1. 26 Jan, 2017 1 commit
  2. 02 Jan, 2017 1 commit
  3. 13 Apr, 2016 1 commit
  4. 08 Apr, 2016 1 commit
  5. 21 Mar, 2016 1 commit
  6. 21 Dec, 2015 1 commit
  7. 18 Dec, 2015 1 commit
  8. 27 Sep, 2015 1 commit
  9. 17 Sep, 2015 12 commits
    • totten's avatar
      CRM-16373 - Simplify UpdateConfigBackend and doSiteMove() · 4f240ac1
      totten authored
      doSiteMove() has been performing a search-and-replace against data
      structures which haven't exist since at least 4.0; it seems quite unlikely
      that this works.  Anecdotally, the workflow around this either incomplete or
      buggy/broken.  Simplifying this is simply a reflection of the reality that
      it doesn't do what it was written to do.
      
      Instead, doSiteMove() now resets to a standard baseline/default.
      4f240ac1
    • totten's avatar
      CRM-16373 - Search Settings - Fix dupes (eg contact_reference_options/autocompleteContactReference) · d434f797
      totten authored
      Note:
       * The `autocompleteContactReference` and `autocompleteContactSearch` settings don't seem to be really stored as such.
       * The form includes various hacks which effectively saves these fields in `contact_reference_options` and `contact_autocomplete_options`.
       * At runtime, the application logic checks for `contact_reference_options` and `contact_autocomplete_options`.
       * This situation predates the declarations of `autocompleteContactReference` and `autocompleteContactSearch` in `Search.setting.php`. It seems likely
         that these settings have never been useful.
      
      The fields still have some hacks, but they should be a bit easier to read/trace/grep.
      d434f797
    • totten's avatar
      CRM-16373 - Localization - Rearrange load/save logic · b72b5fc0
      totten authored
       * Declare missing setting `languageLimit`
       * Allow saving admin form -- even when using makeMultilingual, makeSinglelingual, addLanguage, and/or languageLimit
       * Move onChange logic for `lcMessages` from `ConfigSetting::add()` to `onChangeLcMessages()`
       * Move locale init from `ConfigSetting::retrieve()` to `applyLocale()`
      b72b5fc0
    • totten's avatar
    • totten's avatar
      ConfigSetting::fixParams - Remove legacy handling for countryLimit,... · 2fa3eaf6
      totten authored
      ConfigSetting::fixParams - Remove legacy handling for countryLimit, provinceLimit, defaultContactCountry
      
      This appears to be migration logic for handling a change in the format of
      select config fields.  The migration logic appears to originate before v3.0.
      Since we're only supporting upgrades from v4.0+, there's no point in keeping
      it.
      2fa3eaf6
    • totten's avatar
      CRM_Core_BAO_ConfigSetting::formatParams - Remove silly function · 6e5b5e59
      totten authored
      This function is basically just array_merge().
      
      In theory, there's a side-effect where NULL values in `$values` are unset,
      but neither use-case seems to make any use of `$values` afterward, and that
      smells like an accidental side-effect anyway.
      6e5b5e59
    • totten's avatar
      CRM-16373, CRM-14349 - Config - Disintermediate localeCustomStrings · 234d8f09
      totten authored
      localeCustomStrings has a pretty w.e.i.r.d. lifecycle (involving table `civicrm_word_replacements`
      and column `civicrm_domain.locale_cusotm_Strings`).  This commit does not unweird it, but the
      weirdness no longer involves special-case logic on `CRM_Core_Config*` -- all the weirdness is
      concentrated in `CRM_Core_BAO_WordReplacements` and `CRM_Admin_Form_WordReplacements`.
      234d8f09
    • totten's avatar
      CRM-16373 - Simplify loading of enableComponents · f4fb2d79
      totten authored
      f4fb2d79
    • totten's avatar
    • totten's avatar
      CRM-16373 - Config - Remove `enableComponentIDs` · 3d31a5c6
      totten authored
      Note: `enableComponents` is heavily used, but `enableComponentIDs` is only
      used once.  Generally, it's much easier to write tests based on names
      (`enableComponents`) than IDs (`enableComponentIDs`).
      3d31a5c6
    • totten's avatar
      checksum_timeout - Use consistent naming · 92a8de72
      totten authored
      This was a weird one:
      
       * The admin form rendered HTML element "checksumTimeout" but saved setting "checksum_timeout", which required workarounds.
       * The runtime code read the setting as "checksum_timeout".
       * The metadata defined setting "checksumTimeout". `prefetch` is enabled, but there's no `config_key` or `legacy_key`.
      
      I believe this means that the runtime and admin form worked (because they
      agreed that the name was "checksum_timeout"), but the Setting API would not
      work (because it used "checksumTimeout").
      
      This revision standardizes on "checksum_timeout" on the theory that
      "checksum_timeout" was the actual value used at runtime. This means:
      
       * If you had configured the form and had it working correctly, it should
         still work correctly.
       * If you had a script to set "checksumTimeout" via API, then it wasn't
         working before, and it's still not working now.
      92a8de72
    • totten's avatar
  10. 02 Sep, 2015 1 commit
  11. 17 Aug, 2015 2 commits
  12. 16 Aug, 2015 1 commit
  13. 13 Aug, 2015 1 commit
  14. 14 Jul, 2015 1 commit
  15. 02 May, 2015 1 commit
  16. 27 Mar, 2015 1 commit
  17. 04 Feb, 2015 1 commit
  18. 02 Feb, 2015 1 commit
  19. 20 Jan, 2015 2 commits
  20. 14 Jan, 2015 1 commit
  21. 12 Jan, 2015 1 commit
  22. 11 Jan, 2015 1 commit
  23. 09 Jan, 2015 1 commit
  24. 07 Jan, 2015 1 commit
  25. 06 Jan, 2015 1 commit
  26. 31 Dec, 2014 2 commits