bgeer@javelin.sim.es.com (Bob Geer) (05/29/91)
I have a brand-new Sun IPC after having used a 3/80 for a couple of years. On the 3/80 I used X11R3 (I think), then was upgraded to X11R4 shortly before the arrival of the IPC. So far, I'm not that impressed with openwin on the IPC, especially since the distribution of X includes, libraries, & executables files are into non-X11-typical subdirectories. I mean, I don't think I should have to monkey around with make files to compile such benign programs as xv or xroaches. Making gazillions of file links seems quite unreasonable, too. So, I'm tempted to "off" openwin & load up a "generic" X11R4 & Motif into a more typical directory layout -- /usr/bin/X11, etc. Anyone got comments on this plan -- good, bad, indifferent? Thanks in advance...Bob -- <> Bob `Bear' Geer <> bgeer%javelin.sim.es.com (this *should* work) <> <> cola-zombie <> speaking only for myself, one of my many tricks <> <> Salt Lake City, <> "We must strive to be more than we are, Lal." <> <> Ootah <> -- Cmdr. Data, learning schmaltz <>
rpburry@ncs.dnd.ca (Paul Burry) (05/30/91)
In article <1991May29.144727.5008@javelin.sim.es.com> bgeer%javelin@dsd.es.com writes: : |shortly before the arrival of the IPC. So far, I'm not that impressed |with openwin on the IPC, especially since the distribution of X |includes, libraries, & executables files are into non-X11-typical |subdirectories. I mean, I don't think I should have to monkey around |with make files to compile such benign programs as xv or xroaches. |Making gazillions of file links seems quite unreasonable, too. | |So, I'm tempted to "off" openwin & load up a "generic" X11R4 & Motif |into a more typical directory layout -- /usr/bin/X11, etc. Anyone got |comments on this plan -- good, bad, indifferent? I've also been thinking about running MIT X11R4 on my SS1+, but my motivation is somewhat different. We have a bunch of X-terminals and workstations from other vendors. When we try to display some of the Sun's Openwindows applications on these Xservers, they often complain about missing fonts etc. So, I have been thinking about tossing Openwindows and running MIT X11R4 and one of the available window managers distributed with MIT X. Now, the questions: 1. What Openwindows applications break when used with MIT X11R4? 2. What window manager should we run? 3. Will the Olwm distributed with MIT X11R4 work with my favorite Openwindows applications? Thanks, Paul -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Paul Burry Voice: (613)-991-7325 Internet: rpburry@ncs.dnd.ca Fax: (613)-991-7323 UUCP: ..!{uunet,cunews}!ncs.dnd.ca!rpburry
brossard@sic.epfl.ch (Alain Brossard EPFL-SIC/SII) (05/30/91)
In article <1991May29.230342.15836@ncs.dnd.ca>, rpburry@ncs.dnd.ca (Paul Burry) writes: |> In article <1991May29.144727.5008@javelin.sim.es.com> bgeer%javelin@dsd.es.com writes: |> : |> |shortly before the arrival of the IPC. So far, I'm not that impressed |> |with openwin on the IPC, especially since the distribution of X |> |includes, libraries, & executables files are into non-X11-typical |> |subdirectories. I mean, I don't think I should have to monkey around |> |with make files to compile such benign programs as xv or xroaches. |> |Making gazillions of file links seems quite unreasonable, too. If you install Imake and its configuration files appropriately, you should have no problems compiling programs, you will want to add the symlink /usr/include/X11 though! |> I've also been thinking about running MIT X11R4 on my SS1+, but my motivation |> is somewhat different. We have a bunch of X-terminals and workstations from |> other vendors. When we try to display some of the Sun's Openwindows |> applications on these Xservers, they often complain about missing fonts etc. The solution is not to get rid of the programs! But add the fonts to the other servers. If you have access to the mit server, you have access to the missing fonts you need. Or I think you can convert the OW 2.0 fonts to bdf format, check your manual. |> So, I have been thinking about tossing Openwindows and running MIT X11R4 and |> one of the available window managers distributed with MIT X. The only difference between OW 2.0 and X.V11R4 is the server and the shared libraries: 1- get rid of the OW 2.0 shared libraries and use the one from MIT 2- pick one server depending on your needs, if you don't have a GX accelerator, stick with the server from MIT. |> Now, the questions: |> 1. What Openwindows applications break when used with MIT X11R4? I assume you mean xview based applications, the answer is NONE, since the xview toolkit is pure X11. |> 2. What window manager should we run? Most of my users prefer olwm and now some are moving to olvwm. Those who are not afraid of a configuration file use twm or if they have the swap and memory needed tvtwm. A minority of hackers uses the versatile gwm. |> 3. Will the Olwm distributed with MIT X11R4 work with my favorite |> Openwindows applications? Much better to use the binary from OW 2.0, you will save a lot of time and grief if you don't bother compiling anything under contrib/toolkits/xview2. Just use the binaries already available under OW 2.0. -- Alain Brossard, Ecole Polytechnique Federale de Lausanne, SIC/SII, EL-Ecublens, CH-1015 Lausanne, Suisse brossard@sasun1.epfl.ch
