[oharboe]

Before, when detailed placement failed in CTS, cts.tcl caught the error, wrote 4_1_error.odb (and its .sdc), and raised the error again with a Detailed placement failed in CTS: prefix. It did this twice, once for each detailed_placement call.
Now detailed_placement fails with its own error, and the step fails with it.

This is the open question #4804 left for a separate pull request. The rationale is the same as #4804's:

• A debug checkpoint is an ad-hoc edit now. Debugging happens in a fast loop, often by an AI: a checkpoint is one orfs_write_db line, added for the session and removed again. Nothing reads 4_1_error.odb: no make target, no OPEN_GUI_SHORTCUT, no test.
• cts.tcl should say what CTS does. Ten lines of error plumbing around two calls hid that the step is CTS, then detailed placement, then repair and detailed placement again.
• Fewer merge conflicts. Lines that are not part of the flow still collide with changes that are.

What changes

• cts.tcl: both catch blocks around detailed_placement_helper are gone; the calls stand on their own.
• EndpointSlackTracking.md named 4_1_error.odb as its example of an intermediate .odb that gets an endpoint slack snapshot. The example is now an ad-hoc checkpoint; *-failed ones still get none, as write_endpoint_slack skips them.

Not changed

docs/contrib/MakefileContract.md and TclContract.md say a failing step writes <step>-failed.odb and calls error, as global_place, detail_place, pdn and grt do. After this, CTS writes no failure .odb. It never wrote one under the contract's name either. Whether the rule stays, and CTS gets a 4_1_cts-failed.odb, or the other stages lose theirs the same way, is for a separate pull request.

Tested

On a successful run the step runs the same commands as before; only the failure path changes. tclint and tclfmt --check 0.7.0 pass on cts.tcl. I have not forced a detailed placement failure to see the new log. That, and the flow runs, are left to CI.

🤖 Generated with Claude Code

https://claude.ai/code/session_018PFLaD4tC97s5fueStBjHE

[gemini-code-assist]

Code Review

This pull request removes the CTS_SNAPSHOTS configuration variable and simplifies the CTS script by removing the save_progress procedure and its associated error-catching blocks. The documentation is updated to reflect these changes. The review feedback correctly identifies an inaccuracy in the updated documentation regarding *-failed snapshots, which are not actually created due to an early return in the underlying script, and provides a suggestion to correct it.

[openroad-ci]

🔍 QoR check

Metrics reflect the PR merge build — i.e. what will land on the target branch.

Commit 24b6ab1 · Jenkins build #3 · Baseline: build · View build on dashboard

62 design(s) checked — 0 with regression(s), 0 without a comparable baseline.