jim@EXPO.LCS.MIT.EDU (Jim Fulton) (11/08/88)
Okay, it has been entirely too quiet out there on XPERT on the topic of R3. I'm probably going to be sorry, but.... What do people like? What don't they like (try to be at least a little reasonable on this one)? Try to stay off little details that have been already been discussed, like performance of particular platforms. :-) Also, hackers might want to concentrate on getting the various previewers in contrib working with the new fonts and with the R3 Intrinsics. Jim
phils@ATHENA.MIT.EDU (11/08/88)
Okay, here are a few problems with R3 and things on my wish list for next (just in time for Christmas) release. 1) When logging into the console, it was previously possible to check for this by seeing if the tty was set to /dev/ttyv0. This was useful for disting- uishing whether or not one was remotely logged in or not, in order to bring up other windows, etc. Now we use 'xdm' and ttyp0 is taken on a "first come first serve basis". So how can one's login shell now know whether the logging in is from the console or not? 2) XPutImage though greatly fixed seems to leave holes on expose events but only when refreshing XYBitmaps on the QDSS display. This did not happen before, but the expose regions used to only shift horizontally. Same code and everything. We only have R3 on a Vax/GPX so I can't speak for other machines. 3) Xterm and 'vi' STILL don't get along on long lines that wrap around!? 4) My uwm widow manager is quite a bit slower, and doesn't seem to read mouse motion very well. 5) In this age of toolkits, it would be nice if "uwm" could selectively dump subwindows. (Not really an R3 gripe.) A '-subwindow' option, for example, would be nice. Otherwise, three cheers for the X Consortium!! Philip Thompson (phils@athena.mit.edu) Dept. of Architecture and Planning.
corbet@mead.uucp (Jonathan Corbet) (11/08/88)
jim@EXPO.LCS.MIT.EDU (Jim Fulton):
>What do people like? What don't they like
Well, I was about to post on this anyway...FYI, I am running X on both
Sun 3 and 4 machines, mono and color, SunOS 4.0, nothing special in Sun.macros.
Complaint #1: xman ignores the geometry argument. No matter what I put on
the command line, it still wants me to place the window myself. That makes
it inconvenient to start out of xinit. It is also impossible to start xman
in an iconic state.
#2: When I select "remove this manpage" in an xman manual window, xman
dies with:
X Protocol error: BadDrawable, invalid Pixmap or Window parameter
Major opcode of failed request: 67 (X_PolyRectangle)
Minor opcode of failed request: 0
Resource id in failed request: 0xd00027
Serial number of failed request: 2084
Current serial number in output stream: 2084
#3: I have seen xterm core dump when I hit the Sun "L" keys, but it does not
seem to be reproducible.
#4: I wish there were a window-oriented debugger interface, like dbxtool.
Maybe someday when I have some extra time I will try to implement one :-)
What do I like? I am quite fond of the system as a whole. I won't talk about
my color Suns, though...:-) I like the way xinit works now. I haven't played
with xdm, but it looks useful. Xeyes is cute. I like awm! Xcalendar is
nice, but it core dumps when the pointer enters its window on color machines.
I may try to track that down soon.
Overall, I think it is a very well designed and implemented system. Good work!
