On Tue, Mar 11, 2003 at 01:00:49PM +0100, Grzegorz Nieweglowski wrote: > Dominik Vogt wrote: > > Should be fixed now. Please report back if you notice any > > unexpected behaviour. > > Sure. Well, the buttons work fine now, thank you. > > But some new bugs were introduced. Try doing something like > > Mouse 1 T N Move > > And then dragging the window (by the titlebar). > > Well, normally you don't see it. The 'default' Mouse1<->Titlebar binding > works as expected. But if you redefine it, it goes crazy. > > It's not much of a trouble for most people, but I am using a 5-button mouse > (plus wheel), and have been happily using those two extra side buttons to > move/resize windows.
> > <snippet> > > Mouse 6 WTI A Move > Mouse 7 W A Resize > > </snippet> > > But now it won't work. In the past I could drag the window by simply > pressing and holding Mouse6 anywhere in the window (and moving the mouse, > of course). And that was very nice. Now the "move" function kicks in when > I release the button. And then I have to move the mouse, and press mouse > button _1_ to acknowledge the whole transaction. Which renders the whole > function very ugly. > > It seems there's no simple solution to this. Context-dependency seems a nice > way out, but I don't know if that's that easy as it seems. In titlebar > buttons the functions should (i guess) get executed on a ButtonRelease. In > the "main" titlebar area (titlebar, but not a "normal" button) it should be > either ButtonPress (e.g. "Move"), or ButtonRelease ("Raise/Lower", > especially because the titlebar can be drawn "pressed-in", just like > a button) The only exception seem to be title buttons. In all other cases you want a certain set of actions to be executed immediately. If I did not have to think about backward compatibility, I'd say the current code does the right thing. After all, you can achieve the dragging effect with AddToFunc titlebar_move + M Move Mouse 1 T N Function titlebar_move But of course we have to think about old configurations too. > So it already shows that context-dependency won't do the trick. So _maybe_ > the best thing to do would be to revert the changes, and accept the idea > that some functions work on ButtonPress, and others on ButtonRelease. And > maybe we should simply start thinking which functions should be triggered on > "press", and which on "release". It isn't perfect, but it seemed to work. > Mostly. And it would be a good temporary solution. You forget that the functions working on ButtonPress fail to work without a context window. You can not pick a target window when invoking them from a non-window context. > The *ideal* solution would be to implement correct bindings, like it was in > the past with "Key -FOO". Key release bindings are too unreliable. They *will* screw up the keyboard focus in no time. > Maybe new commands, like KeyPress/KeyRelease, and > MousePress/MouseRelease, so that the user could decide how it should work. > That would a very flexible solution, but I guess also the most difficult to > implement. As I see it's no longer possible to assign a action to > KeyRelease, so suggesting to implement it on an even wider basis (including > Mouse and changing some of the Functions engine) is, well, not so good. This functionality is already available through complex functions. > But I still think that a good KeyPress/KeyRelease, MousePress/MouseRelease > would be the solution to many problems. Sawfish is a nice example - in that > WM the user has access to a set of functions. The user binds events to > contexts not using simple Mouse/Key functions, but a series of commands, > including some that recognize double/triple clicks&releases. Sawfish however > doesn't recognize KeyRelease... It is extremely difficult to code right. I'd be surprised to see a (fairly complex) window manager that can handle key release bindings right. > Alright, so 'Key -FOO' didn't work out and got removed from fvwm. Would doing > a MousePress/MouseRelease variations of the "Mouse" command be just as > difficult? Because such commands could be a nice, clean and flexible solution > to those issues. Keyboard handling could use those features as well, but > I guess that sane mouse handling is more urgent. Complex functions. See above. > I don't know the internals of fvwm, but it would probably be a major change. > It would be a Good Thing, though. > > This mail is more a feature request than a bug report, and I understand that > it has very low priority right now. Quite contrarily, it has the highest possible priority for me. These are the first problems that will surface once a release is made. It is better to catch them before. Bye Dominik ^_^ ^_^ -- Visit the official FVWM web page at <URL:http://www.fvwm.org/>. To unsubscribe from the list, send "unsubscribe fvwm-workers" in the body of a message to [EMAIL PROTECTED] To report problems, send mail to [EMAIL PROTECTED]