Re: Whose error is it anyway

So am losing track of exactly what situations lead to what error but here
is a clue

This error here was cleaned up by correcting the type declaration on the
map to allow the value to contain a sequence of strings.

Now is what you have below what you want to be putting out in response to
what is in effect a run time detection of a typing error. I don't think so.

 XTTE0570  An error occurred matching pattern {element()[let $action :=
  first(tokenize(convertUntyped(data(.)), "\s|-", "")) return
((matches($action, "\i\c*",
  "")) and (Q{http://www.w3.org/2005/xpath-functions/map}contains($actions,
  exactly-one($action))))]}:

On Sat, Aug 15, 2020 at 7:25 PM Ihe Onwuka <[email protected]> wrote:

> There is a type declaration of map(xs:string,xs:string) on the map
> $actions. So if a zero length key or value showed up..........
>
> On Sat, Aug 15, 2020 at 3:43 PM Ihe Onwuka <[email protected]> wrote:
>
>> The value of $action as at time of error contained a forward slash.
>>
>> I can recreate and answer any other questions that may follow now.
>>
>> On Fri, Aug 14, 2020 at 5:09 PM Trevor Nash <[email protected]> wrote:
>>
>>> Does the inconsistency with \W and \s mean that
>>> {map:get($actions,$action)} could evaluate to empty exactly as the (second)
>>> error message suggests?  In other words the match expression does not
>>> guarantee that $action is found in the map.
>>>
>>> Trevor
>>> On 14/08/2020 13:54, Ihe Onwuka wrote:
>>>
>>> I have since changed the code but this is the rest of the template rule
>>> but trying to reconstruct it
>>>
>>> <xsl:template
>>> match="*[map:contains($actions,tokenize(.,'\W')[1][matches(.,'\i\c*')])]">
>>>       <xsl:variable name="action" as="xs:string+"
>>> select="tokenize(.,'\s')[1]"/>
>>>       <xsl:copy>
>>>          <xsl:variable name="encodedAction">
>>>              <xsl:element
>>> name="{map:get($actions,$action)}">{normalize-space(substring-after(.,$action))}</xsl:element>
>>>          </xsl:variable>
>>>          <xsl:apply-templates select="$encodedAction"/>
>>>       </xsl:copy>
>>>    </xsl:template>
>>>
>>> Note an error I overlooked - the first tokenize used '\W' and the second
>>> '\s' which led to the situation.
>>>
>>> I wasn't going to mention this but now that I have pasted the whole code
>>> when I corrected the 2nd tokenize to '\W'  the transform threw a bash shell
>>> error (invalid fork argument)
>>>
>>>
>>>
>>>
>>> On Fri, Aug 14, 2020 at 8:46 AM Ihe Onwuka <[email protected]> wrote:
>>>
>>>> line 67
>>>> <xsl:element
>>>> name="{map:get($actions,$action)}">{normalize-space(substring-after(.,$action))}</xsl:element>
>>>>
>>>> where the @name did not evaluate to a valid XML name.
>>>>
>>>> On Fri, Aug 14, 2020 at 8:20 AM Michael Kay <mike-JkSD5nQpfvpWk0Htik3J/[email protected]> wrote:
>>>>
>>>>> What's on line 67? Is it the declaration of global variable $actions?
>>>>> Otherwise, I can't see why evaluating this pattern should result in a call
>>>>> on xsl:element.
>>>>>
>>>>> Michael Kay
>>>>> Saxonica
>>>>>
>>>>> On 14 Aug 2020, at 13:04, Ihe Onwuka <[email protected]> wrote:
>>>>>
>>>>> Happened again.
>>>>>
>>>>> My code. <xsl:template
>>>>> match="*[map:contains($actions,tokenize(.,'\W')[1][matches(.,'\i\c*')])]">
>>>>>
>>>>> Clearly it doesn't like the zero length string.
>>>>>
>>>>> Warning at char 1 in xsl:template/@match on line 63 column 92 of
>>>>> *****.xsl:
>>>>>   XPTY0004  An error occurred matching pattern
>>>>>   {element()[Q{
>>>>> http://www.w3.org/2005/xpath-functions/map}contains($actions,
>>>>>   exactly-one((first(tokenize(convertUntyped(data(.)), "\W",
>>>>> "")))[matches(., "\i\c*", "")]))]}:
>>>>> Error at char 9 in expression in xsl:element/@name on line 67 column
>>>>> 55 of ******.xsl:
>>>>>   XTDE0820  Supplied element name is a zero-length string
>>>>>