root@fstc-chville.army.mil (Operator) (05/31/91)
rpburry@ncs.dnd.ca (Paul Burry) writes: >In article <1991May29.144727.5008@javelin.sim.es.com> bgeer%javelin@dsd.es.com writes: >|shortly before the arrival of the IPC. So far, I'm not that impressed >|with openwin on the IPC, especially since the distribution of X >|includes, libraries, & executables files are into non-X11-typical >|subdirectories. I mean, I don't think I should have to monkey around >|with make files to compile such benign programs as xv or xroaches. I agree... I don't like the /usr/openwin stuff... I want includes to be in /usr/include/X11 (not /usr/openwin/include/X11) and I want binaries in /usr/bin (not /usr/openwin/bin) etc... So I just: 1) tar /usr/openwin into a file 2) cd /usr and untar the file... 3) rm -r /usr/openwin That solves 99% of the problems... No FONTPATH, MANPATH, LD_LIBRARY_PATH, etc to mess with... Regarding the openwin fonts... I think they are great... Much better than the MIT fonts... I have had very little trouble getting the MIT-developed clients to port to Sun's OPENWIN... And I think the olvwm is super... I've used the MIT-X11 and I've used OPENWIN... I think OPENWIN on the Sun is much better supported... Brian boyter@fstc-chville.army.mil
guy@auspex.auspex.com (Guy Harris) (06/01/91)
> The only difference between OW 2.0 and X.V11R4 is the server and the >shared libraries: > 1- get rid of the OW 2.0 shared libraries and use the one from MIT After, of course, grabbing one of the sets of patches to the MIT Xlib to get it to support the Compose key, if you care about Compose key support, as the vanilla MIT XLookupString doesn't support it but the Open Windows XLookupString does.
elling@eng.auburn.edu (Richard Elling) (06/01/91)
> From: root@fstc-chville.army.mil (Operator) > > I don't like the /usr/openwin stuff... I want includes to be in /usr/include/X11 > (not /usr/openwin/include/X11) and I want binaries in /usr/bin (not /usr/openwin/bin) > etc... This may be fine for a single machine. Or even a small network. But for large networks with multiple servers it is an administration nightmare. Consider the advantages of having one mount point (/usr/openwin) rather than 4 mount points (/usr/bin/X11, /usr/lib/X11, /usr/include/X11, and /usr/man/X11). With a single mount point, there is only one filesystem to worry about mounting. This is a good thing. At Auburn, we use the automounter for mounting /vol/openwin. On machines with only 104Mb drives, /usr/openwin is a symbolic link to /vol/openwin. On machines which have 207Mb drives, /usr/openwin has the goodies in it. This works out very well for network and server load balancing. (We also mount X11 as /vol/X11 with symoblic links from /usr/bin/X11 etc. which looks kinda funny since you have a link from /usr/bin/X11 to /vol/X11/bin :-). It works. We have one mount point. Everyone is happy. Richard Elling Manager of Network Support Auburn University Engineering Administration relling@eng.auburn.edu KB4HB [44.100.0.72] (205)844-2280 "I am not an animal! I am a human bean!" -- Peanutman
root@shawn.uucp (0000-Admin(0000)) (06/02/91)
bgeer@javelin.sim.es.com (Bob Geer) writes: >So, I'm tempted to "off" openwin & load up a "generic" X11R4 & Motif >into a more typical directory layout -- /usr/bin/X11, etc. Anyone got >comments on this plan -- good, bad, indifferent? The virgin X11R4 from MIT is infinitely superior to OpenWindows, both in performance, portability, and quality. You can certainly put Motif in, but the 'twm' window manager that comes free with the MIT stuff is also not bad...perhaps a good intermin solution while saving money to buy Motif... >Thanks in advance...Bob Mark Jeghers
uniel@uts.uni-c.dk (Erik Lawaetz) (06/03/91)
How about widget sets (and other stuff?) not supplied by SUN? Is it easy to use these with OpenWin, and how does one determine what to compile if the whole X11R4 isn't needed? ------------------------------------------------------------------------------- Erik Lawaetz UNI*C Lyngby Danish Computing Centre for Research and Education +45 45 93 83 55 DTH, Building 305 Erik.Lawaetz@uni-c.dk DK-2800 Lyngby -------------------------------------------------------------------------------
mh@roger.imsd.contel.com (Mike Hoegeman) (06/04/91)
