Alexander Leidinger wrote:
Am 2023-09-27 13:04, schrieb Guido Falsi:

I'm not completely sure we;re talking about the same thing. The scccache version you're using is an older version with non standard changes (as vishin pointed out, did not check it or the nature of the changes or differences with newer versions). Maybe that version is not network driven thanks to patches, you should check that before assuming that the sccache-overlay version requires networking.

True, I shouldn't assume it behaves as documented right now. But the operlay has code to start and stop the server...

In the port's bsd.overlay.mk, there are custom make(1) targets for doing these, and they are registered as parts of each phase. Technically speaking, just running sccache as a client starts a server if one is not already started or cannot be reached, but that operation mode would not be suitable in a port build.
I agree that it may be interesting to know what the difference is on a code-level. I do not have the spare cycles right now to find the corresponding original code and compare it against the -overlay version (and to try to understand a programming language I have no clue about).

The port's patches simply change from using network sockets to Unix domain sockets. They are quick and dirty and would need to be re-architected for submission upstream, where a big consideration was given (and actively used) towards using cloud storage for cached objects.

--
Charlie Li
...nope, still don't have an exit line.

Attachment: OpenPGP_signature
Description: OpenPGP digital signature

Reply via email to