Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

Interrupting ctrace leaves the running tool and its run directory behind

Open
#188 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
50/100
Issue type
Bug
Clarity
Clearly specified
Activity status
Active
Tech stack
cpp
Domain
backend, cli

Research direction

Start with the process spawning in include/Process/UnixProcess.hpp around lines 68-77, where the tool gets its own process group, and with the RunDirectory cleanup. Then find where POST /shutdown drains server mode so the signal path can reuse it. Done means that after Ctrl+C or SIGTERM no tool process or run directory remains, the exit codes are 130 and 143, and an integration test with a hung fake tool checks both.

Written by the indexing model from the issue text.

Description

bug severity/major

Interrupting ctrace while an external tool runs stops ctrace but not the tool: the tool keeps running, orphaned, and the run directory stays in the temporary directory.

Observed

main eba7f88, macOS 15.6.1. A fake cppcheck that sleeps 600 s is first in PATH; ctrace --input ok.c --invoke cppcheck is interrupted after 2 s:

Signal Sent to ctrace Tool process Run directory
SIGINT (Ctrl+C) the process group exits 130 within 1 s still running left behind
SIGTERM the process group exits 143 at once still running left behind
SIGINT ctrace only exits 130 within 1 s still running left behind
SIGTERM ctrace only exits 143 at once still running left behind

Each interrupted run leaves one ctrace-cppcheck-* directory in $TMPDIR. A server stopped with SIGINT or SIGTERM also exits at once, without the drain that POST /shutdown performs.

Cause

  • ctrace installs no signal handler: SIGINT and SIGTERM take their default action, so no destructor runs (RunDirectory is not removed).
  • With a timeout, which every external tool has by default, the tool is started in its own process group (include/Process/UnixProcess.hpp:68-77), so the terminal's Ctrl+C does not reach it, and nothing kills it when ctrace dies.

Fix

On SIGINT and SIGTERM, stop the running tools (SIGTERM to their process group, then SIGKILL after a grace period), let the run directories be removed, and exit with 128 + the signal. In server mode, start the same drain as POST /shutdown.

Acceptance

  • After Ctrl+C or SIGTERM during a run, no tool process and no run directory remain.
  • The exit code is 130 for SIGINT and 143 for SIGTERM.
  • An integration test interrupts a run with a hung fake tool and checks both.

Found by the resilience campaign (docs/quality/resilience.md, scenarios C31 to C34, S18, S21).

Dominant language
C++
Stars
5
Forks
4
Avg merge
1h 10m
Merged PRs (30d)
79

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from CoreTrace/coretrace

All issues in CoreTrace/coretrace

Similar issues

More C++ issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.