desktop-apps/eosim and eostudio are skeletons shadowing the real repos; 20 tests assert True

[srpatcha]

desktop-apps/eosim and desktop-apps/eostudio are skeletons that shadow the real repositories by name, and the README documents commands against them that do not work.

What is there

	files	total lines
desktop-apps/eosim	158	162
desktop-apps/eostudio	58	62
desktop-apps/edb	56	4963

138 of eosim's 158 Python files are a single line. Its entire CLI is:

# eApps/desktop-apps/eosim/eosim/cli/main.py
def main(): print('EoSim CLI')

The real one — embeddedos-org/EoSim, eosim/cli/main.py — is 1013 lines and implements 20 commands.

The documented command fails

README line 210:

cd desktop-apps/eosim && pip install -e . && python -m eosim

Step two does not work:

error: Multiple top-level packages discovered in a flat-layout: ['eosim', 'platforms'].

Line 207 documents the same flow for desktop-apps/eostudio.

20 tests that cannot fail

# eApps/desktop-apps/eosim/tests/test_eosim_05.py
# test 5
def test_placeholder_5(): assert True

20 of eApps' 44 test files are of this form. They pass unconditionally and assert nothing. Nothing else in the organisation does this — I scanned all 15 repositories and eApps is the only one.

They are also not currently run: eApps has no root pyproject.toml, so the ecosystem runner detects it as a CMake project, builds it, and reports no tests registered. That report is accurate, which is the problem — the 44 test files are invisible to it.

Why this matters beyond tidiness

§22 assigns eApps the role "Package/application ecosystem" — distribution, not implementation. A skeleton that carries the name of a real product and cannot run is the same failure as the 84 board descriptors of which 71 are generic-*, or a networking subsystem whose connect() returns -1: structure that reads as capability.

Someone following the README concludes EoSim is a one-line stub.

Options, none of which I have taken

This is a product decision rather than a mechanical fix:

1. Delete the skeletons and have desktop-apps/ point at the real repositories — most consistent with §22's distribution role and with §24's rule against two names for one technology.
2. Make them genuinely thin launchers that depend on the published eosim / eostudio packages, so the directory is a packaging shim rather than a copy.
3. Fill them in, if eApps is meant to carry desktop variants that differ from the upstream tools — in which case the placeholder tests should go first, since they currently certify nothing.

Whichever way it goes, the 20 assert True tests should be removed rather than carried forward. A test that cannot fail is worse than no test: it makes a coverage number and a green run mean less everywhere else in the repository.

Verification

All figures measured on 29 Aug 2026: file and line counts by find/wc, the install failure reproduced in a clean venv, the placeholder count by grep across all 15 cloned repositories.