All messages and errors were appended to $msg
and $errormsg, respectively, creating one long
unformatted string - mostly unreadable.
So we make $msg and $errormsg arrays instead
and iterate over those arrays in the template,
displaying each message separately.
Fixes#525.
Signed-off-by: Thomas Hochstein <thh@inter.net>
Initial motivation for this rework was to add support for the responsive thumbnnails (#474). But it also is a re-implementation instead of an enahncement of the existing code because the moveMediaDirectory function had grown into a mess. It was very hard to debug possible renaming bugs, like https://board.s9y.org/viewtopic.php?f=3&t=21185. This approach uses several small functions instead that can be combined and re-used.
See #488 for the discussion.
Merge `dashboardLimit` and `dashboardDraftLimit`
to `dashboardEntriesLimit`
Signed-off-by: Thomas Hochstein <thh@inter.net>
The dashboard will show up to $dashboardLimit future
entries; if the number of future entries is still
less than $dashboardDraftLimit, it will add drafts
up to $dashboardDraftLimit.
If there are _no_ future entries, $entries is no
array, but has a value of "1"; count($entries)
will then be one, too, so one draft less will
be shown.
Closes#465.
Signed-off-by: Thomas Hochstein <thh@inter.net>
Iconfont icons are of no value to screenreader users; in our case,
they get alternative text. By adding 'aria-hidden="true"' to the
<span> holding the iconfont icon, we avoid the screenreader trying
to announce the iconfont icon.
If the current theme was not the same as the backend theme, but the
backend theme was part of a recommended theme, then that backend theme
would get accidentally removed from the list of themes and thus nto set
as the current backend theme, breaking the themes menu
Removes some restrictions previously added. Or else we would need a 'userComments' privilege. But that would not make too much sense, since we already work with the lowest group privilege here.
This still is not ready yet, I presume. It needs more finetuning like answer to comments, etc.
Please help testing.
References #385
We still have the issue that we have set authorID 0 as the standard authorid in ML. This prevents us being more strict than this.
We will have to re-think this, maybe...
References #385
