Re: [sdcc-devel] Picking 1 fix from the next branch to trunk for 4.5.0 RC2

[Benedikt F. <b.f...@gm...>]

Am 12.01.25 um 19:28 schrieb Philipp Klaus Krause:
> Another one, I'd like to pick for RC2:
>
> https://sourceforge.net/p/sdcc/code/15223/
>
> This fixes a bug in the tlcs90 assembler that can result in the silent
> generation of wrong code.
>
> SDCC does not currently generate any of the instructions affected by the
> bug, so this is only be relevant for hand-written assembly.
> tlcs90 is certainly exotic (though a stable port) - I have no idea if
> there are any users left at all -, but while this might reduce the
> importance of the fix, it also reduces the risk: the fix only affects
> assembly for instructions not currently generated by SDCC in an exotic
> port.
>
> Philipp
>
> P.S.: I rebuilt the tlcs90 library, and ran tlcs90 regression tests in
> the next branch on Debian GNU/Linux onamd64. There were no regression
> failures.

Sounds reasonable, but further delays of RC2 should be avoided.

I.e. IMO, this can go into trunk, but RC2 should ideally still happen
today or perhaps tomorrow.

The pending addition of a python installation check for regression
testing may therefore have to be postponed, unless that can be taken
care of today, too, and turns out to be reasonably trivial and thus risk
free.
If both conditions are met, I would proactively favor the inclusion of
that, as well.

Benedikt