There are really multiple things here in the close_all_docviews
function:
1. we reset the Node's tab_offset
2. we ensure the core's active_view is properly set
3. we close LogViews as well as DocViews
Some conditions seems to never happen but we stay safe and try
to cover all possible cases.
Add objc_args in meson when compiling to pass C defines also to
bundle_open.m.
Default "bundle" option to false to have by default a unix-like
build and install.
In the run-local script always expect that "bundle" option is to
false to have a unix-like install.
In the build-package script pass the -Dbundle=true option when
building on macos.
When setting the resouce path revert to original method using
[[NSBundle mainBundle] resourcePath]
to have the real resource path when the bundle option will be
activated. With the recent commit the function
set_macos_bundle_resources will be called only if the "bundle"
option is activate and is not used in unix-like mode.
since 5 months ago (ttps://github.com/libsdl-org/SDL/commit/c289bad9007cb672c994f726d967f6e5682f200d)
SDL2 now reads Xft.dpi. There is no need to link to X11 anymore.
Instead of having a separate start.lua.in file in the scripts directory
and no start.lua file in data/core we use the file data/core/start.lua
as a template for Meson to generate the final start.lua file for release.
In this way people naturally trying to run lite-xl from the source folder
will have a start.lua file albeit without a resolved version number.
Otherwise, when using run-local script or the meson install command the
meson-generated start.lua file will be used as it should be.
- Added version and license metadata
- Configuration data to be used in configured files to set metadata
- Portable binary and directories in the main install directory
- Binary file installed in correct places for all supported platforms
- Freedesktop AppStream support
- Added missing files install rules
In the function Node:get_divider_overlapping_point() we check if we
hit a divider (separator between two nodes). If yes the event is
intercepted and used to set the cursor and drag the separator if
appropriate.
In reality, on mouse move events, when one of the node is a split
and one of its child is not resizable we don't set the cursor to
and we don't intercept the event. However on a mouse pressed event
the event was intercepted regardless of the fact that the child
nodes are resizable or not. This latter behavior was unwanted as it
prevents mouse clicks to be processed because of a divided that is
inactive.
In addition it prevented processing of mouse clicks when the child
node was invisible leading to issue #363. For this latter the issue
was the invisible NagView in the upper part of the window.
To fix the problem we provide a divider with
Node:get_divider_overlapping_point() only if its child node are
resizable. In this way the mouse clicks or movements are intercepted
only if the divider is actually active.