)?
>
> [[[
> * subversion/libsvn_wc/status.c
> (change_dir_prop): Assure that we actually have a value for
> properties sent from the repository before trying to use them.
> ]]]
>
>
> --- subversion/libsvn_wc/status.c (revision 16782)
> +++ subversion/libsvn_wc/status.c (working copy)
> @@ -1504,15 +1504,18 @@
> db->prop_changed = TRUE;
>
> /* Note any changes to the repository. */
> - if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_REV) == 0)
> - db->ood_last_cmt_rev = SVN_STR_TO_REV (value->data);
> - else if (strcmp (name, SVN_PROP_ENTRY_LAST_AUTHOR) == 0)
> - db->ood_last_cmt_author = apr_pstrdup (db->pool, value->data);
> - else if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_DATE) == 0)
> + if (value != NULL)
> {
> - apr_time_t tm;
> - SVN_ERR (svn_time_from_cstring (&tm, value->data, db->pool));
> - db->ood_last_cmt_date = tm;
> + if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_REV) == 0)
> + db->ood_last_cmt_rev = SVN_STR_TO_REV (value->data);
> + else if (strcmp (name, SVN_PROP_ENTRY_LAST_AUTHOR) == 0)
> + db->ood_last_cmt_author = apr_pstrdup (db->pool, value->data);
> + else if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_DATE) == 0)
> + {
> + apr_time_t tm;
> + SVN_ERR (svn_time_from_cstring (&tm, value->data, db->pool));
> + db->ood_last_cmt_date = tm;
> + }
> }
>
> return SVN_NO_ERROR;
Daniel,
I just confirmed that this problem can also occur in change_file_prop.
This variation of your previous patch fixes that as well.
Paul B.
[[[
* subversion/libsvn_wc/status.c
(change_file_prop): Assure that we actually have a value for
properties sent from the repository before trying to use them.
]]]
Index: subversion/libsvn_wc/status.c
===================================================================
--- subversion/libsvn_wc/status.c (revision 16791)
+++ subversion/libsvn_wc/status.c (working copy)
@@ -1697,16 +1697,19 @@
fb->prop_changed = TRUE;
/* Note any changes to the repository. */
- if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_REV) == 0)
- fb->ood_last_cmt_rev = SVN_STR_TO_REV (value->data);
- else if (strcmp (name, SVN_PROP_ENTRY_LAST_AUTHOR) == 0)
- fb->ood_last_cmt_author = apr_pstrdup (fb->dir_baton->pool,
- value->data);
- else if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_DATE) == 0)
+ if (value != NULL)
{
- apr_time_t tm;
- SVN_ERR (svn_time_from_cstring (&tm, value->data,
fb->dir_baton->pool));
- fb->ood_last_cmt_date = tm;
+ if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_REV) == 0)
+ fb->ood_last_cmt_rev = SVN_STR_TO_REV (value->data);
+ else if (strcmp (name, SVN_PROP_ENTRY_LAST_AUTHOR) == 0)
+ fb->ood_last_cmt_author = apr_pstrdup (fb->dir_baton->pool,
+ value->data);
+ else if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_DATE) == 0)
+ {
+ apr_time_t tm;
+ SVN_ERR (svn_time_from_cstring (&tm, value->data,
fb->dir_baton->pool));
+ fb->ood_last_cmt_date = tm;
+ }
}
return SVN_NO_ERROR;
_____________________________________________________________________________
Scanned for SoftLanding Systems, Inc. and SoftLanding Europe Plc by IBM Email Security Management Services powered by MessageLabs.
_____________________________________________________________________________
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Daniel Rall <dl...@finemaltcoding.com>.
On Tue, 18 Oct 2005, Philip Martin wrote:
> John Peacock <jp...@rowman.com> writes:
>
> > Philip Martin wrote:
> >> You don't need to keep it, the crash is caused by a HEAD revison
> >> without an svn:author:
> >> $ svnadmin create repo
> >> $ echo '#!/bin/sh' > repo/hooks/pre-revprop-change
> >> $ chmod +x repo/hooks/pre-revprop-change
> >> $ svn pd --revprop -rHEAD svn:author file://`pwd`/repo
> >> $ svn co file://`pwd`/repo wc
> >> $ svn st -u wc
> >> Segmentation fault
> >
> > Hrm... I wonder if that is in the keywords-as-hash code, then (since
> > you didn't provide a backtrace). I'll see if I can quickly track it
> > down...
>
> It's the new ood stuff:
>
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 16384 (LWP 1847)]
> 0x4006f34b in change_dir_prop (dir_baton=0x80a4cd0,
> name=0x40456357 "svn:entry:last-author", value=0x0, pool=0x8088408)
> at ../svn/subversion/libsvn_wc/status.c:1510
> 1510 db->ood_last_cmt_author = apr_pstrdup (db->pool, value->data);
> (gdb) p value
> $1 = (const svn_string_t *) 0x0
> (gdb)
Peter and I also tracked this to the same spot. I wrote a simple patch a
while ago, and am trying to get my httpd WC into a testable state.
To fix this problem, checking for a NULL value for "svn:author" is really
all that's needed, but perhaps the following patch would be better (given
that it's more defensive against other unexpected NULL values)?
[[[
* subversion/libsvn_wc/status.c
(change_dir_prop): Assure that we actually have a value for
properties sent from the repository before trying to use them.
