[llvm] [MC] Increase asm-macro-max-nesting-depth default to 100 (PR #218455)
Matt Turner via llvm-commits
llvm-commits at lists.llvm.org
Sun Sep 6 18:47:34 PDT 2026
mattst88 wrote:
I accept your point that the original commit message was far too verbose. I am sorry. I will make sure they are only as long as they absolutely need to be.
I hesitate to even say this because I'm afraid you'll blackball me, I think you might be being a bit unfair though.
> I'd like the commit message reduced to the motivating case (gas accepts what we reject; glibc's alpha rt_sigaction.S nests 31 deep) and the new value.
The current commit message is
```
The default of 20 rejects assembly the GNU assembler accepts, e.g. glibc's
sysdeps/unix/sysv/linux/alpha/rt_sigaction.S, which nests 31 deep. Raise the
default to 100 to match the GNU assembler, including at the boundary: 100
levels assemble and 101 do not.
Give macro-max-depth.s a depth parameter so one input covers both the default
and the flag overriding it.
```
That doesn't seem outrageous, but I will happily reduce it to
```
GNU as accepts assembly we reject at the current default of 20, e.g.
glibc's sysdeps/unix/sysv/linux/alpha/rt_sigaction.S, which nests 31
deep. Raise the default to 100, matching GNU as.
```
I've also stripped the three one-line comments from the patch.
> I've noticed that this author has submitted a lot of "AI slop" PRs recently. I cannot be certain whether the author fully understands the code and takes ultimate responsibility for it, rather than simply blindly pasting patch descriptions generated by an LLM. Frankly, this makes me much less inclined to merge the PRs they generate.
I'm sorry you feel that way. Again, I will make sure that commit messages and comments going forward are only as long as they absolutely need to be. I have worked in open source and on compilers for 15 years, but I have only recently begun contributing to LLVM and clang so I am very much still learning how to best engage with this community. For what it's worth, I've never had someone question me personally in this manner before. I hope you'll give me another chance.
https://github.com/llvm/llvm-project/pull/218455
More information about the llvm-commits
mailing list