perl: complete reproducibility — pin all host/time-derived Config fields - #291

packages/perl/build.sh (1)

30-30: 🩺 Stability & Availability | 🔵 Trivial | ⚡ Quick win

Sanitize the nodename by field position rather than regex substitution.

Using $(uname -n) as a sed pattern is fragile: if the nodename is empty the expression degrades to s//builder/g (sed errors on "no previous regular expression", failing the build under set -e), and any regex metacharacter or a / in the nodename would misfire or break the delimiter. Since the nodename is reliably the second whitespace-delimited field of uname -a (the kernel-version field may contain spaces but the nodename never does), replacing field 2 with awk is robust and avoids the regex pitfalls entirely.

♻️ Proposed robust sanitization

-MYUNAME="$(uname -a | sed "s/$(uname -n)/builder/g" | tr '[:upper:]' '[:lower:]' | sed 's#/##g')"
+MYUNAME="$(uname -a | awk '{$2="builder"; print}' | tr '[:upper:]' '[:lower:]' | tr -d '/')"

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@packages/perl/build.sh` at line 30, The MYUNAME assignment uses sed with a
dynamic nodename pattern from uname -n, which is fragile and can fail if the
nodename is empty (causing sed to error on an invalid regex) or contains regex
metacharacters or forward slashes. Instead of using sed to replace the nodename
pattern, use awk to directly replace the second whitespace-delimited field of
the uname -a output with "builder", since the nodename is reliably the second
field and this avoids regex pitfalls entirely. This makes the sanitization
robust and eliminates the dependency on the potentially problematic nodename
value being used as a sed pattern.

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes