On Fri, Jan 8, 2021 at 5:44 PM JonnyRobbie via gegl-developer-list <gegl-developer-list@gnome.org> wrote: > > > Maybe having > > everything in the .c single file with enums and forward declarations > > for the data early and the data after the operation code? > > Locally, I use an external python script in a meta-programming way to > generate the two included files from some input data. It doesn't depend on > any external library, but the main file depens on the two includes. I've > found it much easier to generate the includes from scratch and then simply > include them in the main c file rather then messing with file templats, etc. > in python. The two parts I generate are enums for property_enum() for ui > input (generated in the negative-darkroom-curve-enum.c) and the second > (negative-darkroom-curve-enum.h) is a struct initialization with all the > data. You can see the structure in the three attached files. > But if you insist on having only one file, I might be able to rewrite that. > But messing with templating (locally - the PR for gegl would be clean) is a > pita.
The mailing-list strips attachments, you could generate the c file put into GEGL on your side, though there is some precedent for having external files in a fashion that could be followed while avoiding having non-op sibling files in common/, the perlin-noise op which includes files from gegl/operations/common/perlin. /pippin _______________________________________________ gegl-developer-list mailing list List address: gegl-developer-list@gnome.org List membership: https://mail.gnome.org/mailman/listinfo/gegl-developer-list
