Re: [xsl] Avoiding boneheaded mistakes in XSLT?

Subject: Re: [xsl] Avoiding boneheaded mistakes in XSLT?
From: Wolfgang Laun <wolfgang.laun@xxxxxxxxx>
Date: Thu, 30 Dec 2010 10:52:04 +0100
Rather than asking for an ad-hoc solution for this or that: XSLT might
very well, by definition in the language and not as a vendor
extension, provide a general mechanism for implementing safeguards for
many situations. (I haven't thought long about this, so forgive me if
it's already out there or has a serious flaw.)

A new XSLT-element:
   <xsl:hook name = expression >...</xsl-hook>

This new element xsl:hook is called (as if it were a named template)
whenever a <xsl:instr-name> instruction is executed. Within its body,
a function hook-attr( "attr-name" ) can be used to access the actual
values of <xsl:instr-name>. Also, hook-line() might be useful, and, if
a general @id were permitted for every <xsl:*>, hook-id() could be
used to filter certain instantiations of the category.

Happy New Year (and thank you all for the help)

On 30 December 2010 10:05, Dave Pawson <davep@xxxxxxxxxxxxx> wrote:
> On Wed, 29 Dec 2010 23:49:53 -0500
> Graydon <graydon@xxxxxxxxx> wrote:
>> > Even without schema awareness/context, surely the processor can
>> > tell when an xpath expression will result in zero match?
>> But a zero match is completely legitimate, isn't it?
>> If I'm matching on para/b, and much of my input doesn't have it, and
>> some of it does, I shouldn't get a warning because I'm looking for
>> something that isn't in this particular input file.
>> Or am I completely mis-understanding what you want, there?
> Agreed, there are cases where this might be wanted as a response
> from an xpath expression. Hence a warning. Hence 'not all the time'.
> But consider the typo that we all make from time to time .... to time.
> If it can help find that, it helps debugging.
> Once you've stared at a stylesheet for minutes on end and read what
> you thought you wrote, rather than what you actually wrote....
> Then any debug info helps!
> --
> regards
> --
> Dave Pawson

Current Thread