summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2009-01-16Also test operator=.Daniel Burrows1-0/+6
2009-01-16Add a sanity-check that the copy constructor does something sensible on hints.Daniel Burrows1-0/+11
2009-01-16Fix parsing version selectors that have a ">" or ">=" condition.Daniel Burrows1-0/+2
2009-01-16Add some basic tests of resolver hints.Daniel Burrows2-0/+212
2009-01-16Implement operator!= on hints and version selectors.Daniel Burrows1-0/+2
2009-01-16Add a version selector type that only selects versions that are not equal to ↵Daniel Burrows2-0/+9
the given version.
2009-01-16Finish writing a parser for resolver hints.Daniel Burrows2-7/+151
2009-01-16Reverse the order of arguments we use when making a version comparison selector.Daniel Burrows1-2/+2
It makes more sense to me to put the operation first (as in, e.g., ">5.0").
2009-01-16Give the comparison operators better names.Daniel Burrows2-6/+6
2009-01-16Add a "inst" version selector that selects only versions that represent ↵Daniel Burrows2-0/+19
installing the package. This will be used, e.g., to describe a hint that says to reject a package entirely.
2009-01-16Add a version selector that selects all versions of a package.Daniel Burrows2-0/+17
2009-01-16Move cmdline_is_search_pattern() to aptitude::matching::is_pattern().Daniel Burrows10-25/+29
2009-01-16Create a structured data type to store version comparison in resolver hints.Daniel Burrows2-20/+295
This commit also introduces the ability to select ranges of versions, not just individual versions, and the ability to compare two hints. All untested.
2009-01-15Don't crash when the user asks for a virtual package in "why-not". (Closes: ↵Daniel Burrows1-1/+4
#511634)
2009-01-14Eliminate all uses of s?printf in favor of the safer s?sprintf. (Closes: ↵Daniel Burrows1-34/+22
#511559)
2009-01-13Initialize relevance to 0 (arbitrarily).Daniel Burrows1-1/+6
It should never have been used uninitialized, but it's easy enough to make this safe.
2009-01-13Copy verbosity in operator=.Daniel Burrows1-0/+1
2009-01-13Copy verbosity with the other fields of justification_search.Daniel Burrows1-2/+3
2009-01-11Always display the package version that's currently selected in the version ↵Daniel Burrows2-22/+65
list. This may be too low a trigger to switch versions -- maybe we should consider using double-clicks instead (but then users can't open a new tab) or adding a radio column for the currently shown version.
2009-01-11Warn the user when they're seeing a file list for a different version of the ↵Daniel Burrows1-0/+15
package.
2009-01-11Internationalize the error about not-installed packages.Daniel Burrows1-1/+1
2009-01-11Clear the list of files before displaying a new version.Daniel Burrows1-1/+3
2009-01-10Be more explicit about what's wrong when a package has a non-Debian origin.Daniel Burrows1-6/+29
2009-01-08Make the history entries model the things that change even more closely, and ↵Daniel Burrows1-35/+136
flesh them out further.
2009-01-08Tweak the description of the "archive" column.Daniel Burrows1-1/+1
2009-01-08Add feature to display the origin Archive alongside the version.Obey Arthur Liu4-2/+40
The layout might need some tweaking to accomodate the added width.
2009-01-07Redesign the history entry class to be more practical.Daniel Burrows1-61/+62
I like the idea of having a much cleaner internal model of proposed actions -- but that's not what we have and it's not clear that it should be. The history entries were describing aptitude's actions in a very high-level way, but at the cost of not exactly describing what happened in the aptitude cache. The new model is a direct representation of the extended state structure, and should work better. I also added some notes in places where we need to do more work in order to be able to track something correctly (e.g., the "auto" flag can't be compared between sessions at present -- that will be necessary in order to know when it was changed).
2009-01-07Don't misleadingly claim that candidate versions aren't written to the state ↵Daniel Burrows1-3/+0
file.
2009-01-06Merge in and push a set of patches I forgot. Sorry.Obey Arthur Liu15-234/+624
2009-01-05Flesh out the history entry code a little more and start compiling it.Daniel Burrows5-6/+126
2009-01-04Merge the backout of the reversion patch with my other work.Daniel Burrows20-699/+1342
2009-01-04Improve the note on how to do this better.Daniel Burrows1-1/+2
2009-01-04Add radio buttons to choose between showing the actions contained in a ↵Daniel Burrows3-12/+104
solution, and showing an explanation of the whole solution. The button text and tooltips still need some work IMO.