Jonathan Corbet
National Center for Atmospheric Research, Field Observing Facility
corbet@stout.ucar.edujim@EXPO.LCS.MIT.EDU (Jim Fulton) (11/08/88)
> 1) [how to distinguish xdm logins] Here is what I do with the default setup. I have my ~/.xsession #!/bin/csh -f source $HOME/.env.csh xrdb -load $HOME/.Xres xmodmap $HOME/.xmodmaprc xsetroot -solid '#888' xset b 100 400 c 50 s 1800 r on uwm & xclock -geometry 48x48-1+1 & exec /usr/bin/X11/xterm -geometry 80x66-0-0 -name hpx $* If you want to do anything special when you start up an xterm, you could set a magic environment variable and to an xterm -ls (so that your .login could look for your variable). Or any number of things. It depends on what you want to do. One of the goals for xdm was that it would be flexible enough to allow you to tailor quickly and easily without writing C code. The next step is to write a real session manager. > 2) XPutImage though greatly fixed seems to leave holes on expose events > but only when refreshing XYBitmaps on the QDSS display. XYBitmaps are still broken on some displays. As always, send small test programs to xbugs. > 3) Xterm and 'vi' STILL don't get along on long lines that wrap around!? Nobody contributed fixes for this and we didn't have time to look into it ourselves (nobody here uses vi). Attention vi-hackers: you will make new friends if you fix this and then give it back to us! > 4) My uwm widow manager is quite a bit slower, and doesn't seem to read > mouse motion very well. Definitely send a note to xbugs with your .uwmrc, any resources, and explicit instructions for how to demonstrate the situation. Uwm does have to do a lot more work these days, so expect it to be a little slower (window manager writers, time to start thinking about tracking color map changes among other things). We hadn't noticed, but then that's not very surprising. > 5) In this age of toolkits, it would be nice if "uwm" could selectively > dump subwindows. (Not really an R3 gripe.) A '-subwindow' option, for > example, would be nice. I don't understand what you mean by "dump subwindows". Do you mean xlswins? Or xpr? Or xmag? In any case, uwm is rapidly reaching the point of needing a major overhauling. Once the ICCCM is adopted, we plan to attack it with very sharp knives.... Thanks for the feedback, Jim
mtr@mace.cc.purdue.edu (Miek Rowan) (11/08/88)
In article <8811072007.AA10984@EXPO.LCS.MIT.EDU>, jim@EXPO.LCS.MIT.EDU (Jim Fulton) writes: > > What do people like? What don't they like (try to be at least a little > reasonable on this one)? I would say, as far as the core R3 dist., that you guys did a really professional job. The problems I ran into were minor. The contrib dist. is a little better than it was with R2, but that doesn't say too much :-( But remember that the X Consortium people never claimed anything about the contrib stuff. I imagine most of the complaints will be about the same though (about contrib that is). xdm is nice. I haven't used it for anything but a Sun console. There are some security questions that come to mind, but I will leave those till after I have played with it more. (Especially since people are in a "security hysteria" mode nowadays :-) Hats off. mtr
eli@haddock.ima.isc.com (Elias Israel) (11/08/88)
In article <8811080021.AA13611@EXPO.LCS.MIT.EDU> jim@EXPO.LCS.MIT.EDU (Jim Fulton) writes: >> 3) Xterm and 'vi' STILL don't get along on long lines that wrap around!? > >Nobody contributed fixes for this and we didn't have time to look into it >ourselves (nobody here uses vi). Attention vi-hackers: you will make new >friends if you fix this and then give it back to us! If this is the bug that I am thinking of, the source of the problem is actually a shortcoming in the termcap entry for xterm. I have added the auto-margin ('am') attribute to all of my termcaps and now vi and xterm seem to get along just fine. I don't know if SysV systems have this attribute or an analog, but you have to do something like this to the terminfo file for xterm to make it work there. Elias Israel | "Justice, n. A commodity which in more or Interactive Systems Corp. | less adulterated condition the State sells Boston, MA | to the citizen as a reward for his allegiance, ..!ima!haddock!eli | taxes, and personal service." | -- Ambrose Bierce, _The Devil's Dictionary_
jlevy@arisia.Xerox.COM (Jacob Levy) (11/08/88)
The X11R3 I have is running on a variety of B&W Sun 3's. Today we are about to attempt color and Sun4's, so beware an following flames.. :-) I like it. There are (as usual) some small things I don't like. #1. Not sure that this has been discussed before, but it seems unreasonable that the X server exits when it is out of memory. This happens a lot on the local configuration since people tend to run other big jobs, running out of swap space fast. Couldn't the request causing the memory exhaustion just be refused? #2. xterm comes up with VERY thick scrollbars, if I have xterm*scrollbar: yes in the .Xdefaults file. It's easy to work around this, by specifying -sb on the command line, and then somehow it comes up OK. Maybe something wrong in the way I set up the resource manager or xrdb? #3. Minor complaint: xhost (or xinit) prints out what hosts are being added to the permission table, ON THE CONSOLE. makes your screen garbled when X comes up. Also easy to work around, by redirecting the output of xinit to a file, for later viewing - this is a good idea anyway,... All in all, a lot of improvement over R2, and the right people are to be congratulated. --Jacob Levy jlevy.pa@xerox.com
