Hi Mathieu,
Please understand, that if we follow this path, we have to make a new
release whenever the PSL data changes. And that is not my intention.
What you could do is
a)
Update the git submodule and then rebuild libpsl. You can automate this
with a schedule that suits you best.
b)
Just build and update the DAFSA data file every now and then (use the
src/psl-make-dafsa script for that). In your application, use
psl_load_file().
The loading of DAFSA files is very lightweight - the file is read into
the memory without parsing. The PSL operations are the same for built-in
and for DAFSA data (we use a DAFSA blob internally as well).
c)
Similar to b) except that you load the original PSL file.
It includes a parsing step and a bunch of memory allocations, so it will
have a larger memory footprint.
But the operations on the data are faster than using DAFSA. This only
plays a role when you are doing a significant number of operations per
second.
I really suggest b). But of course it depends on your use case.
Regards, Tim