libpcp_web: accept NaN/Inf in series_pmAtomValue_conv_str on float/double

[kurik]

The first-character digit check rejected pmAtomStr_r() output like "nan" from 0.0/0.0 on s390x (positive quiet NaN), breaking pmseries division tests 1886 and 1906.

For some reason the FPU on s390x returns NaN with sign bit clear which is different from most of the other architectures which return negative NaN (sign bit set). Both ways are allowed by IEEE 754 which states only that Nan should be returned (no sign defined).

Edit: I can reproduce this on ppc64le arch as well, so it is not specific for s390x FPUs.

[coderabbitai]

📝 Walkthrough

Walkthrough

The PR strengthens numeric validation in the series_pmAtomValue_conv_str() function by adopting a safer unsigned char cast in the isdigit() check for float and double conversions, improving the robustness of string-to-atom value conversion while maintaining sign prefix handling.

Changes

String Conversion Validation

Layer / File(s)	Summary
Float/Double conversion string validation src/libpcp_web/src/query.c	The PM_TYPE_FLOAT/PM_TYPE_DOUBLE branch now assigns the pmAtomStr_r() result and validates the string prefix with isdigit((unsigned char)*s) instead of isdigit(*s), enforcing stricter type safety while still permitting leading +/- characters.

A digit check so fine, 🐰
Unsigned char cast, designs align,
No more sign confusion,
Just numeric precision,
Atoms convert by design.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately describes the main change: accepting NaN/Inf in the float/double conversion path of series_pmAtomValue_conv_str function.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description check	✅ Passed	The pull request description clearly explains the issue being fixed: rejecting NaN output from pmAtomStr_r() on s390x, and provides technical context about IEEE 754 compliance and architectural differences.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[natoscott]

@kurik LGTM.