summaryrefslogtreecommitdiff
path: root/includes
AgeCommit message (Collapse)AuthorFilesLines
2026-06-06kernel: load mConfig before checkEnvironment writes to itLester Caine1-2/+4
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>
2026-06-06kernel: revert installer login redirect attemptsLester Caine1-6/+0
The signin/loginfrom session approach didn't work reliably due to session name inconsistency. Revert to original force_installer.tpl display. Use gOverrideLoginFunction for upgrade-time admin access. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06kernel: redirect non-admin to signin when INSTALLER_FORCE is activeLester Caine1-2/+6
Replace the inline login form (which suffered from session cookie name inconsistency) with a direct redirect to signin.php. The post-login redirect to install.php is now handled by validate.php checking the version mismatch directly, which is session-independent. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06kernel: set loginfrom to installer before force_installer.tpl displayLester Caine1-0/+2
After login via validate.php, redirect back to the installer rather than the site homepage, so the admin can proceed with the upgrade without having to navigate back manually. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06kernel: fix site_closed login bypass for nginx (SCRIPT_URL not set)Lester Caine1-1/+3
SCRIPT_URL is an Apache-only variable — nginx does not set it unless explicitly added to fastcgi_params. When site_closed is active, validate.php needs to bypass the gate so admins can log in, but the SCRIPT_URL check always failed on nginx, making login impossible while the site is closed. Fall back to PHP_SELF (set by nginx) when SCRIPT_URL is absent. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06BitSystem: replace TikiWiki placeholder with architecture docblockLester Caine1-9/+54
Covers the six concerns: package registry (registerPackage/scanPackages/ registerAppMenu), configuration (kernel_config, getConfig/storeConfig), display pipeline (display/fatalError/fatalPermission), permissions and access guards, installer schema registration (register* family / verifyInstalledPackages), and utilities (MIME, versioning, date/time, navigation). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-06BitBase: replace TikiWiki placeholder with architecture docblockLester Caine1-2/+34
Covers the four concerns: DB connection, mInfo/mErrors data hash, APCu object caching, and list utilities (prepGetList/postGetList). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-01Bump BIT_SUB_VERSION to 1 — getBitVersion() was still returning 5.0.0Lester Caine1-1/+1
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-01Bump version to 5.0.1Lester Caine1-2/+2
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-01Return early from bit_redirect() if URL is null or emptyLester Caine1-0/+4
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-24Allow null default in getConfig() for PHP 8.5 compatibilityLester Caine1-2/+2
string|null $pDefault coerced to '' internally; return type stays string to avoid breaking callers. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-24Drop finfo_close() deprecated in PHP 8.5Lester Caine1-1/+0
finfo objects are freed automatically; explicit close is a deprecation warning in 8.5. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-18Actually check tk return :)Lester Caine1-1/+1
2026-05-18Move requirements to a base of 5.0.0 since the schema structure needs a more ↵Lester Caine1-1/+4
modern base
2026-05-18Retire mochikit in favour of jquery replacementsLester Caine1-1/+0
2026-05-18Remove legacy jscalender in favour of modern replacementLester Caine1-22/+33
2026-05-15Remove unused group model codeLester Caine2-25/+0
2026-05-15Further trimming of role_model selectionLester Caine2-7/+1
2026-05-15Namespace fixesLester Caine1-3/+4
2026-05-14move phpmailer to external package managementLester Caine1-2/+2
2026-05-14move dBug into local include to manage with packageLester Caine2-1/+532
2026-05-14php-cs-fixer tidies to php8.5 standardsLester Caine20-1049/+1018
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-05-10Tidy loose ends from 6 years ago and rework to be php8.5 compliantLester Caine1-38/+37
2026-05-10(interger) -> (int) for php8.5Lester Caine1-2/+2
2026-05-10Ensure $ADODB_CACHE_DIR exists in the web services /tmp directory ( and ↵Lester Caine1-1/+1
check adodb uses it )
2026-05-10Use BIT_INSTALL to block database access until database is populatedLester Caine1-1/+1
2026-05-10Restore BIT_PHP_ERROR_REPORTING setting for error_reporting, bit_error ↵Lester Caine2-3/+4
reports are now married to this properly
2026-05-08add_link_ticket is not working in the current build of smarty so having to ↵V5-php84lsces1-1/+1
manually add tk tokens to forms as required
2026-04-19PHP8.4 style tweakslsces1-14/+14
2026-04-16checkEnvironment expanded to manage server style flags and isFeatureActive ↵lsces1-42/+35
checks for a particular feature rather than yust a y/n responce
2026-04-16PHP8.4 tidies and restore missing defaultslsces1-7/+7
2026-04-06Tidying links to bitweaver.org to avoid redirect from httplsces1-6/+6
2026-04-06doc.bitweaver.org no longer exists, reset to main websitelsces1-1/+1
2026-04-06verifyId needs to allow content_id of -1 for anonymous user trafficlsces1-1/+1
2026-04-06Drop hard coded session name in favour of site specific onelsces1-4/+4
2026-04-05bitweaver.txt is not returning a value and in any case it's out of sync with ↵lsces1-2/+5
this rebuild. Making a clean start from bitweaver5 but previous databases should just work.
2026-04-05Use namespace calls rather than require_oncelsces1-1/+1
2026-04-05If the pdo driver is being used, then db_name is part of the db_host string ↵lsces1-15/+8
and needs to be null. The problem remaining is that datadict does not get the right driver when pdo is used, so hard coded for now.
2026-04-05verifyId should only return true for valid positive content_id numbers, but ↵lsces1-1/+1
-1 is used for the anonymous user which can mess things.
2026-04-05Solving a little chicken and egg situation when installing from scratch. ↵lsces4-9/+9
Once the database is created autoload is messing things up until the tables are also created. Need a tidier way of doing this but it works for now by check tables exist before using them.
2026-04-05ADOdb needs a few little changes to properly handle pdo drivers. Currently ↵lsces1-1/+1
hard coded for pdo-firebird but should get the right driver from the Hostname
2026-04-05Renamed as BitDb to match with the autoloaded naming conventionlsces1-1132/+0
2026-03-26Namespace tidieslsces2-2/+4
2026-03-26String/null confusion change ... empty string is tidier than nulllsces1-3/+3
2026-02-02Fix duplicating _PATH defines ... cluge needed for VSCode to see the defines ↵lsces1-1/+3
so it can actually work.
2026-02-02Modernize code style.lsces1-2/+2
2025-09-11Merge pull request #1 from BitweaverCMS/masterLester Caine2-0/+231
add BitCli and BitLogger
2025-09-04Modernize code with defaults and code style. Also add missing layout tagslsces1-1/+1
2025-08-30add BitCli abd BitLoggerspiderr2-0/+231
2025-08-29Tidied the class layout and removed more adodb links, but still needs more ↵lsces1-173/+197
work to eliminate adodb-time and also pick up current php classes. Is probably obsolite really