You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cocoon.apache.org by Nicola Ken Barozzi <ni...@apache.org> on 2004/09/03 08:35:30 UTC
cocoon:/ protocol bug in 2.2?
I've updated Cocoon to use 2.2-dev to test the new pass-through stuff,
and one thing is not working as before. In a subsitemap, I do a cocoon:/
call and it does not work. It works when I put the things to call in the
base sitemap, but wasn't that cocoon:// (which BTW seems to work correctly)?
--
Nicola Ken Barozzi nicolaken@apache.org
- verba volant, scripta manent -
(discussions get forgotten, just code remains)
---------------------------------------------------------------------
Re: cocoon:/ protocol bug in 2.2?
Posted by Andreas Hartmann <an...@apache.org>.
Andreas Hartmann wrote:
> Nicola Ken Barozzi wrote:
>
>>
>> I've updated Cocoon to use 2.2-dev to test the new pass-through stuff,
>> and one thing is not working as before. In a subsitemap, I do a
>> cocoon:/ call and it does not work. It works when I put the things to
>> call in the base sitemap, but wasn't that cocoon:// (which BTW seems
>> to work correctly)?
>>
[...]
> Maybe a slash finds its way into the URI so that
> it results in a cocoon:// call?
Apparently not:
DEBUG (2004-09-22) 15:57.59:476 [sitemap]
(/default/authoring/info-sitetree/sitetree.js)
PoolThread-1/SitemapSourceFactory: Creating source object for
cocoon:/default/authoring/sitetree.xml
OK, I'll do some further investigation.
-- Andreas
Re: cocoon:/ protocol bug in 2.2?
Posted by Andreas Hartmann <an...@apache.org>.
Andreas Hartmann wrote:
> Andreas Hartmann wrote:
>
>> Nicola Ken Barozzi wrote:
>>
>>>
>>> I've updated Cocoon to use 2.2-dev to test the new pass-through
>>> stuff, and one thing is not working as before. In a subsitemap, I do
>>> a cocoon:/ call and it does not work. It works when I put the things
>>> to call in the base sitemap, but wasn't that cocoon:// (which BTW
>>> seems to work correctly)?
I filed a bug:
http://issues.apache.org/bugzilla/show_bug.cgi?id=31363
-- Andreas
Re: cocoon:/ protocol bug in 2.2?
Posted by Andreas Hartmann <an...@apache.org>.
Carsten Ziegeler wrote:
> Looks good to me; and you beat me by one minute in submitting the patch :(
commit wars :)
Sorry for that,
-- Andreas
RE: cocoon:/ protocol bug in 2.2?
Posted by Carsten Ziegeler <cz...@apache.org>.
Looks good to me; and you beat me by one minute in submitting the patch :(
Thanks
Carsten
> -----Original Message-----
> From: news [mailto:news@sea.gmane.org] On Behalf Of Andreas Hartmann
> Sent: Wednesday, September 22, 2004 7:20 PM
> To: dev@cocoon.apache.org
> Subject: Re: cocoon:/ protocol bug in 2.2?
>
> Vadim Gritsenko wrote:
> > Andreas Hartmann wrote:
> >
> >> - Is it allowed to mount sitemaps with an empty URI prefix?
> >> If not, the problem will not appear.
> >> If yes, what can I do about it?
> >
> >
> > Yes, it should be. Can you make a patch fixing it? :)
>
> It seems that the SitemapSource cannot rely on the
> SitemapSourceInfo.prefix to determine the target processor.
>
> That's why I would add a boolean field
>
> SitemapSourceInfo.processFromRoot
>
>
> // does the uri point to this sitemap or to the root sitemap?
> if (sitemapURI.startsWith("//", position)) {
> position += 2;
> info.processFromRoot = true;
> info.prefix = "";
> } else if (sitemapURI.startsWith("/", position)) {
> position ++;
> info.processFromRoot = false;
> info.prefix = env.getURIPrefix();
> } else {
> throw new MalformedURLException("Malformed cocoon URI: "
> + sitemapURI); }
>
>
> SitemapSource:
>
> // does the uri point to this sitemap or to the root sitemap?
> if (info.processFromRoot) {
> this.processor =
> EnvironmentHelper.getCurrentProcessor().getRootProcessor();
> } else {
> this.processor = EnvironmentHelper.getCurrentProcessor();
> }
>
>
> That seems to do the trick, at least in my application.
> If you think it is appropriate and sufficient, I'll
> commit it to the trunk.
>
> -- Andreas
>
>
Re: cocoon:/ protocol bug in 2.2?
Posted by Andreas Hartmann <an...@apache.org>.
Vadim Gritsenko wrote:
> Andreas Hartmann wrote:
>
>> - Is it allowed to mount sitemaps with an empty URI prefix?
