Commit 51e94cf
fix(release): tolerate executable-bit flips in tag-script status checks
The auto-tag workflow failed on the first run after flowglad-main got the
fix for the empty-ArrayBuffer dbBuffer regression:
Error: Validation changed files. Commit the build output first, then rerun:
M packages/just-bash/dist/bin/just-bash.js
M packages/just-bash/dist/bin/shell/shell.js
esbuild emits dist/bin/*.js at mode 100644 on Linux, but those files were
originally committed at mode 100755 and the create-flowglad-package-tag.mjs
status checks could not tell the difference between an executable-bit flip
(spurious) and a real content change (a legitimately uncommitted dist).
Two-part fix:
- Pass `-c core.fileMode=false` to the pre-validation and post-validation
`git status --porcelain` calls so the script ignores executable-bit
diffs between index and working tree. Real content drift still aborts
the run.
- Chmod the committed `dist/bin/just-bash.js` and
`dist/bin/shell/shell.js` index entries from 100755 to 100644, matching
what esbuild produces. The blob hashes are unchanged. npm/Bun set the
executable bit at install time via the package.json `bin` field, so
consumers are unaffected.
After this, an auto-tag run on a flowglad-main HEAD whose dist matches a
clean rebuild will pass; an auto-tag run on a HEAD with stale dist will
still abort with the existing error message.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>1 parent 75353b8 commit 51e94cf
3 files changed
Lines changed: 15 additions & 4 deletions
File mode changed.
File mode changed.
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
91 | 91 | | |
92 | 92 | | |
93 | 93 | | |
94 | | - | |
| 94 | + | |
| 95 | + | |
| 96 | + | |
| 97 | + | |
| 98 | + | |
| 99 | + | |
| 100 | + | |
| 101 | + | |
| 102 | + | |
| 103 | + | |
95 | 104 | | |
96 | 105 | | |
97 | 106 | | |
| |||
143 | 152 | | |
144 | 153 | | |
145 | 154 | | |
146 | | - | |
147 | | - | |
148 | | - | |
| 155 | + | |
| 156 | + | |
| 157 | + | |
| 158 | + | |
| 159 | + | |
149 | 160 | | |
150 | 161 | | |
151 | 162 | | |
| |||
0 commit comments