Re: Suggestions before a new lightning-2.1.4 release?

[Marc Nieper-Wißkirchen]

Am Mo., 31. Okt. 2022 um 19:18 Uhr schrieb Paulo César Pereira de
Andrade <paulo.cesar.pereira.de.andrade@gmail.com>:

>   I would also suggest sanitizing the jit_skip argument. Otherwise it
> could cause an assertion, or, patch the _nop() functions to calculate
> the proper alignment, adding any required extra padding.

By sanitizing, do you mean that the argument is not too large?

>   Maybe could also consider merging jit_align and jit_skip, as they
> serve mostly for the same purpose. Something to think, maybe
> have an extra argument to jit_align. It would be a minor abi
> breakage, but jit_align should not be much used, and easy to
> patch.

I feel that having two different instructions is not too bad (the GNU
Assembler also has a separate align and a skip (pseudo) instruction).
The nice thing about the current instruction formats is that they are
already supported (e.g. by the printer).

> Also note that the reorder of jit_code_* already did break
> the abi. So far this should be a minor issue as lightning is not
> widely used, and patching, if every required should be trivial.

Ah, this is an important point I missed.  Next time, I will add new
instructions at the end of the list.

Marc