Subject: Re: [xsl] Tree Comparing Algorithm|
From: "Martin Honnen martin.honnen@xxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Date: Mon, 3 Feb 2020 20:38:52 -0000
Thanks both. Martin's solution sort of worked, but it only gave me 21 children, but I had around 21000 nodes in the xml. I am not sure to what depth the comparisonB is happening.
It was solely an attempt to try to find some way to recursively process two documents with streaming at the same time, not an attempt to implement your particular algorithm.
I have tested my code now on a large files, it seems to process lots of nodes judging by the output to the console and the length of processing, but it doesn't seem to use streaming when I look at the memory consumption (600MB of input needed more than 2GB of memory), even if Saxon nowhere shows any -t message that input trees were built.
Michael's comment on the way streaming is implemented in Saxon suggests that the whole attempt is futile, even if the code somehow manages to get by the streamability analysis.