]]]
--- subversion/libsvn_wc/status.c (revision 16782)
+++ subversion/libsvn_wc/status.c (working copy)
@@ -1504,15 +1504,18 @@
db->prop_changed = TRUE;
/* Note any changes to the repository. */
- if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_REV) == 0)
- db->ood_last_cmt_rev = SVN_STR_TO_REV (value->data);
- else if (strcmp (name, SVN_PROP_ENTRY_LAST_AUTHOR) == 0)
- db->ood_last_cmt_author = apr_pstrdup (db->pool, value->data);
- else if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_DATE) == 0)
+ if (value != NULL)
{
- apr_time_t tm;
- SVN_ERR (svn_time_from_cstring (&tm, value->data, db->pool));
- db->ood_last_cmt_date = tm;
+ if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_REV) == 0)
+ db->ood_last_cmt_rev = SVN_STR_TO_REV (value->data);
+ else if (strcmp (name, SVN_PROP_ENTRY_LAST_AUTHOR) == 0)
+ db->ood_last_cmt_author = apr_pstrdup (db->pool, value->data);
+ else if (strcmp (name, SVN_PROP_ENTRY_COMMITTED_DATE) == 0)
+ {
+ apr_time_t tm;
+ SVN_ERR (svn_time_from_cstring (&tm, value->data, db->pool));
+ db->ood_last_cmt_date = tm;
+ }
}
return SVN_NO_ERROR;
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Philip Martin <ph...@codematters.co.uk>.
John Peacock <jp...@rowman.com> writes:
> Philip Martin wrote:
>> You don't need to keep it, the crash is caused by a HEAD revison
>> without an svn:author:
>> $ svnadmin create repo
>> $ echo '#!/bin/sh' > repo/hooks/pre-revprop-change
>> $ chmod +x repo/hooks/pre-revprop-change
>> $ svn pd --revprop -rHEAD svn:author file://`pwd`/repo
>> $ svn co file://`pwd`/repo wc
>> $ svn st -u wc
>> Segmentation fault
>
> Hrm... I wonder if that is in the keywords-as-hash code, then (since
> you didn't provide a backtrace). I'll see if I can quickly track it
> down...
It's the new ood stuff:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 16384 (LWP 1847)]
0x4006f34b in change_dir_prop (dir_baton=0x80a4cd0,
name=0x40456357 "svn:entry:last-author", value=0x0, pool=0x8088408)
at ../svn/subversion/libsvn_wc/status.c:1510
1510 db->ood_last_cmt_author = apr_pstrdup (db->pool, value->data);
(gdb) p value
$1 = (const svn_string_t *) 0x0
(gdb)
--
Philip Martin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by John Peacock <jp...@rowman.com>.
Philip Martin wrote:
> You don't need to keep it, the crash is caused by a HEAD revison
> without an svn:author:
>
> $ svnadmin create repo
> $ echo '#!/bin/sh' > repo/hooks/pre-revprop-change
> $ chmod +x repo/hooks/pre-revprop-change
> $ svn pd --revprop -rHEAD svn:author file://`pwd`/repo
> $ svn co file://`pwd`/repo wc
> $ svn st -u wc
> Segmentation fault
Hrm... I wonder if that is in the keywords-as-hash code, then (since
you didn't provide a backtrace). I'll see if I can quickly track it down...
John
--
John Peacock
Director of Information Research and Technology
Rowman & Littlefield Publishing Group
4501 Forbes Boulevard
Suite H
Lanham, MD 20706
301-459-3366 x.5010
fax 301-429-5748
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Stefan Küng <to...@gmail.com>.
Philip Martin wrote:
> You don't need to keep it, the crash is caused by a HEAD revison
> without an svn:author:
>
> $ svnadmin create repo
> $ echo '#!/bin/sh' > repo/hooks/pre-revprop-change
> $ chmod +x repo/hooks/pre-revprop-change
> $ svn pd --revprop -rHEAD svn:author file://`pwd`/repo
> $ svn co file://`pwd`/repo wc
> $ svn st -u wc
> Segmentation fault
Thanks a lot. I closed access to the server now.
Stefan
--
___
oo // \\ "De Chelonian Mobile"
(_,\/ \_/ \ TortoiseSVN
\ \_/_\_/> The coolest Interface to (Sub)Version Control
/_/ \_\ http://tortoisesvn.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Crash with 'svn st -u' and Version 1.3 (1.3.x branch)
Posted by Philip Martin <ph...@codematters.co.uk>.
Stefan Küng <to...@gmail.com> writes:
> svn co https://194.208.64.49/svn/DUnit/ crashwc
> svn cd crashwc
> svn st -u
> --> crash
>
> Those of you who are able to fix this or want to reproduce the crash:
> username: svn, password: SVN
> (please only use that ip for this, I don't have that much
> bandwith/datatransfer available).
>
> This doesn't crash with the latest release (1.2.4) but only with the
> upcoming 1.3 (1.3.x branch).
> It also crashes if you access that repository via file:/// or svn://,
> so it's not a problem with the 0.25.3 neon I built svn and TSVN with.
> And I don't know if it crashes on Linux too, but it definitely crashes
> on Win2k and XP.
>
> Stefan
>
> P.S. Please let me know as soon as you don't need the repository
> anymore so I can close down the access.
You don't need to keep it, the crash is caused by a HEAD revison
without an svn:author:
$ svnadmin create repo
$ echo '#!/bin/sh' > repo/hooks/pre-revprop-change
$ chmod +x repo/hooks/pre-revprop-change
$ svn pd --revprop -rHEAD svn:author file://`pwd`/repo
$ svn co file://`pwd`/repo wc
$ svn st -u wc
Segmentation fault
--
Philip Martin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org