Oh well, re-reading this whole thread, I've rallied to Thomas' proposed
new "stack" operator. Its main advantage over named registers is indeed
that it avoids potential issues with register naming when integrating
such steps in very complex pipelines.
Hence I've implemented push, pop and swap modes (the later swaping the
top and antepenultimate elements of the stack), which should address
"my" use case and Jochem's. Actually our both use cases are the same, it
is just that steps are done in the reverse way (and thus the final
omit_fwd pop of Jochem's pipeline is equivalent to the initial omit_inv
push of "mine).
==> https://github.com/OSGeo/PROJ/pull/4036
--
http://www.spatialys.com
My software is free, but my time generally not.
_______________________________________________
PROJ mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/proj