toddb@tekcrl.crl (Todd Brunhoff) (11/09/88)
>> > 3) Xterm and 'vi' STILL don't get along on long lines that wrap around!? >> >> Nobody contributed fixes for this and we didn't have time to look into it >> ourselves (nobody here uses vi). Attention vi-hackers: you will make new >> friends if you fix this and then give it back to us! Here's the version from my Tek workstation that works. Don't ask my why it works. Just believe... vs|xterm|4316x|4317x|vs100|xterm terminal emulator (X window system):\ :cr=^M:do=^J:nl=^J:bl=^G:le=^H:ho=\E[H:\ :LC:co#80:li#65:cl=\E[H\E[2J:bs:am:cm=\E[%i%d;%dH:nd=\E[C:up=\E[A:\ :ce=\E[K:cd=\E[J:so=\E[7m:se=\E[m:us=\E[4m:ue=\E[m:\ :md=\E[1m:mr=\E[7m:me=\E[m:\ :ku=\EOA:kd=\EOB:kr=\EOC:kl=\EOD:kb=^H:\ :k1=\EOP:k2=\EOQ:k3=\EOR:k4=\EOS:ta=^I:pt:sf=\n:sr=\EM:\ :al=\E[L:dl=\E[M:ic=\E[@:dc=\E[P:\ :MT:ks=\E[?1h\E=:ke=\E[?1l\E>:\ :is=\E[r\E[m\E[2J\E[H\E[?7h\E[?1;3;4;6l:\ :rs=\E[r\E<\E[m\E[2J\E[H\E[?7h\E[?1;3;4;6l:xn:\ :AL=\E[%dL:DL=\E[%dM:IC=\E[%d@:DC=\E[%dP:\ :ti=\E7\E[?47h:te=\E[2J\E[?47l\E8:\ :hs:ts=\E[?E\E[?%i%dT:fs=\E[?F:es:ds=\E[?E: v2|xterms|4315x|vs100s|xterm terminal emulator (small)(X window system):\ :co#80:li#24:tc=xterm: --------------- Usenet: {ucbvax,decvax,allegra,uw-beaver,hplabs}!tektronix!crl!toddb {CS,ARPA}net: toddb%tekcrl.crl.tek.com@relay.cs.net c--Q Q US: Todd Brunhoff; Visual Systems Lab; Tektronix, Inc. ` Box 500 MS 50-662, Beaverton OR 97077 - Phone: (503) 627-1121
wicinski@nrl-cmf.UUCP (Tim Wicinski) (11/09/88)
In article <8811072007.AA10984@EXPO.LCS.MIT.EDU>, jim@EXPO.LCS.MIT.EDU (Jim Fulton) writes: > > Okay, it has been entirely too quiet out there on XPERT on the topic of R3. > > I'm probably going to be sorry, but.... > > What do people like? What don't they like (try to be at least a little >reasonable on this one)? Try to stay off little details that have been already > been discussed, like performance of particular platforms. :-) > the main problem in the building of the core is that if you want to put the fonts and rgb databse someplace else (like /usr/local/lib) you have to remember about server/include/site.h! no where in imake/imake.includes is that mentioned (maybe in server/ddx/<make>/README), but you have to track that down, and shove the right stuff in site.def. DO something about that - make ALL possible choices in imake.includes, not 95% of them. sort of defeats the purpose... That, and the xterm and vi bug (major lame-o), and the xman bugs are the most noticable/painful. xmh - adding buttons for the rest of mh stuff like burst, packf. (i'll probably add burst myself when i get chance...whenever that is) contrib distribution - here's something I'm not sure about - does the contributed distribution contain the same version of the stuff on expo in the ~contrib directory? I mean, why should i blow away ANOTHER disk when all I might want is a lousy window manager or some screen previewer. everything else i might want when i get around to getting disk space. ...just bitchin' and moanin'.... tim wicinski, naval research labs
burzio@mmlai.UUCP (Anthony Burzio) (11/09/88)
In article <8811072007.AA10984@EXPO.LCS.MIT.EDU>, jim@EXPO.LCS.MIT.EDU (Jim Fulton) writes: > What do people like? What don't they like (try to be at least a little > reasonable on this one)? Try to stay off little details that have been already > been discussed, like performance of particular platforms. :-) After installing X11R3 on a SUN, I think it would be nice if there was a decent installation/instruction manual. Sure it compiles (quite amazing) but the user is dumped at that point with a resounding "so now what?" If I didn't have X10 we'd have never got even XTERM running. Now if I can just figure out how to set default attributes... ****************************************************************************** Tony Burzio * X is catching on!!! But, what are we going to do Martin Marietta Labs * with all these VT terminals, let alone the card punch! ******************************************************************************
bob@allosaur.cis.ohio-state.edu (Bob Sutterfield) (11/09/88)
In article <957@ncar.ucar.edu> corbet@stout.UCAR.EDU (Jonathan Corbet) writes: >#4: I wish there were a window-oriented debugger interface, like dbxtool. >Maybe someday when I have some extra time I will try to implement one :-) Has anyone spent any time bringing xgdb up to X11R3? Any war stories or helpful hints? I'll try to get to it, but probably not until next week.
