| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
string, add parser support for using it on trailing bare strings, and make incremental search in the cwidget UI use it. (Closes: #506651)
|
|
Originally, I thought these would be useful in the analysis that's used
to generate the Xapian search. However, I ended up not using them, and
they don't have a good definition with the prefix-search stuff that's
coming up, so I've just removed them for now.
|
|
|
|
settings, and dispatch Undo commands to the current tab.
|
|
Undo command.
|
|
This will be useful for updating the states of GUI buttons / menu items.
|
|
The design I settled on relies on the TabsManager to signal when the
currently active view's Package menu changes and on the main window
to rebuild the Package menu. Something similar will be done with Undo
when I have time to address the issues around that menu.
|
|
The main impetus here is that tab.h will need PackagesAction, but gui.h
needs tab.h.
|
|
I'm working on some code that will require having a constructed and
accessible main window available when new tabs are added (to update
the menu bar appropriately).
|
|
|
|
For some reason, GTK+ gets unhappy when we remove these in order
to repopulate the menu.
|
|
|
|
|
|
|
|
object, and store it as a PkgEntity.
The use of a bare reference was a holdover to decrease the number of
lines changed by the last patch. But since we have a proper object and
not just a sigc++ binding thunk, it's safe to use a ref_ptr to store
the encapsulated entity. Also, storing a PkgEntity instead of an Entity
eliminates a lot of unnecessary dyn_downcasting.
|
|
description area.
This makes memory management of the buttons cleaner; it will also make
it possible to add new buttons (we were at the limit of how many
arguments could be bound via libsigc++).
|
|
that in the package action menu).
|
|
installed.
|
|
We were using what's technically a "hard disk" icon, but looks like an
envelope to me; either way, I never was able to associate it with "this
is installed and OK!" The green sphere does that better for me.
|
|
Specifically, autoresolution of dependencies is deferred until after all
the package states chosen by the user have been updated. This means
that, for instance, packages that the user selects won't be marked as
autoinstalled even if another selected package requires them, and
the selected packages will be used to fulfill alternatives rather than
installing other packages.
|
|
entities at once.
We can't test this because of the lack of an "undo" operation at
present.
|
|
|
|
entity view.
|
|
|
|
|
|
build_package_menu.
|
|
|
|
|
|
get_menu() was essentially serving both purposes; fill_menu is the least
common denominator of the two.
|
|
|
|
|
|
This is a reversal from my previous commit.
I still have to figure out how to remove old entries from a menubar.
Also there are cases where the focused package has changed but it is not detected correctly.
|
|
Secondary actions should be placed in the description pane or info tab.
The first intention was to have a duplicate of the contextual popup menu in the menu bar as Synaptic or the ncurses version of Aptitude for that matter. The tabbed design of the GTK Aptitude however makes this very difficult. Determining which packages in which entityview in which tab is supposed to get in and out of focus precisely would involve adding quite a lot of code.
|
|
|
|
|
|
|
|
|
|
This is partly a replacement for the top-level trees in the old
aptitude; it seems like people find it handy to have a few simple
searches that they can quickly combine with anything else.
Eventually there should be a way for the user to interactively add, edit
and remove entries.
|
|
This has the unfortunate side effect that showing the tab again just
adds it to the visible list and doesn't make it the currently selected
tab. One thing at a time, though...
|
|
This took a lot longer than it should have, mostly due to my own
stupidity in missing the GtkNotebook::remove_page() binding in
TabsManager::TabsManager().
|
|
Maybe the "done applying changes" message should be bolded and/or in a
larger font? Or maybe centered so it doesn't show up in a different
place from the messages in the progress bar?
|
|
This straightens out some of the back-and-forth calls and replaces them
with more straightforward creation/connection/execution semantics.
|
|
I'm about to introduce a class named DpkgTerminal, and this avoids name clashes.
|
|
Some of the code is a little gross and will be refactored, but it all
seems to work. Currently the conffile prompt is shown in a tab, but it
might be better to use a dialog instead.
|
|
corrupt later messages with the first one.
|
|
the program.
|
|
implementation.
The implementation is mostly based on reverse-engineering the message
syntax from the synaptic source code, since I can't find up-to-date
documentation on it. The one hole in the parser is that it will do the
wrong thing if a conffile filename contains a single quote, but so will
synaptic and I don't know how dpkg escapes the inner quote (backslashes
maybe?).
|
|
|
|
status pipe.
|
|
|
