considering to refactor gen_*_xml functions.

[Chabrier]

Hi,
In some project, we are facing large experimental plan.

In this case using, the std function, this is what we want to, of the package to generate is almost impossible.

The process is very slow and the degradation of the process i far less good than linear (quadratic? exponential?)

Most of the time we have to rewrite the function to directly write the xml as simple text files..That is much more efficient.

We do believe that this does not provide the same confidence from the ingenireeng point of view.

We believe also that refactoring the use of a xml library is very important.

Contributing is possible.

https://record.pages-forge.inrae.fr/-/benchmarkingsticsrfiles/-/jobs/2354540/artifacts/public/index.html to illustrate.

If you don't have access to the bench let me know.

[plecharpent]

Hi Patrick,
I am aware of that performance bottle neck.
For the moment we still use the XML library, which is no longer maintained as I know.
We have had the intent to replace it with the xml2 library but we didn't investigate performances and what is the refactoring consequences (lack of time). But we still want to change.

We are about to make a new release of the package, so we can plan to refactor xml generation function later.

I would like to have access to the benchmark page you mentionned, and of course interested for contibutions.