GitLab

Change-Id: Ia16d7f3e59409838685d5ebee813f83e1f62468f
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

This reverts commit 8e8132fb.

That commit was not ready to be merged.  As it says in the commit
message, "if the session directory goes away somehow between updates,
there is currently _no_ way to restore it without one of these two
options. This may need to be considered before merging in this
change."

We can revert this revert when we have a good alternative for creating
the session directories.

Change-Id: I5141832da3225e202feef19dccabfea9ff35beeb

http://www.sophos.com/en-us/rss/sophos-security-news.xml

 for example
is rather poorly constructed and just has closing title tags rather
than an actual title. This patch adds some fallback so that links
are actually clickable and not lost to the UI.

Change-Id: Ida3b61f3cf60ddaa5f4b96795cd43ee443c111e4
Signed-off-by: Melissa Draper <melissa@catalyst.net.nz>

Sometimes a custom theme wants to override homeinfo.tpl with some
different content for the homepage.  There's some javascript on the
homepage (to hide the homeinfo boxes) that assumes an element with the
id 'hideinfo', but custom themes should not have to define this.

Change-Id: I1a296ae7f2ab8cc532ec443eb042f3f8e9e7d977
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

Change-Id: I87560e7a6b37de99edcdf2a9b5371d766eb95b60
Signed-off-by: Kristina D.C. Hoeppner <kristina@catalyst.net.nz>

When the site or institution has disabled public views, there's a
notice on the view access page to that effect, but the user can
happily create new secret urls with no indication that they won't
work.  This patch removes the generate new secret url button, and adds
a notice to the secret url page to inform the user when they're not
allowed to create public pages.  Also adds a sentence to the site
setting to make it clearer that disabling public pages also disables
secret urls.

Change-Id: I93a36df425bf06c4748cfdbfaf62613017518872
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

When cron finds a function with a stale (day-old) lock, just remove
the lock and give up rather than trying to execute the function again.
It's already pretty late, and waiting another minute for the next cron
to start won't make a difference.

This should avoid the case where two copies of cron both find the stale
lock at the same time, and restart simultaneously.  It won't matter if
cron_lock fails to get a $started value because the row has been
deleted: the next instance of cron to run will be able to insert the
row.

Change-Id: I025aaf89d64f47466f1ba4c5bb8178317277ec2c
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

Change-Id: I5e440374e952fbda50a79ff4141683b63b628c3c
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

This broke after changes were made to the Goals, Skills and
Interests tabs in 3282ebbf

.

Change-Id: I3f9a1db9d41945d0442c63613ae4c384d0b39d99
Signed-off-by: Francois Marier <francois@catalyst.net.nz>

Change-Id: I631d2d048154d995f84e5b57438da27a0e795afe
Signed-off-by: Francois Marier <francois@catalyst.net.nz>

* changes:
  Remove unnecessary global declaration
  Fix error messages for group access dates (bug #894592)
  Fix warnings when rendering access type after error (bug #894592)
  Display stopdate fields which have errors in them (bug #894592)
  Fix display of view access rows after error on submission (bug #894592)

Sometimes when creating a thumbnail image, getimagesize doesn't return
a value for bits per pixel on certain gifs.  This creates warnings in
the logs when approximating the required memory consumption, and also
refuses to create the thumbnail.  We can check the mimetype of the
image when 'bits' is not set, and when it's a gif, just assume the
maximum value (8 bpp).

Change-Id: Ie4a7161f8e17fa17291448084a57fa13873e0e4a
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

Change-Id: I349d2dde48e14cc59a51729cfefe4f360a70a858
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

These were using undefined strings and gave messages like "The end
date for '[[group/view]]' access cannot be in the past".

Change-Id: If087014fc608747eee8c9533e36df1fd90856456
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

When the view access page is submitted with an error, php warnings
are emitted when the rendering the access rows.

First because the pieform element assumes that an 'accesstype' index
exists in the accesslist value, but this is only true when the
accesslist still contains database records rather than the list
submitted by the form.

Second,  disabled public access record causes a warning because
'public' isn't in the list of 'presets' allowed to be added to the
view, so we need to keep track of the allowed presets and the full
list of presets that may have been allowed in the past.

Change-Id: Ic4db20738d53a52caeb3ebef0b44660dcb7c9ee6
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

When you get an 'end date in the past' error from the view access page,
the offending record is not displayed because the form removes stale
records as it pulls them out of the database.  However, when we're just
validating a bad stop date entered by the user, we need to make sure
it's still displayed to avoid confusion over the error.

Change-Id: I22c8876a599a2f00280058f32b4757d8eb523ea0
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

When you submit the view access page and there's an error in the form
submission (due to bad start/stopdates, etc.), sometimes an item is
removed from the array of access records by unsetting the array key.
Messing with the ordering of the array keys causes a javascript error
which stops the access list from being displayed.

Change-Id: Iaa686c8adecc322a6b779ee3247b28c945031854
Signed-off-by: Richard Mansfield <richard.mansfield@catalyst.net.nz>

We removed info from the email subject and added it to headers. However
people generally don't see the headers, so this patch adds the info back
as pseudoheaders, while removing duplicate info such as poster name,
topic and time.

Change-Id: I1063c70693e3b6345d41225b8fed53e0c20e7d95
Signed-off-by: Melissa Draper <melissa@catalyst.net.nz>