diff options
| author | Lester Caine <lester@lsces.co.uk> | 2026-06-06 23:35:20 +0100 |
|---|---|---|
| committer | Lester Caine <lester@lsces.co.uk> | 2026-06-06 23:35:20 +0100 |
| commit | 8a2962f8d9cc5a2160e82d06f2978002003e9694 (patch) | |
| tree | 475eb61d7447d8c570af55cd41622872633c070a /templates/top_bar.tpl | |
| parent | 88a8fbd1ffca15e0ee6a83fc0cbbf8c7f19464ce (diff) | |
| download | kernel-8a2962f8d9cc5a2160e82d06f2978002003e9694.tar.gz kernel-8a2962f8d9cc5a2160e82d06f2978002003e9694.tar.bz2 kernel-8a2962f8d9cc5a2160e82d06f2978002003e9694.zip | |
kernel: load mConfig before checkEnvironment writes to it
checkEnvironment() called setConfig() to record site_server_type before
mConfig had been loaded from the database. This left mConfig non-empty
with only the server-type entry, so the subsequent empty() guard in
loadConfig() short-circuited and site_title was never loaded.
Result: getSiteCookieName() fell back to 'bit-user-bitweaver' on every
request where checkEnvironment() ran first, creating a second spurious
cookie alongside the real 'bit-user-<site>' cookie and splitting sessions.
Fix: call loadConfig() at the top of checkEnvironment() if mConfig is
not yet populated, matching the pattern used elsewhere. The bit-user-bitweaver
fallback is now only reached during a genuine cold install.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Diffstat (limited to 'templates/top_bar.tpl')
0 files changed, 0 insertions, 0 deletions