…k other themes.
…. Will not break other themes
…reak other themes
…t break other themes
…ue (to mimic old behaviour). If set to false, then all e-mail will come from the site address defined in the Mail From field (or automatically calculated if not set). I also split the code that generates the site e-mail address into a getSiteEmailAddress() function, so that we encapsulate the logic to generate it better - it can now be called from elsewhere, if needed. This will involve a change to the DB data (to add the new option to the config_options table, and to modify the ordering of a subsequent item). I have added the necessary changes to initial_data.sql, but I don't know how to manage updates to existing systems. Here is the code that will need to be run, but I have no idea where to put it, I'm afraid! INSERT INTO `<?php echo $table_prefix ?>config_options` (`category_name`, `name`, `value`, `config_handler_class`, `is_system`, `option_order`, `dev_comment`) VALUES ('mailing', 'mail_expose_user_emails', '1', 'BoolConfigHandler', 0, 12, 'Enable to expose user e-mails in the from/reply-to field, or disable to always use the site address, below.'); UPDATE `<?php echo $table_prefix ?>config_options` SET `option_order` = 13 WHERE `category_name` = 'mailing' AND `name` = 'mail_from';
…is automatically created/updated by the software, and so should not be under version control (it always gets marked as changed, which is incorrect and confusing, and may end up causing the changes to be committed to the repo accidentally). We also don't want any of the files that the application generates in this directory (the uploads themselves) to be accidentally committed to the repository, so I've added a .gitignore file to ignore everything except itself. Now that this is present, we no longer need empty.txt, which was originally added to make sure that Git created the directory (as there is no way to create truly empty directories in Git). I've added a check to FilesController::repair() so that it works correctly when the file is not there, however the code which is used for all normal usage is nicely encapsulated in the FileRepository_Backend_FileSystem class, which already handles things correctly if the file is missing, so that was the only place that needed updating. The only other reference to attributes.php is in tools/db2fs.php, but this is referring to a local version of the file in that directory, and so is unaffected by this change.
… in the config settings, then no URL is generated at all. I believe this was intended behaviour (as there is one location where the code checks whether the URL is blank and only adds an option if it isn't) but in all cases the config value had a query string appended, and therefore it would never be blank. As well as ensuring that the contact's getReserveParkingSpaceUrl() function works properly, I have updated the code on the list_contacts view so that it only outputs the 'reserve parking space' option when a URL is available. This was the only other place where the function is currently used.
…tus, and auto-assign upon creation instead of having to subsequently edit the user.
If there is an error during uploading of a file revision, display it to the page
…quotes in it.
…eed to be there: * /cache/autoloader.php * /cache/log.php * /cache/trace.txt These files ship as empty place-holder files, but will be written to by the application in certain situations. This causes problems with revision control, as the files show up as 'modified' and may end up being committed to the source repository. This is not just a problem in the main ProjectPier repository, but it has down-stream implications too. The code already creates these files automatically if they don't already exist, but there are a couple of places where errors may be raised if they are not present. This commit contains the necessary modifications to remove these errors: * When including the autoloader.php, we suppress errors using @ prior to the fopen() statement. * The __production_error_handler() was logging all errors (except E_STRICT), even if they were deliberately suppressed using @. This patch updates the function so that it ignores suppressed errors. I've also added a comment to explain what 2048 means, and why the E_STRICT constant isn't being used. This has been logged as a bug at http://www.projectpier.org/node/3245.