Update gemfile locks via sed instead of bundle lock

The rails_main and rails_main_head appraisals target ruby-head / a
future Ruby (e.g. 4.1.0.dev). Running `bundle lock` on them from the
release job's Ruby 4.0 — even with BUNDLE_IGNORE_RUBY_VERSION=1 —
produces lockfiles that CI on ruby-head then wants to rewrite, and
rejects because Bundler is running in frozen mode:

  Cannot write a changed lockfile while frozen.
  The running version of Bundler (4.0.16) does not match the version
  of the specification installed for it (4.1.0.dev).

The only thing that actually changes in a lockfile during a release is
the view_component version pinned by the path gemspec (e.g. 4.13.0 →
4.14.0). Replace the bundle lock loop with a targeted sed that rewrites
that version in every Gemfile.lock / gemfiles/*.gemfile.lock. This is
robust across mismatched Ruby versions and avoids running bundler at
all for appraisals whose Ruby isn't installed.

Verified locally: sed rewrites both the PATH-section spec line and the
bottom DEPENDENCIES pin, matching only the view_component entry.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Copilot-Session: 217f0a0d-a9f2-4607-9643-5fcb3e34f20f

Author: joelhawksley

script/release

@@ -61,16 +61,22 @@ update_ruby_version() {
 }
 
 update_gemfiles() {
-  # Update Gemfile.lock. Use `bundle lock` (not `bundle install`) so we
-  # only rewrite the lockfile with the new version — no gem installs
-  # needed, and it avoids re-triggering the bundler-cache frozen check.
-  bundle lock
-
-  # Update appraisal gemfile locks.
-  # BUNDLE_IGNORE_RUBY_VERSION is needed because the gemfiles declare ruby
-  # versions that don't match the local Ruby (e.g. 3.2, 3.3, 4.1.0.dev).
-  for gemfile in gemfiles/*.gemfile; do
-    BUNDLE_GEMFILE="$gemfile" BUNDLE_IGNORE_RUBY_VERSION=1 bundle lock
+  major=$1
+  minor=$2
+  patch=$3
+  new_version="$major.$minor.$patch"
+
+  # The only lockfile change during a release is the view_component
+  # version pinned by the path gemspec. Update it in place across every
+  # lockfile. Running `bundle lock` from a mismatched Ruby (e.g. ruby-head
+  # appraisals from Ruby 4.0) produces lockfiles CI cannot consume in
+  # frozen mode, so we edit them directly instead.
+  for lockfile in Gemfile.lock gemfiles/*.gemfile.lock; do
+    [ -f "$lockfile" ] || continue
+    sed -i \
+      -e "s/^\(    view_component \)([0-9][0-9]*\.[0-9][0-9]*\.[0-9][0-9]*)/\1($new_version)/" \
+      -e "s/^\(  view_component \)([0-9][0-9]*\.[0-9][0-9]*\.[0-9][0-9]*)/\1($new_version)/" \
+      "$lockfile"
   done
 }