docs(openspec): defer the markdown-suites change

Postponed until Robot Framework reports file-accurate positions for tests
in Markdown files.

Author: d-biehl

openspec/changes/markdown-suites/proposal.md

@@ -1,5 +1,7 @@
 # Proposal: markdown-suites
 
+> **Status (2026-09-22): deferred.** Robot Framework reports the positions of tests in embedded suites relative to the code block, not to the file, so the test explorer already lists them with shifted lines; remapping them in RobotCode would only paper over it. Maintainer decision: postponed as a whole — also the editor and `analyze code` part, which could map positions itself — until Robot Framework reports file-accurate positions (model, tokens, listener attributes, `output.xml`); re-implementing the block extraction and a position table in RobotCode is hacking that diverges from Robot Framework's parsing. The requirements are part of the Robot Framework issue. Markdown is the goal; whether reStructuredText (older, docutils-based, more complex) gets supported at all is undecided. No Robot Framework issue exists for this yet (tracker searched 2026-09-22).
+
 ## Why
 
 Robot Framework 7.5 executes tests embedded in Markdown files (`*.robot.md`, and `*.md`/`*.markdown` via `--extension`/`--parseinclude`) and accepts Markdown resources, alongside the reStructuredText support it has had since 3.1. RobotCode only nominally accepts these files: the raw Markdown/reST text is tokenised as Robot syntax, so an imported `keywords.md`/`keywords.rst` resource yields a false "resource is empty" warning and "keyword not found" errors, `.robot.md`/`.robot.rst` suites get no hover, completion, diagnostics or navigation in the editor, `analyze code` silently skips them, `discover files` does not list them, and the test explorer navigates to wrong lines because Robot Framework reports line numbers relative to the concatenated code blocks (an upstream report is planned in `support-rf75`, task 6.1; once filed, its URL is the trigger for switching the remap off).