the mail menu is a client action, so not subject to the same template
as regular views. so, it had to be patched to properly scroll like
other views (meaning: the 'header', left menu and top menu are fixed)
Also, changes the width of the searchbbar into a min-width css property,
to allow it to grow more. The reason is that we don't want the searchbox
to take two lines (or more), which will force the view_manager_header to be
higher, and which will cause slight issues with the scrollbar (absolutely
positioned)
(1) it seems that position:relative behaviour is undefined in a td,
so it has been moved to the div below oe_application
(2) the left menu bar can now scroll vertically if necessary.
when a view_manager is contained inside another, the
css rule for defining a scrollable zone (with position:absolute)
was also applied to the inside view manager.
correct a few problems introduced by the change to the scrolling area.
Note: body is in overflow:auto to allow scrolling when there is no view
manager (for example, client action)
Now, only the main view is supposed to scroll (the part below the
header with the view selector). The left menu, top bar and header
are supposed to stay fixed. This is done by carefully using
position:relative and absolute. However, this is likely to
introduce some (many?) issues, perhaps with tooltips position and
such. The bug hunt is on!
add a div parent to oe_view_manager_body, and move the searchview
drawer inside oe_view_manager_body. This is in preparation to the
goal of making the oe_view_manager_body div scrollable
to accomodate the fact that the searchview has been changed.
Two changes need to be accounted:
(1) the drawer html structure has changed (oe_selected class is
replaced by badge, some li tags are replaced by spans, ...) and
(2) the CustomFilters has been split into CustomReports and
SaveFilter, so there is an extra filter in the drawer inputs.
Also, the searchview need a parent to react to the switch_mode event,
so the tests add a fake parent to the view in make_search_view
after some thought, the extra method insert was not really necessary.
It was at first a way to force the user to give two nodes to determine
where the searchview and the drawer were to be inserted. However,
the second node can be omitted, a perfectly reasonable default is
to append the drawer just after the searchview.
Then, the role of insert is exactly the same as appendTo, so it makes
sense to override appendTo and remove insert. Simpler, and the job is done.
this commit splits the widget CustomFilters in two widgets:
CustomReports and SaveFilter. CustomReports is the widget
displaying the list of reports, and SaveFilter obviously is the
Save Current Filter widget.
The goal was to put the custom reports in the first column and the
Save Current Filter form in the second.
Before, the searchview and the searchview drawer had to be instantiated by
the user and appended separately to their correct place. Now, the
searchview is responsible for creating/destroying the drawer, and the
user is responsible for correctly using the insert method with
the node where the searchview/drawer are to be inserted.
the searchview has been splitted in two widgets: SearchView and
SearchViewDrawer. They are dependent of each other, so a drawer
has to be created for each searchview. This was not the case in
some form views (inline, created by a wizard).
the drawer wasn't closed when switching views, which might be a
problem when the user switches to a form view. Also, renames the
variable 'searchview_drawer' to 'drawer'.
incomplete work. So far, the search view widget has been split in
two widgets: SearchView and SearchViewDrawer. The SearchViewDrawer
has been inserted inline, and some preliminary work has been done
to improve the layout.
Server-side, view extension is done via xpath. This includes "template" views
full of HTML.
HTML elements often have a bunch of classes, sometimes even semantic
(!). XPath is generally great, but specifically lousy at dealing with
space-separated values: in standard XPath 1.0 to know if an element has a
class 'foo' the predicate is:
contains(concat(' ', normalize-space(@class), ' '), ' foo ')
and this has to be fully duplicated if there's a second class involved.
Things are slightly better with EXSLT/XPath 2.0 and tokenize, but still not
great:
tokenize(@class, '\s+') = 'foo'
and the equality check is very weird when unaware of XPath's evaluation rules.
``hasclass`` makes this much simpler to deal with: to get any ``foo`` node
with the class ``bar`` is as simple as:
//foo[hasclass('bar')
and it can take multiple class, as with e.g. jquery it will return elements
with all specified classes.
Beware though, the predicate function will be called once for each element to
check, since it's implemented in pure python and not profiled elements should
be filtered as much as possible before this point.
Géry Debongnie