use STANDARD_TREE as model for saxon tree

[cmangeat]

1500 ms less for edit.xsl transformation.

for the same editor form, tested edit.xsl with:
0: linked -> 900 ms as minimum
1: tiny, the default one -> 2350 ms as minimum
2: tinyc -> 850 ms as minimum

To choose between 0 and 2, it should be considered that 0 could have drawback (memory consumption ?), cf. https://www.saxonica.com/documentation10/index.html#!sourcedocs/choosingmodel ? (in my opinion, there are no extra cost for 'slower tree construction' as xsl transformations are cached with geonetwork).

I do not know if another transformation somewhere could suffer from this (going from 1 to 2 seems to also improve indexation a little bit, once transformation has been cached).

Checklist

• I have read the contribution guidelines
• Pull request provided for main branch, backports managed with label
• Good housekeeping of code, cleaning up comments, tests, and documentation
• Clean commit history broken into understandable chucks, avoiding big commits with hundreds of files, cautious of reformatting and whitespace changes
• Clean commit messages, longer verbose messages are encouraged
• API Changes are identified in commit messages
• Testing provided for features or enhancements using automatic tests
• User documentation provided for new features or enhancements in manual
• Build documentation provided for development instructions in README.md files
• Library management using pom.xml dependency management. Update build documentation with intended library use and library tutorials or documentation

[jodygarnett]

This is a change that camptocamp has been using, would like to share it for others to try as a QA activity with the community.

Q: Can this be a configuration setting, a system property to allow folks to "revert" if it has a regression for them?

[cmangeat]

This is a change that camptocamp has been using, would like to share it for others to try as a QA activity with the community.

Q: Can this be a configuration setting, a system property to allow folks to "revert" if it has a regression for them?

Yes, thanks, looking for making a system property from it, but as trees are cached, it should be decided if a cache clear handle has to be provided or if setting change implies a restart.

This said, maybe can transformation cache live with two or three trees for the same transform and use the asked one when needed, I have to test.

[cmangeat]

Not so clear indeed, sorry for the noise.