Re: [xsl] Tool that measures the performance of an XSLT program at a fine granularity?

Subject: Re: [xsl] Tool that measures the performance of an XSLT program at a fine granularity?
From: "Michael Kay mike@xxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Date: Mon, 27 Jul 2020 10:18:32 -0000
> I guess, saving value of current-dateTime() at two different places within
an XSLT stylesheet, and computing the difference later might give us the time
spent by the enclosed XSLT instructions.
> For your use case, I think to do this, extracting the component expressions
into their own variables might help to utilize the approach I've suggested.
>
>

No, current-dateTime() is explicitly defined to give constant results
throughout an evaluation. That's because there isn't any defined order of
execution, so "later" has no meaning.

There is a function saxon:timestamp() designed to get around this. But at this
level of granularity the results are going to be meaninless.

Michael Kay
Saxonica

Current Thread