Bill.Chiles@WB1.CS.CMU.EDU (09/20/88)
First, I must say that I fully understand the policy of fighting with window managers and trying to play window manager yourself (you being a client). Okay, now. I have an application that now and then (at the user's request) wants to split one of its windows into two, shrinking the original to consume half its previous space. The new window occupies the newly vacant screen area. When one of these two halves is deleted (again by the user), I want to grow the other window to take up the space. Growing the top one works fine, but when you grow the botton one, you must also move it up to where the top one was. Under twm, this seems impossible. It does not honor basic reconfiguration with new positions, and it ignores the "Inter-Client Communication Conventions Manual" (that is, setting the normal hints as well as sending repositioning requests). I understand that it is legal for the window manager to ignore this activity, but I find it plainly offensive that it fights so hard when other reasonable window managers are willing to comply. Has anyone modified twm's code to be like awm's code in this respect? Am I correct in assuming that twm is purposefully ignoring repositioning requests? Or is this a twm bug? Is there a fix for it? I would very much like to get a patch for this. I do not hack C code, so I cannot do this for myself with the complexity of the twm implementation. Our local X maintainers are not interested in patching twm from awm's sources until the "Inter-Client Communication Conventions Manual" is law, and it is shown that twm is really at fault. Basically, they don't care about doing what's reasonable unless they have to, and there is no reason that I can tell why twm does not support repositioning. This must be an oversight or bug. If I could supply them with a fix, then they would install it. Can anyone help? Are there any comments on this other "Your client should not ..."? Bill
murthy@rati.cs.cornell.edu (Chet Murthy) (09/20/88)
In article <5916.590696332@CHILES.SLISP.CS.CMU.EDU> Bill.Chiles@WB1.CS.CMU.EDU writes: >Under twm, this seems impossible. It does not honor basic reconfiguration >with new positions, and it ignores the "Inter-Client Communication >Conventions Manual" (that is, setting the normal hints as well as sending >repositioning requests). I understand that it is legal for the window >manager to ignore this activity, but I find it plainly offensive that it >fights so hard when other reasonable window managers are willing to comply. > >Has anyone modified twm's code to be like awm's code in this respect? Am I >correct in assuming that twm is purposefully ignoring repositioning >requests? Or is this a twm bug? Is there a fix for it? > Yep. I have the same problem with twm. my application runs fine under uwm, and puts up windows in the same place, time after time. Under twm, it does an xlib:set-standard-window-properties call (in CLX) with the same valies of x,y,width,height, each time, but the window is put in different places in the screen each and every time. It's really a pain. Sometimes, the window is created off the screen, even though I set DontMoveOff. Under awm, my program doesn't seem to be getting the exposure events it needs to know when to redraw, so I can't run it there. I would like fices to twm to stop it from positioning windows in such weird places, or to make it position windows completely on-screen if possible when DontMoveOff is set. I'm gonna try to fix it myself, but the code is sort of large, so it may take a while. --chet-- murthy@svax.cs.cornell.edu
toml@hpfcdq.HP.COM (Tom LaStrange) (09/20/88)
> Under twm, this seems impossible. It does not honor basic reconfiguration > with new positions, and it ignores the "Inter-Client Communication > Conventions Manual" (that is, setting the normal hints as well as sending > repositioning requests). I understand that it is legal for the window > manager to ignore this activity, but I find it plainly offensive that it > fights so hard when other reasonable window managers are willing to comply. > Has anyone modified twm's code to be like awm's code in this respect? Am I > correct in assuming that twm is purposefully ignoring repositioning > requests? Or is this a twm bug? Is there a fix for it? This was most definitely a twm problem. I have fixed this and windows can now reposition/resize themselves with no problems. I have sent a new copy of twm to comp.sources.x but Mike is having problems getting it to run. This new copy of twm has also been submited to MIT for R3. -- Tom LaStrange Hewlett Packard Co. ARPA: toml@hpfctel.hp.com Graphics Technology Division UUCP: hplabs!hpfcla!hpfctel!toml Fort Collins, CO
bw@hpcvlx.HP.COM (Bill Wilhelmi) (09/24/88)
As a work-around, you might consider putting all these windows within one window and managing them yourself. Since you seem to be using a fixed amount of space, this shouldn't be hard to do. TWM isn't going to care about how you reconfigure the children of top-level windows. Bill Wilhelmi