Even Rouault via PROJ <[email protected]> writes: > Basically CMake concatenates a few .cpp file until it generates a > translation unit that is big enough but not too big. That can result > in twice faster builds (at least for libproj, excluding tests or > proj.db generation). But as code is not always ready to be > concatenated with other files, I had to do quite a lot of changes for > that (like making sure to #undef stuff at end of .cpp files, or that > structures in anonymous namespaces or static functions have unique > names), which by themselves should be harmless (and also makes it > easier to debug, as instead of having >10 static setup() functions, > they have a distinct name now) when doing regular builds.
This strikes me as very strange, and a departure from the documented semantics of the language. Without thinking too much, I would be -1 on committing workarounds to code to make this work, or support for it. It feels like a lot of complexity for no real gain. Assuming you have set up ccache if you are concerned about build times, do you find that it's a real issue? _______________________________________________ PROJ mailing list [email protected] https://lists.osgeo.org/mailman/listinfo/proj