>>>>> On Thu, Aug 13, 2020 at 6:40 PM Ihe Onwuka <[email protected]>
>>>>> wrote:
>>>>>
>>>>>> I tried but cannot reconstruct a snippet that reproduces. The hint as
>>>>>> to what happened in my use case was the best I could offer.
>>>>>>
>>>>>> On Thu, Aug 13, 2020 at 6:15 PM Michael Kay <mike-JkSD5nQpfvpWk0Htik3J/[email protected]>
>>>>>> wrote:
>>>>>>
>>>>>>> If you post a repro that we can run, that's useful as we can examine
>>>>>>> the error messages, see what's causing them, and test any improvements.
>>>>>>> With only snippets like this, there's not really any effective action we
>>>>>>> can take.
>>>>>>>
>>>>>>> Michael Kay
>>>>>>> Saxonica
>>>>>>>
>>>>>>> On 13 Aug 2020, at 18:53, Ihe Onwuka <[email protected]> wrote:
>>>>>>>
>>>>>>> I can tell you that the error went away when I eliminated the
>>>>>>> possibility of the context element being empty. I struggle to correlate
>>>>>>> that back to the error message that was displayed.
>>>>>>>
>>>>>>> On Thu, Aug 13, 2020 at 10:16 AM Graydon <[email protected]>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> Both?  The hints about what the optimizer thinks is happening are
>>>>>>>> useful right up until I can't tell where the error is.
>>>>>>>>
>>>>>>>> On August 13, 2020 1:40:48 PM UTC, Michael Kay <mike-JkSD5nQpfvpWk0Htik3J/[email protected]>
>>>>>>>> wrote:
>>>>>>>> >Yes, we should probably retain the original pattern as a string
>>>>>>>> rather
>>>>>>>> >than trying to reconstitute it from the compiled form.
>>>>>>>> >
>>>>>>>> >Michael Kay
>>>>>>>> >Saxonica
>>>>>>>> >
>>>>>>>> >> On 13 Aug 2020, at 13:52, Ihe Onwuka <[email protected]>
>>>>>>>> wrote:
>>>>>>>> >>
>>>>>>>> >> This is not my code.
>>>>>>>> >>
>>>>>>>> >> Warning at char 1 in xsl:template/@match on line 62 column 72 of
>>>>>>>> >******.xsl:
>>>>>>>> >>   XPTY0004  An error occurred matching pattern
>>>>>>>> >>
>>>>>>>> >{element()[Q{
>>>>>>>> http://www.w3.org/2005/xpath-functions/map}contains($actions
>>>>>>>> ><http://www.w3.org/2005/xpath-functions/map%7Dcontains($actions>,
>>>>>>>> >>   exactly-one(first(tokenize(convertUntyped(data(.)), "\s",
>>>>>>>> ""))))]}:
>>>>>>>> >>
>>>>>>>> >> My code was
>>>>>>>> >>
>>>>>>>> >>
>>>>>>>> >>    <xsl:template
>>>>>>>> >match="*[map:contains($actions,tokenize(.,'\s')[1])]">
>>>>>>>> >> _______________________________________________
>>>>>>>> >> saxon-help mailing list archived at http://saxon.markmail.org/
>>>>>>>> >> [email protected]
>>>>>>>> >> https://lists.sourceforge.net/lists/listinfo/saxon-help
>>>>>>>> >
>>>>>>>> >_______________________________________________
>>>>>>>> >saxon-help mailing list archived at http://saxon.markmail.org/
>>>>>>>> >[email protected]
>>>>>>>> >https://lists.sourceforge.net/lists/listinfo/saxon-help
>>>>>>>>
>>>>>>>> --  Sent from my p≡p for
>>>>>>>> Android._______________________________________________
>>>>>>>> saxon-help mailing list archived at http://saxon.markmail.org/
>>>>>>>> [email protected]
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/saxon-help
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> saxon-help mailing list archived at http://saxon.markmail.org/
>>>>>>> [email protected]
>>>>>>> https://lists.sourceforge.net/lists/listinfo/saxon-help
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>
>>>
>>> _______________________________________________
>>> saxon-help mailing list archived at http://saxon.markmail.org/[email protected]://lists.sourceforge.net/lists/listinfo/saxon-help
>>>
>>> _______________________________________________
>>> saxon-help mailing list archived at http://saxon.markmail.org/
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/saxon-help
>>
>>

_______________________________________________
saxon-help mailing list archived at http://saxon.markmail.org/
[email protected]
https://lists.sourceforge.net/lists/listinfo/saxon-help