>> If not, the problem will not appear.
>> If yes, what can I do about it?
>
>
> Yes, it should be. Can you make a patch fixing it? :)
It seems that the SitemapSource cannot rely on the
SitemapSourceInfo.prefix to determine the target processor.
That's why I would add a boolean field
SitemapSourceInfo.processFromRoot
// does the uri point to this sitemap or to the root sitemap?
if (sitemapURI.startsWith("//", position)) {
position += 2;
info.processFromRoot = true;
info.prefix = "";
} else if (sitemapURI.startsWith("/", position)) {
position ++;
info.processFromRoot = false;
info.prefix = env.getURIPrefix();
} else {
throw new MalformedURLException("Malformed cocoon URI: " + sitemapURI);
}
SitemapSource:
// does the uri point to this sitemap or to the root sitemap?
if (info.processFromRoot) {
this.processor =
EnvironmentHelper.getCurrentProcessor().getRootProcessor();
} else {
this.processor = EnvironmentHelper.getCurrentProcessor();
}
That seems to do the trick, at least in my application.
If you think it is appropriate and sufficient, I'll
commit it to the trunk.
-- Andreas
Re: cocoon:/ protocol bug in 2.2?
Posted by Vadim Gritsenko <va...@reverycodes.com>.
Andreas Hartmann wrote:
> - Is it allowed to mount sitemaps with an empty URI prefix?
> If not, the problem will not appear.
> If yes, what can I do about it?
Yes, it should be. Can you make a patch fixing it? :)
Vadim
Re: cocoon:/ protocol bug in 2.2?
Posted by Andreas Hartmann <an...@apache.org>.
Andreas Hartmann wrote:
> Nicola Ken Barozzi wrote:
>
>>
>> I've updated Cocoon to use 2.2-dev to test the new pass-through stuff,
>> and one thing is not working as before. In a subsitemap, I do a
>> cocoon:/ call and it does not work. It works when I put the things to
>> call in the base sitemap, but wasn't that cocoon:// (which BTW seems
>> to work correctly)?
>>
>
> I can confirm the problem.
>
> Test case: Lenya 1.4-dev (trunk) with Cocoon 2.2 (trunk)
>
> The sub-sitemap is mounted using
>
> <map:match pattern="*/*/info-sitetree/**">
> <map:mount uri-prefix="" src="lenya/info.xmap"
> check-reload="true" reload-method="synchron"/>
> </map:match>
>
> (BTW, is it allowed to have an empty URI prefix here?
> Carsten just told me it has to end with a slash)
>
>
> The cocoon:/ call in info.xmap
>
> <map:match pattern="*/*/info-sitetree/sitetree.js">
> <map:aggregate element="lenya" label="aggregate">
> <map:part src="cocoon:/{1}/authoring/sitetree.xml"/>
> ...
>
> goes directly to the root sitemap instead of staying in
> info.xmap.
I think I tracked it down.
--------------------------------------------------------
SitemapSourceInfo.parseURI() (I added some logging):
// does the uri point to this sitemap or to the root sitemap?
if (sitemapURI.startsWith("//", position)) {
position += 2;
info.prefix = ""; // start at the root
} else if (sitemapURI.startsWith("/", position)) {
position ++;
info.prefix = env.getURIPrefix();
} else {
throw new MalformedURLException("Malformed cocoon URI: " + sitemapURI);
}
logger.debug("Sitemap URI: [" + sitemapURI + "]");
logger.debug("URI prefix: [" + env.getURIPrefix() + "]");
logger.debug("Prefix: [" + info.prefix + "]");
--------------------------------------------------------
results in:
SitemapSourceInfo: Sitemap URI: [cocoon:/default/trash/sitetree.xml]
SitemapSourceInfo: URI prefix: []
SitemapSourceInfo: Prefix: []
I guess the URI prefix is empty because the sitemap was mounted
using an empty URI prefix.
--------------------------------------------------------
SitemapSource:
// does the uri point to this sitemap or to the root sitemap?
if (info.prefix.length() == 0) {
this.processor =
EnvironmentHelper.getCurrentProcessor().getRootProcessor();
} else {
this.processor = EnvironmentHelper.getCurrentProcessor();
}
--------------------------------------------------------
This means the root sitemap processor is used.
Some questions:
- Is this behaviour intended?
- Is it allowed to mount sitemaps with an empty URI prefix?
If not, the problem will not appear.
If yes, what can I do about it?
Thanks in advance,
-- Andreas
Re: cocoon:/ protocol bug in 2.2?
Posted by Andreas Hartmann <an...@apache.org>.
Nicola Ken Barozzi wrote:
>
> I've updated Cocoon to use 2.2-dev to test the new pass-through stuff,
> and one thing is not working as before. In a subsitemap, I do a cocoon:/
> call and it does not work. It works when I put the things to call in the
> base sitemap, but wasn't that cocoon:// (which BTW seems to work
> correctly)?