jim@EXPO.LCS.MIT.EDU (Jim Fulton) (11/09/88)
> #1. [server exits when it runs out of memory] We had wanted to make headway on this for R3 but ran out of time. It was one of the first things put on the R4 todo list. > #2. xterm comes up with VERY thick scrollbars Yeah, they look really ugly, don't they. I like the following (although it is a little thin): XTerm*scrollbar*thickness: 5 > #3. Minor complaint: xhost (or xinit) prints out what hosts are being added > to the permission table, ON THE CONSOLE. Sounds like your startup script isn't right. Send me (but not the list) a copy of your .xinitrc and explicit directions of how you start up X. I'll try to figure out what's up. Thanks for the comments, Jim
jim@EXPO.LCS.MIT.EDU (Jim Fulton) (11/09/88)
> site.h Hmmm, you should *never* have to touch site.h. The Imakefile should take care of passing the right definitions to globals.h. Could you send me your site.def so that I can try to reproduce it here? Also, note that you have to rebuild Makefiles after editing site.def and ought to do a full make World. > xterm vi problems See Todd's suggestion for a termcap entry to use. > xman bugs Yup. They're being worked on. > contrib Expo's ftp directories and the contrib tapes are not necessarily related. With only a few exceptions, things in expo:~ftp/contrib didn't get put onto the tape(s) unless somebody sent me a specific request to do so. People with limited disk space are definitely encouraged to list out the tapes (or tar files) and only keep the things they want on disk. Jim
ehrlich@shire.cs.psu.edu (Dan Ehrlich) (11/10/88)
In article <8811072007.AA10984@EXPO.LCS.MIT.EDU>, jim@EXPO (Jim Fulton) writes: > >Okay, it has been entirely too quiet out there on XPERT on the topic of R3. > >I'm probably going to be sorry, but.... > We have noticed that on an IBM RT 6152 Academic System running AOS 4.3 Release 2 (aka 4.3 BSD UNIX) that a number of the demos result in blank windows. So far this is true of xbiff, xeyes. When trying to run xtrek (in color) there is no cursor in the window. Beyond trying the demos I haven't ventured yet. I am still a noviciate and just getting my feet wet. -- Dan Ehrlich <ehrlich@shire.cs.psu.edu> | Disclaimer: The opinions expressed are The Pennsylvania State University | my own, and should not be attributed Department of Computer Science | to anyone else, living or dead. University Park, PA 16802 |
janssen@titan.sw.mcc.com (Bill Janssen) (11/10/88)
I am working with the Andrew toolkit in the contrib section, and have really got to applaud the ITC, IBM, and CMU, for donating such a fantastic set of tools. The Andrew "ez" editor is *almost* GNU Emacs in function, with styled text and multi-media objects. R3 finally provides a decent raster editor, and a picture editor that is suitable for slides. The Andrew console tool is one of the best ways of watching your system that I have ever seen. What's really nice is to see that all of the Andrew multi-media documents can be sent via E-mail, and posted to bboards, and so forth. There is a sketchily documented integrated mail-and-bboard scheme in the system. Extremely impressive stuff. *But*... in drawing round-cornered boxes, the thing is glacial. I don't think that it's Andrew, I think it's the server. It appears to be caused by slowness of drawing in the "mi" portion of the sample server (running on a Sun). Can't really complain, because the rest of the server seems quicker than the R2 release. Also ran across an infinite loop in the server Arc code, when a particularly bogus set of args is given to it (which will be posted later when it is more definitely identified). Bill