1,396 posts

[email protected]

@HaskRay

Maintainer of LLVM MC, Clang driver, lld/ELF, and libunwind

California, USA

[email protected]
@HaskRay
Sep 1
The time from "issue filed" to "plausible patch posted" is now shorter than the time a maintainer needs to understand the bug properly. a stale-analysis crash, from a change I co-authored. A patch showed up clearing the cache at one of the sites that mutates the CFG.
2
[email protected]
@HaskRay
Aug 30
Two kinds of LLM-assisted patches showing up in some areas I maintain in LLVM. First: wrong root cause. Second: issue is real and patch is fine, but the commit message is 4 times longer than it needs to be with comments that restate the code beneath them. mckayla.blog/posts/yap.html
5
[email protected]
@HaskRay
Aug 27
When code is this cheap to produce, saying no is the scarce resource. I don't think maintainers can absorb the volume that's coming....
2
[email protected]
@HaskRay
Aug 27
Another hash table optimization merged today (for FoldingSet, a hash-consing container). stage2-O0-g -0.28%, stage2-O3: -0.20%. Added notes to
maskray.me
Recent LLVM hash table improvements
LLVM has several hash tables. They used quadratic probing with in-band sentinel keys (empty, tombstone); recent work has been replacing that with linear probing with tombstone key removed. DenseMap (
[email protected]
@HaskRay
Aug 27
The bottleneck moved from writing patches to converging on designs