>
I can confirm the problem.
Test case: Lenya 1.4-dev (trunk) with Cocoon 2.2 (trunk)
The sub-sitemap is mounted using
<map:match pattern="*/*/info-sitetree/**">
<map:mount uri-prefix="" src="lenya/info.xmap"
check-reload="true" reload-method="synchron"/>
</map:match>
(BTW, is it allowed to have an empty URI prefix here?
Carsten just told me it has to end with a slash)
The cocoon:/ call in info.xmap
<map:match pattern="*/*/info-sitetree/sitetree.js">
<map:aggregate element="lenya" label="aggregate">
<map:part src="cocoon:/{1}/authoring/sitetree.xml"/>
...
goes directly to the root sitemap instead of staying in
info.xmap. Maybe a slash finds its way into the URI so that
it results in a cocoon:// call?
-- Andreas
Re: cocoon:/ protocol bug in 2.2?
Posted by Andreas Hartmann <an...@apache.org>.
Nicola Ken Barozzi wrote:
> Carsten Ziegeler wrote:
>
>> Hmm, I just tested latest trunk from SVN and it works for me.
>> For example our samples:
>>
>> http://localhost:8888/samples/sources/xsl-cocoon
>> or the portal work very well and they all use the cocoon:/
>> protocol.
>>
>> Can you provide a test-case?
>
>
> In the base sitemap I put:
>
> <map:pipeline internal-only="false">
> <map:match pattern="*.xml">
> <map:select type="exists">
> <map:when test="subsitemap.xmap">
> <map:mount uri-prefix=""
> src="subsitemap.xmap"
> check-reload="yes"/>
OK, this seems to affirm my results, you mount the sitemap
using an empty URI prefix as well.
-- Andreas
Re: cocoon:/ protocol bug in 2.2?
Posted by Nicola Ken Barozzi <ni...@apache.org>.
Carsten Ziegeler wrote:
> Hmm, I just tested latest trunk from SVN and it works for me.
> For example our samples:
>
> http://localhost:8888/samples/sources/xsl-cocoon
>
> or the portal work very well and they all use the cocoon:/
> protocol.
>
> Can you provide a test-case?
In the base sitemap I put:
<map:pipeline internal-only="false">
<map:match pattern="*.xml">
<map:select type="exists">
<map:when test="subsitemap.xmap">
<map:mount uri-prefix=""
src="subsitemap.xmap"
check-reload="yes"/>
</map:when>
</map:select>
</map:match>
</map:pipeline>
Then in the subsitemap I put:
<map:pipelines>
<map:pipeline internal-only="false">
<map:match pattern="test2.xml">
<map:generate src="index2.xml" />
<map:serialize type="xml" />
</map:match>
</map:pipeline>
<map:pipeline internal-only="false">
<map:match pattern="test.xml">
<map:generate src="cocoon:/index2.internal" />
<map:serialize type="xml" />
</map:match>
</map:pipeline>
<map:pipeline internal-only="true">
<map:match pattern="*.internal">
<map:generate src="{1}.xml" />
<map:serialize type="xml" />
</map:match>
</map:pipeline>
</map:pipelines>
Then test2.xml gives output, test.xml does not.
Note that if I remove the match in the base sitemap, both tests work
even if the internal pipeline in the subsitemap is still hidden from
direct calls.
--
Nicola Ken Barozzi nicolaken@apache.org
- verba volant, scripta manent -
(discussions get forgotten, just code remains)
---------------------------------------------------------------------
RE: cocoon:/ protocol bug in 2.2?
Posted by Carsten Ziegeler <cz...@apache.org>.
Hmm, I just tested latest trunk from SVN and it works for me.
For example our samples:
http://localhost:8888/samples/sources/xsl-cocoon
or the portal work very well and they all use the cocoon:/
protocol.
Can you provide a test-case?
Carsten
> -----Original Message-----
> From: news [mailto:news@sea.gmane.org] On Behalf Of Nicola Ken Barozzi
> Sent: Friday, September 03, 2004 8:36 AM
> To: dev@cocoon.apache.org
> Subject: cocoon:/ protocol bug in 2.2?
>
>
> I've updated Cocoon to use 2.2-dev to test the new
> pass-through stuff, and one thing is not working as before.
> In a subsitemap, I do a cocoon:/ call and it does not work.
> It works when I put the things to call in the base sitemap,
> but wasn't that cocoon:// (which BTW seems to work correctly)?
>
> --
> Nicola Ken Barozzi nicolaken@apache.org
> - verba volant, scripta manent -
> (discussions get forgotten, just code remains)
> ---------------------------------------------------------------------
>
>