ponder-lab / Hybridize-Functions-Refactoring / #3473
90%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 46
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: 88.369%. Remained the same
#3473

push

github

web-flow 
Pin the untraceable-Python check's two misses on a branch condition (#1084)

* Pin the untraceable-Python check's two misses on a branch condition

The check asks whether a branch condition is derived from a parameter,
when what makes a branch a tf.cond is that its condition is a tensor.
Measured on TensorFlow 2.9.3, two cases follow from that and are pinned
as wrong today:

- mask_or_none tests a parameter with `is None`, a Python bool decided
  while tracing, and traces with and without the argument, but is
  refused (FIXME).
- dataset_size branches on dataset.cardinality(), a tensor, with paths
  disagreeing about None, so the traced function raises ValueError, but
  it is not refused (TODO; also wala/ML#1072).

Part of #1083.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NTnpjPjJj2aZy1ZkUD1eAn

* Assert the check runs on dataset_size before pinning its miss

The untraceable-Python verdict is left undetermined when the check does
not run, as it would if dataset_size's parameter failed the tensor-
parameter gate. Asserting that the verdict is a determinate false shows
the check reaches the function and misses its branch, rather than never
running, so a later fix cannot leave this pin green while the check
still skips it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NTnpjPjJj2aZy1ZkUD1eAn

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

5136 of 5812 relevant lines covered (88.37%)

0.88 hits per line

Jobs
ID Job ID Ran Files Coverage
1 #3473.1 46
88.37
Source Files on build #3473