Mailing-List: contact dev-help@forrest.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@forrest.apache.org
Received-SPF: pass (athena.apache.org: domain of sjurnm@mac.com designates
 17.148.16.98 as permitted sender)
MIME-version: 1.0
Content-transfer-encoding: 7BIT
Content-type: text/plain; charset=us-ascii
Subject: Re: ElemTemplateElement error: carry-body-attribs caused by r887050
From: Sjur Moshagen <sjurnm@mac.com>
In-reply-to: <829D58CB-CF0B-4443-87DA-3ACC94ADF383@mac.com>
Date: Wed, 19 May 2010 23:48:49 +0300
Message-id: <12921DC4-090E-4084-B0DC-7BF63E7D7385@mac.com>
References: <1E18B5FA-8DBA-421C-AEE9-7EC620C2E1E0@mac.com>
 <20100513015440.GC3186@beaver.dehavilland> <20100513085019.GB66169@igg.local>
 <829D58CB-CF0B-4443-87DA-3ACC94ADF383@mac.com>
To: dev@forrest.apache.org

Den 13. mai. 2010 kl. 15.36 skrev Sjur Moshagen:

> Den 13. mai. 2010 kl. 11.50 skrev David Crossley:
>> In default sites, "linkmap.html" is the first document to
>> be processed.
>> 
>> Remember that "linkmap" is a special pipeline.
>> 
>> To help isolate the problem, try setting the "start-uri"
>> to be something else (in forrest.properties)
>> e.g. project.start-uri=index.html
> 
> I will try that as soon as I have some more time.

Now I got the time, and here is the result:

X [0]                                     index.html	BROKEN: ElemTemplateElement error: carry-body-attribs

I tried with a couple of other start pages as well, and they all gave the same result.

>> Don't know if that is of any help.
>> 
>> ---oOo---
>> 
>> Sjur, you are talking about a Dispatcher-based site, correct?
> 
> Yes, but the bug is showing up whether or not I have dispatcher enabled. I stripped the plugins down to only pdf, and I still got that bug.
> 
>> Are you sure that r887050 is the cause?
>> Remember that the svn merge of the new Dispatcher branch
>> happened one day prior to that date, 03 December 2009.
> 
> I know, and that was what I suspected as well. But so far r887050 is the one that triggers the bug. That doesn't mean that the bug hasn't been introduced earlier, but somehow was hidden until this revision.

I found the bug - it is on our side. It turned out that at some point the default skin directory tree was copied from the forrest code to our project code, and perhaps slightly modified. This local skin copy of course was not kept in sync with the improvements and bug fixes made to the common skin files were transfered to our local copy. After I updated all local files to be in sync with the Forrest ones, the bug disappeared.

I guess that before that specific commit, only parts of the local skin files were used, as the pre-lm solution probably didn't as thoroughly ckeck for local overrides of all files. That would explain why the site had worked before, and why not after this commit.

Anyway - this is a lession to our group, and perhaps a relief regarding the Forrest code: there was no bug in Forrest:)

Best regards,
Sjur