added an example of pure bash coverage#989
Conversation
WalkthroughAdds a Git submodule at Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor Dev as Developer
participant Run as run_cov.sh
participant Cov as pure_bash_coverage/generate_coverage_report.sh
participant Ent as entry.sh
Dev->>Run: Execute run_cov.sh
note right of Run #f0f4c3: set -euo pipefail\nexport INPUT_VERBOSE=true\necho $INPUT_VERBOSE
Run->>Cov: invoke generator with path to ../entry.sh
Cov->>Ent: execute entry.sh under coverage
note right of Ent #e1f5fe: export PS4='+${BASH_SOURCE}:${LINENO}:'\necho $INPUT_VERBOSE\nif INPUT_VERBOSE=='true' then set -x
Ent-->>Cov: completes execution
Cov-->>Run: generates coverage report
Run-->>Dev: exit status returned
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Poem
✨ Finishing Touches
🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 4
🧹 Nitpick comments (8)
.gitmodules (3)
3-3: Trim trailing whitespace in URL line.There’s a trailing space after “.git”. Harmless but noisy in diffs.
Apply:
- url = https://github.com/ostvld/pure_bash_coverage.git + url = https://github.com/ostvld/pure_bash_coverage.git
1-3: Consider shallow submodule to speed CI clones.Adds depth=1 by default; safe for read-only usage.
Apply:
[submodule "calc_coverage/pure_bash_coverage"] path = calc_coverage/pure_bash_coverage url = https://github.com/ostvld/pure_bash_coverage.git + shallow = true
1-3: Confirm repository source/ownership.URL points to a personal fork (ostvld). If this is intended—fine. If not, prefer an org-owned upstream to reduce supply-chain risk.
calc_coverage/pure_bash_coverage (1)
1-1: Guard against missing submodule at runtime.Add a quick existence check in
calc_coverage/run_cov.shto fail fast with guidance if the submodule isn’t initialized.Example snippet to prepend in run_cov.sh:
SUBMOD="${DIR}/pure_bash_coverage/generate_coverage_report.sh" if [[ ! -x "$SUBMOD" ]]; then echo "pure_bash_coverage submodule not found. Run: git submodule update --init --recursive" >&2 exit 1 fientry.sh (3)
5-6: PS4 is fine; consider richer trace prefix (optional).Current PS4 works. If helpful, include function for easier triage.
-export PS4='+${BASH_SOURCE}:${LINENO}:' +export PS4='+${BASH_SOURCE}:${LINENO}:${FUNCNAME[0]}:'
9-11: Make verbose check case-insensitive and use [[ ... ]].More robust to inputs like "True"/"TRUE".
-if [ "${INPUT_VERBOSE}" == 'true' ]; then - set -x -fi +if [[ "${INPUT_VERBOSE,,}" == 'true' ]]; then + set -x +fi
29-31: Duplicate set -x block; remove or guard.You already enable tracing above; this second block is redundant.
-if [ "${INPUT_VERBOSE}" == 'true' ]; then - set -x -fi +# Tracing already enabled earlier when INPUT_VERBOSE=truecalc_coverage/run_cov.sh (1)
3-4: Don’t overwrite pre-set INPUT_VERBOSE; label the echo.Allow callers to override and keep logs consistent with entry.sh.
-export INPUT_VERBOSE="true" -echo "${INPUT_VERBOSE}" +export INPUT_VERBOSE="${INPUT_VERBOSE:-true}" +echo "INPUT_VERBOSE: ${INPUT_VERBOSE}"
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (4)
.gitmodules(1 hunks)calc_coverage/pure_bash_coverage(1 hunks)calc_coverage/run_cov.sh(1 hunks)entry.sh(1 hunks)
🧰 Additional context used
🧬 Code graph analysis (1)
entry.sh (1)
makes/entries.sh (1)
run_test(11-27)
🪛 GitHub Actions: copyrights
calc_coverage/run_cov.sh
[warning] 1-1: Missed: calc_coverage/run_cov.sh.
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: make
🔇 Additional comments (1)
calc_coverage/pure_bash_coverage (1)
1-1: Submodule pinned; license and supply-chain review required
The submodulecalc_coverage/pure_bash_coverageis athttps://github.com/ostvld/pure_bash_coverage.gitpinned to commit60ef6a7795eb67c1dc2d3c82cc486fd589c93eb5. Verify its license compatibility and, if this dependency is critical long-term, consider forking under our org or vendoring it and managing updates via PRs.
| [submodule "calc_coverage/pure_bash_coverage"] | ||
| path = calc_coverage/pure_bash_coverage | ||
| url = https://github.com/ostvld/pure_bash_coverage.git |
There was a problem hiding this comment.
💡 Verification agent
🧩 Analysis chain
Ensure CI fetches submodules (actions/checkout).
