Reorder dataflow to be earlier in fixed-point pipeline

Dataflow is capable of making large simplifications so its good to have it earlier if we can.

PiperOrigin-RevId: 698542726

Rename sample_cc_test.cc to sample_test.cc

This is testing sample.{h,cc}, so the test should be named
accordingly (probably a leftover to disambiguate from an earlier
Python test ?)

PiperOrigin-RevId: 698163964

github_workflows: clarify docs workflow name

PiperOrigin-RevId: 697829788

In the LSP adapter, properly reference the text to be formatted.

PiperOrigin-RevId: 696971926

github_workflows/doc: publish doc continuously

- enable on push
- enable on pull_request (this will skip deploy per: https://docs.github.com/en/pages/getting-started-with-github-pages/configuring-a-publishing-source-for-your-github-pages-site#creating-a-custom-github-actions-workflow-to-publish-your-site)

follow-up to #1463

PiperOrigin-RevId: 696733559

Add fuzzer-crasher finding an issue in strength reduce.

Issues: https://github.com/google/xls/issues/1711
PiperOrigin-RevId: 696318413

Update LLVM version

PiperOrigin-RevId: 695887700

Fix logic for flag count.

Previously, a span of exactly one value (as all the individuals
were AND-ed together) was passed to c_count(); Replace that with the
logic that was actually meant.

PiperOrigin-RevId: 695533639

Merge pull request #1699 from xlsynth:cdleary/2024-11-06-unhandled-fo…

In the parser, properly capture the start position of (possibly publi…