2009-01-04Implement a first cut of an explanation mode for the GUI resolver.Daniel Burrows2-2/+113
2009-01-04Extract the logic to build a tree that groups the actions of a resolver ↵Daniel Burrows2-183/+199
solution by type into its own method.
2009-01-04Backed out changeset f3e18253f9b9: reinstating changes to make Status ↵Daniel Burrows3-8/+9
autosize and to disable fixed-height mode. I mistakenly asked for this changeset to be reverted last week. This was due to some confusion on my part vis-a-vis what was causing status icons to disappear, and also due to an apparent faulty recollection of why the status columns were fixed-width in the first place. This patch doesn't seem to cause problems, so I'm sticking it back in.
2009-01-04Make set_text() on a package search entry implicitly activate the entry -- ↵Daniel Burrows2-2/+6
this means that set_limit() works as expected, and so entering a search in the dashboard works again.
2009-01-04Invoke post_process_model when the model of a package list is changed so ↵Daniel Burrows1-5/+6
that the reverse store gets rebuilt properly. This means that changes to package states show up again.
2009-01-04Add support to the Preview tab for the new pakage search entry object.Daniel Burrows2-12/+23
This incidentally fixes how the preview tab is built so that it's compatible with the new search code in PkgView, so previews have information again instead of being empty.
2009-01-04Use the new PackageSearchEntry object to improve error reporting in the ↵Daniel Burrows3-10/+43
dashboard tab.
2009-01-03Use Undo stock which underlines the U.Luca Bruno1-8/+3
* aptitude.glade: Having the same accelerator as Update Package List is not a problem as they are inside different menus and doesn't affect key bindings. People would expect the underlined U from a default GUI.
2009-01-03Uniform all menu item labels.Luca Bruno1-10/+10
* aptitude.glade: capitalize all menu item labels words
2009-01-03Add icons to "View" menu.Luca Bruno1-4/+25
* aptitude.glade: - menu_view_edit_columns: _EDIT - menu_item_find_dependency_paths: _FIND - menu_view_apt_errors: _ERROR
2009-01-03Change internal image size of menu items to ICON_SIZE_MENU.Luca Bruno1-2/+13
* aptitude.glade: - All fhe "File" menu items - Undo menu item
2008-12-31Revert two previous commits.Obey Arthur Liu3-9/+8
See <http://lists.alioth.debian.org/pipermail/aptitude-devel/2008-December/001136.html>
2008-12-29Refactor the package tab "conductor" to make it more readable and reusable.Daniel Burrows5-257/+340
Instead of a bunch of complicated functions that implicitly pass values around in slot bindings and manage memory awkwardly, we now have two "manager" objects that store a bunch of widgets and handle their signals. One deals with managing the widget to enter a search (colorizing it, showing errors, parsing the pattern, etc), and the other one embeds that object and uses it to manage a list of search results.
2008-12-26Use autosize for some columns (Status).Luca Bruno1-5/+9
* src/gtk/entityview.cc (EntityView::setup_column_properties): if size < 0 use AUTOSIZE sizing (EntityView::init): setup Status column properties with size = -1
2008-12-26Do not use fixed height rows anymore.Obey Arthur Liu2-3/+0
The performance gain doesn't seem significant enough, even for a full packages listing and fixed height prevents using autosized columns.
2008-12-26Use EditColumns column itself as Expander column.Luca Bruno3-23/+42
This will both cleanup the code and remove extra unused space in treeviews (like Dependencies). * src/gtk/info.cc (InfoTab::disp_package): removed expander references * src/gtk/entityview.cc (EntityView::EditColumnsDialog::hidden_property, set_hidden, get_hidden): added (EntityView::EditColumnsDialog::add_column_to_list): return if col has been hidden (EntityView::column_drop_handler): enforce "..." column to be the first (EntityView::init): removed Expander column, hide EditColumns in the columns editor for real (EntityView::set_model): cleaned up has_expandable_rows trick * src/gtk/entityview.h (EntityView::get_expander_column): removed (EntityView::EditColumns): column added
2008-12-26Expose the AutomaticallyInstalled object and hide the column.Luca Bruno2-1/+4
* src/gtk/info.cc (InfoTab::disp_package): hide AutomaticallyInstalled columns * src/gtk/entityview.h (EntityView::get_automatically_installed_column): added