Refactor: sources compilation using sub-processes
Currently libvala
is used in the same process to compile code and extract symbols for LSP services; but found a lot of memory in use.
-
Reduce as low as possible the memory foot print before refactor -
Per file compilation and symbol scan should be saved to a on-disk database for access from main process -
Remove the use of CodeContext
for symbol access, the context should be removed immediately after a compilation was completed -
Compilation should be implemented using a subprocess, that will push symbols to a database and then will release all its resources at quit
Edited by Daniel Espinosa Ortiz