Now the number of runs each can be set, and the failures and successes are
summarized in the same way as the core Properties.hs. There is some duplicated
code which could be avoided by modifying Properties.hs.
Daniel's change which broke -Wall (adding an import for haddock only) was
somehow removed. Instead we can just modify the sample code to add the import.
Based on Adam Vogts recommendation on the mailing list. I had to give explicit
type signatures to get rid of warnings, but nearly verbatim to his version.
LayoutBuilderP is similar to LayoutBuilder (and is based on it), but LayoutBuilderP places windows matching given X.U.WindowProperties.Property (or any other predicate) into one rectangle, instead of fixed number of windows.
Somewhere between my creating the original version of this script and
adding it to the tree, a backslash got lost. It appears to have been
lost in the version I put on the wiki, so I suspect a copy-paste
problem at that point.
Spotted by arlinius in #xmonad, and this only shows up for xinerama setups.
Using an algorithm that scales better with number of workspaces would probably
be shorter too (visiting them in turn, rather than doing random access), but
probably not worth the effort.
See <http://code.google.com/p/xmonad/issues/detail?id=441>
> I have run into programs that fail when run by safeSpawn but succeed with spawn.
> I tracked it down in one (python) and it seems to be due to uninstallSignalHandlers.
> When run by safeSpawn, the program reports errors from wait.
dylan did not supply a patch and his version doesn't match the declared type signature;
since I don't want to break every `safeSpawn` user, I tossed a `>> return ()` in to make
the type right, although I'm troubled at removing the exception functions.
It is possible that this atom should be defined in the X11 library, but fix the
build of contrib for now. In any case, this would have to wait for a change and
release of the X11 binding.
rolling back:
Wed Jan 5 22:38:39 EST 2011 Adam Vogt <vogt.adam@gmail.com>
* Remove accidental atom_WM_TAKE_FOCUS from H.ICCCMFocus
The XMonad module exports this already
M ./XMonad/Hooks/ICCCMFocus.hs -7 +1
This patch adds support for calling countScreens in arbitrary places. Prior to
this patch one would end up with an open display for each call of the
countScreens function with would eventually mess up X. This patch ensures that
the display that is no longer needed is closed after the operation and thus
using the function without side effects.
A couple removed constructors need to be replaced by the lowercase versions
(ex. State = StateT Identity now). But it isn't clear that mtl-1 should be
dropped.
Now TrackFloating remembers focus for the given layout when the other window is
also tiled, but not fed to the given layout: this helps with X.L.IM, among
others.
The X.L.Maximize layout modifier does not track whether the window it stores as
maximized does still exist. The X server reuses window IDs. As a result, I
was able to reproduce the following behaviour (e.g., by opening and closing
xpdf windows): Create a window, maximize it, close it without restoring it to
its normal state. Open a new window with the same window ID (e.g., an xpdf
window after just closing an xpdf window). The new window will open maximized,
which is not what one would expect. This patch addresses this problem,
removing the ID of the maximized window from the layout when the maximized
window is closed.
Addresses the comment that:
If the focus goes from the floating layer to tiling by deleting a floating
window, it's again the master window that gets focus, not the remembered
window.
