dheller@cory.Berkeley.EDU (Dan Heller) (08/11/88)
In article <1701@hoqax.UUCP> twb@hoqax.UUCP (T.W. Beattie) writes: >PLEASE PLEASE at least a brief indication of what the posting does, why >anyone would want it, etc. Good idea -- and how about at least 5 postings a day instead of 3? :-) >So what is an Ardent Window Manager(X11)?? The newsgroup, of course, is comp.windows.x, so the software posted it going to be X-windows related ...oops, excuse me..."X Window System" At any rate, a window manager is something that manages the windows on your screen (or do you need to know more than that). More specifically, however... AWM is something written by Jordan Hubbard formerly of Ardent Computer. He hacked up a "version" of uwm (I don't know which version) to allow title bars on windows and so forth... For the most part, it's pretty nice and functional for the common user. However, as a programmer, I'm finding it rather tough, but I don't know how bad it is because I haven't used any other window manager other than uwm (and I need title bars too much to bother with that too long). The problems I'm having with awm (sorry, Jordan -- since you've moved away, I have to complain to _someone_) are mostly that it dies too easily at the first sign of trouble. For example, I start up X via the alias xinit =80x11+0+715 -fn r14 -C -T console -e ~/.Xinit This gives me a startup shell which uses the font "r14", places it at 0,715 on my screen, is 80 columns by 11 lines, is a console window (-C) has the name "console" and the startup file to run is the file .Xinit in my home dir (be sure to make this file executable, or you won't last too long :-). In .Xinit, I start "awm &" and run other applications (my clock, my "biff" and 2 xterms). So far, fine... I have no idea why this happens, but even tho I can go in and out of other programs in that window and so on... but the *first* ^C that I do in that window kills awm. When I start awm again, all the windows that existed on the screen do _not_ respond to the "autoraise" feature. If I start new applications, regardless of what they are, they will get autraised if I enter the window, but only the new ones... not the old ones.. ^C in that window no longer kills awm. I can't seem to turn the title bars off on windows anymore. I used to have things like "*showTitle: off" and that used to work for the programs I didn't want to have title bars, but now everything has title bars and there's no way to turn them off. Yes, I know about the change to "decorate" as documented in the latest version. however, I can't even use f.nodecorate to turn them off... Most f.* functions hardly work as they used to (e.g. I click in a window and the function should apply to that window, but literally nothing happens). If I write a program which has a simple form widget and some command wigets in it and I use my mouse button within the _form_ widget, AWM seems to think I'm on my _root_ window and tries to translate events ignoring the fact that I'm not on my root window... AWM seems to core dump a lot more often than it used to. Also, unknown events seem to kill it pretty easily. I still use awm despite my grumblings, but I've noticed that each version posted is just a little buggier than the previous release and there's no backwards compatibility with previous versions (for example, the titlebar changes). however, I'd love to get feedback on other people's experiences with awm and/or other window managers. I haven't run TWM, but I haven't seen many bug reports on it... is that cuz no one uses it, or cuz it's bugless? :-) Dan Heller <island!argv@sun.com>
moraes@godzilla.ele.toronto.EDU (Mark Moraes) (08/12/88)
In article <4961@pasteur.Berkeley.EDU> pasteur!cory.Berkeley.EDU!dheller@ames.arc.nasa.gov (Dan Heller) writes: > however, I'd love to get feedback >on other people's experiences with awm and/or other window managers. >I haven't run TWM, but I haven't seen many bug reports on it... is >that cuz no one uses it, or cuz it's bugless? :-) I've been using twm ever since version 1.0 - for the record, I think its been getting better with every release, and is still small (little over half the size of awm), and reasonably functional. There are two bugs in twm that annoy - one sounds like what you report for awm: >If I write a program which has a simple form widget and some command >wigets in it and I use my mouse button within the _form_ widget, AWM >seems to think I'm on my _root_ window and tries to translate events >ignoring the fact that I'm not on my root window... twm does the same for popup widgets - when you set override-redirect - which makes it kind of hard to use dialogs - your keystrokes go into never-never land, or the root window to be more precise. I've tried to track this down, but haven't been able to find this. As far as I can tell, twm has no idea that the override-redirect window is there, and isn't involved in any way - but it still seems to get events that have the Root Window as the window field. I'm not sure why this happens - does anyone have a fix/reason for this behaviour? The other bug is that it doesn't deal with zaphod mode (dual screen) on color Suns. It works fine on one screen, but doesn't allow anything to happen in the other screen even if you run another wm there - two awms are fine and two uwms also fine. Since I only use the mono screen anyway, it doesn't matter usually. But when I want to test something in color...... Otherwise I'm quite happy with it - doesn't dump core very much, and since version 3.2, doesn't die too often when an application dies abruptly.
jkh@pcsbst.UUCP (Jordan K. Hubbard) (08/17/88)
Having finally (more or less) settled down in Germany, I can answer some of the points raised by Dan. >I have no idea why this happens, but even tho I can go in and out >of other programs in that window and so on... but the *first* ^C that >I do in that window kills awm. This was due to system V ignorance on my part. I didn't realize that many sysV's deal with tty generated signal handling in backgrounded processes by simply ignoring the signals in question. Since awm trys to be sociable by catching these signals and exiting as gracefully as possible, it loses in these cases. I learned that there's a long standing tradition of doing.. if (signal(SIGFOO, DieNicely) == SIG_IGN) signal(SIGFOO, SIG_IGN); on system V systems and have patched awm accordinly. I'll send out a patch at the end of the week. I've also fixed a few other minor nits and one major memory leak. Since I'm going to the Xhibition, I'll send this out in a timely fashion; I.E. in a couple of days. (sigh, my feet have barely touched the ground and already I'm flying back across the pond...) I won't be sending this patch out via xpert as comp.sources.x has assumed this role. Let me know (*real soon*) whether you need it mailed direct. Dan can't turn his title bars off anymore because "showTitle" isn't the proper method anymore. I did mention this in the CHANGES file, which should *always* be read. I do stand guilty of not remaining backwardly-compatible, but I do try to warn everyone. awm also hasn't coredumped on me in quite some time. Dan, please give me more details on your problems and I'll do my best to help. Anyone else with questions, or an axe to grind, can talk to me at Xhibition. Otherwise, please send mail to the address below... Chow.. ----- Jordan Hubbard uunet!unido!pcsbst!jkh PCS Computer Systeme GmbH Pfaelzer-Wald-Str. 36 D-8000 Muenchen 90. West Germany