[PATCH PR gas/33030] gas: accept leading zeros on dollar local labels in z80 sdcc compat mode

[Maciej W. Rozycki]

On Sun, 6 Jul 2025, Jan Beulich wrote:

> >> Worth noting, the commit that first broke SDCC compatibility did run the
> >> z80 tests and saw the test failure. They just modified the test to pass.
> > 
> >  Weird, the need to adjust a testcase should have been a red flag!
> 
> I indeed was surprised to find something like that. Without things being
> explained anywhere that I could sensibly spot, I could only take it as
> "odd", and hope for someone to reply to the patch submission if they knew
> more than I did. _That_ imo is the main problem - patches aren't being
> looked at while they're still uncommitted. (The situation is of course
> generally worse for targets without any maintainers.)

 It depends on the situation with specific ports, but I wouldn't overall 
expect port maintainers to look at all the generic patches posted unless 
cc'd on them.  I certainly don't, though I can spot things by chance from 
time to time.  There's just too much traffic to handle to have a look at 
each and every individual message and I can reasonably expect maintainers 
to be duly dilligent in making/approving the changes that go in.

 Since there's no Z80 port maintainer there's nobody to cc, though in this 
case I'd have tried asking the original author of the test case.  Things 
happen though and it's good the issue has now been spotted and covered.  
It can take decades sometimes for things to get noticed and fixed, as I 
have experienced with the MIPS patch series I have just pushed, where the 
handling of the generic linker has been broken for some relocations since 
forever.

  Maciej