The code was so involved there was even a comment asking for a refactor.
Now, File_redirection::where always returns a nice File_redirection
object instead of an array or string or nothing. The object is
either one which already existed or else a new, unsaved object.
Instead of duplicating "does it exist" checks everywhere, do it in
File_redirection::where. You either get what exists or something to save.
An unsaved File_redirection may be paired with an unsaved File.
You will want to save the File first (using ->saveFile()) and put the
id in File_redirection#file_id before saving.
Also made some changes in the password "munging" function call
common_munge_password to accept a profile instead of user ID (which
was only there because stoneage StatusNet used the ID to generate a
not-very-random salt, but nowadays we primarily use AuthCrypt plugin).
A feature we use of parent notices is that if you use the same @user
as the parent notice, the same @user will be notified, regardless if
there might be @user@site.com as well as @user@example.com and you're
subscribed to just one of them (or both, or none of them!).
But this threw an exception since we tested this on new notice threads.
Because we don't want to auto-fetch items from a remote server. Such
items should be delivered as attachment metadata and portrayed in the
way the local instance chooses.
Choices for portrayal are either simply nullifying this and embedding
the data, linking the file remotely requiring a manual click or maybe
use remote oEmbed data etc. to download files locally so no remote
requests have to be made.
It's too farfetched to assume any text.com in a notice is an HTTP URL.
For example stuff like pasting from log entries, with domain.com:1234
where 1234 is a _PID_ or something, not a port number for http://...
There were problems with queries that were executed but didn't seem to
be committed. Trying to patch that up by calling a ROLLBACK on transactions
where the loading of the page isn't stopped after the BEGIN statement's
intended function fails (like with the rememberme cookie in this commit).
We should get another form of URL identifier for interpreting links on notices...
It was hard editing this line in vim even, because of wide, multibyte characters...
Bring common_path() back into harmony with common_local_url(),
which started doing this 2013-03-25.
Shouldn't need to spread "StatusNet::isHTTPS()" logic all over
wherever common_path() is called; just DTRT automatically instead.
No validation has been attempted yet. Lots of changes left. This
is visibly not (very) different from the previous CSS layout. But
some simplifications have been made.
Might cause issues with local changes to themes and CSS. Also maybe
javascript which depends on certain legacy microformats elements.
The move to microformats2 is motivated by the announcement that all
microformats should be migrated to version 2, as of 2014-06-20 at:
http://microformats.org/2014/06/20/microformats-org-turns-9-upgrade-to-microformats2
Read more at http://microformats.org/
Also, tooltip text on time representation for humans has been improved.
Unfortunately no standardised representation (like "RFC850") had 4-digit years.
spl_autoload_register now calls the GNUsocial_class_autoload function
instead of us replacing the magic __autoload($cls). This means we can
queue up other autoload functions, such as the one now used for extlib
functions which exist directly in the 'extlib/' folder or have proper
namespacing (which our new Markdown class does).
At the same time we remove the "filecommand" setting, since we will
likely not have use of it thanks to PECL fileinfo.
Also the "supported" list for attachment mime types has changed
format, so we can keep track of at least some known file extensions.
Also removed the entirely unused saveGroups function.
Now avoiding multiGet and using listFind in Profile->getGroups()
so we don't have to deal with ArrayWrapper.