If your workflows don’t set submodules: true/recursive, builds will miss this content.
Run this to verify and show offending jobs:
If missing, add:
- uses: actions/checkout@v4
with:
submodules: recursive
fetch-depth: 0🏁 Script executed:
#!/bin/bash
set -euo pipefail
# List workflow steps using actions/checkout and whether they enable submodules
rg -nC3 --glob '.github/workflows/**/*.{yml,yaml}' '\buses:\s*actions/checkout@'
rg -nC2 --glob '.github/workflows/**/*.{yml,yaml}' '\bsubmodules:\s*(true|recursive)'Length of output: 251
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# List all files under .github directory to locate workflows
find .github -type fLength of output: 569
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Search for actions/checkout usage in workflow files
rg -nC3 'uses:\s*actions/checkout@' .github/workflows/*.yml
# Search for submodules:true or recursive in those files
rg -nC2 'submodules:\s*(true|recursive)' .github/workflows/*.ymlLength of output: 7003
Enable submodule fetching in CI workflows
All workflows under .github/workflows use actions/checkout@v5 but omit submodule settings, so the calc_coverage/pure_bash_coverage submodule isn’t cloned. In each workflow’s checkout step, add:
- uses: actions/checkout@v5
with:
submodules: recursive
fetch-depth: 0🤖 Prompt for AI Agents
In .gitmodules around lines 1 to 3, CI workflows currently do not fetch the
calc_coverage/pure_bash_coverage submodule because actions/checkout@v5 steps
lack submodule settings; update every GitHub Actions workflow file under
.github/workflows to modify the checkout step to include submodules: recursive
and fetch-depth: 0 so the submodule is cloned and full history is available for
coverage tasks.
| @@ -0,0 +1 @@ | |||
| Subproject commit 60ef6a7795eb67c1dc2d3c82cc486fd589c93eb5 | |||
There was a problem hiding this comment.
💡 Verification agent
🧩 Analysis chain
Submodule added and pinned — ensure CI initializes it.
Pinning to a commit is good for reproducibility. Make sure your workflows/checkouts fetch submodules; otherwise run_cov.sh will fail when calling ${DIR}/pure_bash_coverage/generate_coverage_report.sh.
- In GitHub Actions, configure checkout:
- uses: actions/checkout@v4
with:
submodules: recursive
fetch-depth: 0
- uses: actions/checkout@v4
Run to confirm CI is set up or to audit locally:
🏁 Script executed:
#!/bin/bash
# Check .gitmodules and submodule status
test -f .gitmodules && cat .gitmodules
git submodule status || trueLength of output: 282
🏁 Script executed:
#!/bin/bash
# Check GitHub Actions workflows for submodule checkout configuration
if [ -d .github/workflows ]; then
grep -R "uses:.*actions/checkout" -n .github/workflows
grep -R "submodules:" -n .github/workflows
else
echo "No .github/workflows directory found"
fiLength of output: 1201
Configure CI to fetch submodules
Your GitHub Actions workflows use actions/checkout@v5 but lack the with: submodules: recursive and fetch-depth: 0 settings. Add these to every checkout step in .github/workflows/*.yml that runs run_cov.sh, for example:
- uses: actions/checkout@v5
with:
submodules: recursive
fetch-depth: 0Without this, calc_coverage/pure_bash_coverage/generate_coverage_report.sh will fail.
🤖 Prompt for AI Agents
In calc_coverage/pure_bash_coverage at lines 1-1: CI workflows that invoke
run_cov.sh may fail because actions/checkout@v5 is used without submodule and
full history settings; update every .github/workflows/*.yml workflow that runs
run_cov.sh to set the checkout step with with: submodules: recursive and
fetch-depth: 0 so submodules are fetched and full git history is available for
the pure_bash_coverage/generate_coverage_report.sh script to run correctly.
| export INPUT_VERBOSE="true" | ||
| echo "${INPUT_VERBOSE}" | ||
| DIR="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" &>/dev/null && pwd)" | ||
| "${DIR}/pure_bash_coverage/generate_coverage_report.sh" "${DIR}/../entry.sh" |
There was a problem hiding this comment.
@ostvld what will it generate? Will the file be suitable for codecov.io?
| export INPUT_VERBOSE="true" | ||
| echo "${INPUT_VERBOSE}" | ||
| DIR="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" &>/dev/null && pwd)" | ||
| "${DIR}/pure_bash_coverage/generate_coverage_report.sh" "${DIR}/../entry.sh" |
There was a problem hiding this comment.
@ostvld, The script interface is generally in Russian. I believe this could a problem for international developers to read logs and coverage statistics.
The script should be full in English.
|
@ostvld Hey there! 👋 Heads up about your branch name " |
There was a problem hiding this comment.
