You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Clark Rawlins <cl...@bit63.org> on 2002/10/07 11:41:50 UTC
[Issue 907] libxml broken?
I haven't had a chance to look into this for a few days but last night I
started poking arround the libsvn_ra_dav code to figure out where the
escaped & were comming from.
It looks like they are comming from the attribute list provided at the
begining of element parsing through neon in the xml parsing
layer. But, it also looks like neon is just passing the attribute lists
straight from the xml parsing library (in my case libxml).
Reconfiguring/recompiling with configure ... --with-expat solved the
problem with the svn update problem.
Can someone confirm the workarround?
Does anyone disagree that general (i.e. &, ", etc) entities
found in attributes should be dereferenced before passing them to the
application layer?
Clark
--
"They that can give up essential liberty to obtain a little temporary safety
deserve neither liberty nor safety."
- Benjamin Franklin, Historical Review of Pennsylvania, 1759.
Re: [Issue 907] libxml broken?
Posted by Clark Rawlins <cl...@bit63.org>.
When I get a few minuites I will attempt to work up a test case
and send it to the maintainer.
Clark
On Mon, Oct 07, 2002 at 01:38:36PM +0100, Joe Orton wrote:
> On Mon, Oct 07, 2002 at 06:41:50AM -0500, Clark Rawlins wrote:
> > I haven't had a chance to look into this for a few days but last night I
> > started poking arround the libsvn_ra_dav code to figure out where the
> > escaped & were comming from.
> ...
> > Can someone confirm the workarround?
>
> Yes, confirmed that libxml 2.4.23 isn't expanding "&" in attribute
> values; unless there is something extra that neon should be doing with
> the libxml SAX interface I'd presume this is a libxml2 bug.
>
> > Does anyone disagree that general (i.e. &, ", etc) entities
> > found in attributes should be dereferenced before passing them to the
> > application layer?
>
> No disagreement here - can you report this bug to the maintainer?
>
> Regards,
>
> joe
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: dev-help@subversion.tigris.org
>
--
"They that can give up essential liberty to obtain a little temporary safety
deserve neither liberty nor safety."
- Benjamin Franklin, Historical Review of Pennsylvania, 1759.
Re: [Issue 907] libxml broken?
Posted by Joe Orton <jo...@manyfish.co.uk>.
On Mon, Oct 07, 2002 at 01:38:36PM +0100, Joe Orton wrote:
> On Mon, Oct 07, 2002 at 06:41:50AM -0500, Clark Rawlins wrote:
> > I haven't had a chance to look into this for a few days but last night I
> > started poking arround the libsvn_ra_dav code to figure out where the
> > escaped & were comming from.
> ...
> > Can someone confirm the workarround?
>
> Yes, confirmed that libxml 2.4.23 isn't expanding "&" in attribute
> values; unless there is something extra that neon should be doing with
> the libxml SAX interface I'd presume this is a libxml2 bug.
I asked Daniel Veillard about this, and you do have to turn this on
explicitly with libxml, here's the neon patch:
--- src/ne_xml.c 22 Sep 2002 08:40:25 -0000 1.69
+++ src/ne_xml.c 8 Oct 2002 13:09:35 -0000
@@ -714,6 +714,7 @@
if (p->parser == NULL) {
abort();
}
+ p->parser->replaceEntities = 1;
#endif
return p;
}
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: [Issue 907] libxml broken?
Posted by Joe Orton <jo...@manyfish.co.uk>.
On Mon, Oct 07, 2002 at 06:41:50AM -0500, Clark Rawlins wrote:
> I haven't had a chance to look into this for a few days but last night I
> started poking arround the libsvn_ra_dav code to figure out where the
> escaped & were comming from.
...
> Can someone confirm the workarround?
Yes, confirmed that libxml 2.4.23 isn't expanding "&" in attribute
values; unless there is something extra that neon should be doing with
the libxml SAX interface I'd presume this is a libxml2 bug.
> Does anyone disagree that general (i.e. &, ", etc) entities
> found in attributes should be dereferenced before passing them to the
> application layer?
No disagreement here - can you report this bug to the maintainer?
Regards,
joe
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org