ponder-lab / Hybridize-Functions-Refactoring / #2650
83%
main: 84%

LAST BUILD BRANCH: fed0cbf564e64cc505c53a264e244d915a2c46bd
DEFAULT BRANCH: main
Ran
Jobs 1
Files 36
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 83.491% (+0.03%) from 83.462%
#2650

push

github

web-flow 
Bump Ariadne to 0.52.76 (#828)

* Bump Ariadne to 0.52.76

Advances the pin forty releases (0.52.36 to 0.52.76); ML master still targets WALA 1.8.0, so the target's WALA entries, the frozen-importlib fragment (whose host is version-free), and the reactor are untouched, and the full build compiles without an API break.

The suite caught exactly two behavioral deltas, both strict precision improvements, and their pins advance with it. `testTargetedCfaDepthSoundness`: the wildcard (null-dims) over-approximation that covered the test call site through 0.52.36 now resolves to the concrete `(10000, 10)` shape, so the depth-invariant sound set is the two concrete call-site shapes. `testGpt2GetLossVendored`: `real` and `pred` now type with the batched dataset element's batch axis concrete (`padded_batch(32, ([-1], [-1]))` makes the runtime element rank-2 `(32, None)`, so the old rank-1 pin had codified the unbatched element), with the extent symbolic in the context that does not recover it; `pred` additionally sheds the spurious rank-4 `(Dynamic, Dynamic, 8, 8)` member, the attention-internal d_model shape that had leaked into the call result.

The bump also makes #807 actionable (`LIST_APPEND_CONTENTS_FIELD`, wala/ML#773, exists from 0.52.58), which follows separately.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GcqABsog1rtb4m2CDZdcRD

* Address review: bring the pinned tests' Javadoc to 0.52.76

The assertions advanced with the bump but three prose spots still described the 0.52.36 behavior: the depth test's set composition, the gpt-2 rank-4 member, and the 0.52.36-era comment reading as current. Each now scopes its history and states the pinned 0.52.76 behavior.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GcqABsog1rtb4m2CDZdcRD

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

2918 of 3495 relevant lines covered (83.49%)

0.83 hits per line

Jobs
ID Job ID Ran Files Coverage
1 #2650.1 36
83.49
Source Files on build #2650