Hi Steven,

I will address this type of bug report once 1.6 is released.

Right now, I prefer to concentrate my energy on getting the release done
right.

Cheers,

Antoine
----- Original Message -----
From: <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Tuesday, December 16, 2003 7:21 PM
Subject: Bug 25564 - CVS tagdiff does not escape XML characters


> Server is down at the moment, so I cannot add to
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=25564
>
> Hi All,
>
> I have been looking through the problem source and the XML is generated by
> hand with code like
>
>             writer.print("<tagdiff ");
>             if (mystartTag != null) {
>                 writer.print("startTag=\"" + mystartTag + "\" ");
>             } else {
>                 writer.print("startDate=\"" + mystartDate + "\" ");
>             }
>             if (myendTag != null) {
>                 writer.print("endTag=\"" + myendTag + "\" ");
>             } else {
>                 writer.print("endDate=\"" + myendDate + "\" ");
>             }
>
> Which is all quite nasty.
>
> Firstlly, the problem could be fixed by adding a CDATA to the offending
> text, but this still leaves potential problems.
>
> My favoured way of generating XML from a class is to extend XMLFilterImpl
> and fire SAX events to a Transformer with no stylesheet and StreamResult
> of the required output file, inspired by Cocoon 2.
>
> Given that frequently the XML output of this task would be transformed to
> HTML using taggdiff.xsl then the processing would be more efficient to use
> that transformer from the outset; this would of course have to be optional
> in order not to break existing build files.
>
> The cvschangelog task is also guitly of handcrafting XML albeit using a
> CDATA for certain elements, again this task is also frequently followed by
> a style task.
>
> CvsTagDiff extends AbstractCvsTask yet ChangeLogTask does not, but both
> tasks are very similar, i.e:
>
>         execute cvs command
>         parse resultant file to memory
>         output memory contents to XML
>
> And are both frequently followed by a style task, such similarity IMO
> should be exploited.
>
> Anyway, I am new to this open source apache stuff so appologies if I have
> omitted and stages or information. I am happy to perform this work myself
> of course.
>
> Steve
>
>



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to