Actionable comments posted: 1
♻️ Duplicate comments (1)
entry.sh (1)
6-6: Gate the noisy “INPUT_VERBOSE” banner behind verbose mode.Avoids extra log chatter when not debugging.
-echo "INPUT_VERBOSE: ${INPUT_VERBOSE}" +if [ "${INPUT_VERBOSE}" = 'true' ]; then + echo "INPUT_VERBOSE: ${INPUT_VERBOSE}" +fi
🧹 Nitpick comments (1)
entry.sh (1)
5-5: Make xtrace prefix shorter and easier to scan.Using only the basename reduces noise; adding a trailing space improves readability of traces.
-export PS4='+${BASH_SOURCE}:${LINENO}:' +export PS4='+${BASH_SOURCE##*/}:${LINENO}: '
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (3)
.gitmodules(1 hunks)calc_coverage/pure_bash_coverage(1 hunks)entry.sh(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (2)
- .gitmodules
- calc_coverage/pure_bash_coverage
🧰 Additional context used
🧬 Code graph analysis (1)
entry.sh (1)
makes/entries.sh (1)
run_test(11-27)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: make
| if [ "${INPUT_VERBOSE}" == 'true' ]; then | ||
| set -x | ||
| fi | ||
|
|
There was a problem hiding this comment.
💡 Verification agent
🧩 Analysis chain
set -x can leak tokens/PII into logs — localize tracing.
Even with GitHub’s masking, echoed CLI args like --token=… may slip into traces. Keep tracing opt-in but disable it around sensitive commands.
Minimal change here to track trace state:
-if [ "${INPUT_VERBOSE}" == 'true' ]; then
- set -x
-fi
+TRACE_ON=false
+if [ "${INPUT_VERBOSE}" = 'true' ]; then
+ set -x
+ TRACE_ON=true
+fiThen, around sensitive calls elsewhere (pull/download/push), bracket with:
# before a command that prints tokens/credentials:
if [ "${TRACE_ON}" = 'true' ]; then set +x; fi
${JUDGES} "${gopts[@]}" pull --timeout=0 "--token=${INPUT_TOKEN}" "--owner=${owner}" "${name}" "${fb}"
if [ "${TRACE_ON}" = 'true' ]; then set -x; fiOptionally route xtrace to a dedicated FD for cleaner logs:
# near the trace enablement
exec 3>&2
export BASH_XTRACEFD=3Run to locate sensitive lines to bracket:
🏁 Script executed:
#!/bin/bash
# Find likely sensitive usages to guard when tracing is on
rg -nP --type=sh -C1 '(--token=|INPUT_GITHUB-TOKEN|INPUT_TOKEN|github_token)' entry.shLength of output: 1525
Localize tracing around sensitive token usage
Wrap any ${JUDGES} invocation that passes --token=${INPUT_TOKEN} (and similar --option=github_token…) with tracing off/on using a TRACE_ON flag:
• Top of entry.sh (replace lines 9–12):
- if [ "${INPUT_VERBOSE}" == 'true' ]; then
- set -x
- fi
+ TRACE_ON=false
+ if [ "${INPUT_VERBOSE}" = 'true' ]; then
+ set -x
+ TRACE_ON=true
+ fi• For each sensitive call at lines 154–156, 165–167, 208–210, 233–235, wrap as:
if [ "$TRACE_ON" = 'true' ]; then set +x; fi
${JUDGES} "${gopts[@]}" … "--token=${INPUT_TOKEN}" …
if [ "$TRACE_ON" = 'true' ]; then set -x; fi(Optional) route xtrace to a dedicated FD for cleaner logs:
exec 3>&2
export BASH_XTRACEFD=3🤖 Prompt for AI Agents
In entry.sh around lines 9–12, and for the sensitive invocations at lines
154–156, 165–167, 208–210, 233–235: introduce a TRACE_ON flag and localize
xtrace around any command that passes secrets (e.g. --token=${INPUT_TOKEN}) by
turning tracing off before the call (set +x) and restoring it after (set -x)
only if TRACE_ON is true; replace the current global set -x handling at lines
9–12 with initialization of TRACE_ON based on INPUT_VERBOSE and (optionally) set
up a dedicated xtrace FD (exec 3>&2 and export BASH_XTRACEFD=3) if you want
cleaner logs, then wrap each sensitive ${JUDGES} invocation with the conditional
set +x / set -x guards so tokens never appear in traces.
| @@ -0,0 +1,5 @@ | |||
| # SPDX-FileCopyrightText: Copyright (c) 2024-2025 Zerocracy | |||
There was a problem hiding this comment.
I think this file should be added to Reuse.toml instead. It's odd to see copyright notice in configs
4 participants
added an example of pure bash coverage
Summary by CodeRabbit
New Features
Chores
Notes