In article <1991Jun01.232522.786@shawn.uucp> root@shawn.uucp (0000-Admin(0000)) writes: >bgeer@javelin.sim.es.com (Bob Geer) writes: > >>So, I'm tempted to "off" openwin & load up a "generic" X11R4 & Motif >>into a more typical directory layout -- /usr/bin/X11, etc. Anyone got >>comments on this plan -- good, bad, indifferent? > >The virgin X11R4 from MIT is infinitely superior to OpenWindows, both in >performance, portability, and quality. You can certainly put Motif in, >but the 'twm' window manager that comes free with the MIT stuff is also >not bad...perhaps a good intermin solution while saving money to buy >Motif... > >>Thanks in advance...Bob > >Mark Jeghers Infinitely ?? Give me a break. You probably have valid reasons on why you prefer the MIT X over xnews . You list none of them however. You merely state it is "infintely superior". What a crock. I 've used both on a sun 4/370 and on various sparc workstations and it is not infinitely worse. Yes it does use more memory. Yes it is slower on 8 meg and less machines but the difference is somewhat less than infinite. About Portability. I have doubts about your claim that the MIT server is more portable. That people run it on more machines there is no doubt. But is it more portable? I suspect you cannot even make an informed opinion on that. I apologize in advance in you have the source and have made an informed comparison but I doubt it. You do have to fiddle around with symbolic links at times for X source that uses hardcoded path names and the set of fonts avaiable between the two are not the same so sometimes you have to convert fonts from one to the other. You may put these under the category of portability I suppose but in the case of the symbolic links needed this is a strike against the MIT server if anything. The font issue is no big deal (for me at least you mileage may vary on this) Now what about quality? How to measure it is difficult task at best. I'll just leave it at the fact that I have found the openwindows (xnews) server in my own personal use just as robust as the MIT server. It also has NeWS built into it which I *love* having. It is display PostScript done _right_. I'm not trying to give any kind of "proof" that xnews is "better" than MIT X but simply stating that I have access to both and prefer xnews over the MIT server and given some reasons why. xnews is certainly not orders of magnitude worse than the MIT server in any respect and there is a lot to be said in favor of it. I suspect that you just don't like the fact that xnews source is not free and that the MIT server source is. This seems to offend you deeply and thus xnews is trash and MIT X is God's gift to windowdom.
bill@polygen.uucp (Bill Poitras) (06/04/91)
In article <1991May29.230342.15836@ncs.dnd.ca> rpburry@ncs.dnd.ca (Paul Burry) writes: >So, I have been thinking about tossing Openwindows and running MIT X11R4 and >one of the available window managers distributed with MIT X. > >Now, the questions: > 1. What Openwindows applications break when used with MIT X11R4? If you don't have availible in some sub-directory libolgx.a and libxview.a, then all the "tool" applications (mailtool, dbxtool etc) will not be able to find these libraries. If you leave the openwin directory on disk, then setting your LD_LIBRARY_PATH to $OPENWINHOME/lib will fix this. You might not be able to find certain fonts when running them. Converting Openwin fonts to .bdf format and installing them should fix this possible problem. > 2. What window manager should we run? What's your favorite? Basically its what you want. > 3. Will the Olwm distributed with MIT X11R4 work with my favorite > Openwindows applications? Yes. > >Thanks, Your welcome > Paul Bill +-----------------+---------------------------+-----------------------------+ | Bill Poitras | Polygen Corporation | {princeton mit-eddie | | (bill) | Waltham, MA USA | bu sunne}!polygen!bill | | | FAX (617)890-8694 | bill@polygen.com | +-----------------+---------------------------+-----------------------------+
root@shawn.uucp (0000-Admin(0000)) (06/06/91)
mh@roger.imsd.contel.com (Mike Hoegeman) writes: >In article <1991Jun01.232522.786@shawn.uucp> root@shawn.uucp (0000-Admin(0000)) writes: >>bgeer@javelin.sim.es.com (Bob Geer) writes: >> >>>So, I'm tempted to "off" openwin & load up a "generic" X11R4 & Motif >>>into a more typical directory layout -- /usr/bin/X11, etc. Anyone got >>>comments on this plan -- good, bad, indifferent? >> >>The virgin X11R4 from MIT is infinitely superior to OpenWindows, both in >>performance, portability, and quality. You can certainly put Motif in, >>but the 'twm' window manager that comes free with the MIT stuff is also >>not bad...perhaps a good intermin solution while saving money to buy >>Motif... >> >>>Thanks in advance...Bob >> >>Mark Jeghers >Infinitely ?? Give me a break. You probably have valid reasons on why >you prefer the MIT X over xnews . You list none of them however. You >merely state it is "infintely superior". What a crock. I 've used both >on a sun 4/370 and on various sparc workstations and it is not >infinitely worse. .....lots of rhetorical flaming deleted... In answer: I did not elaborate because I have limited time to read news and even less time to write long dissertations (sp?). I still hold to my opinion that the MIT release is superior in performance, portability, and quality. And it is *not*, as you put it, a "crock". It is simply my opinion, which you need not accept. Performance: I have used both. The MIT stuff is faster by my perception. Portability: OpenWin has little or none of the standard utilities of X-Windows. It is largely a politically motivated product, meant to continue looking as much like Sun's old technology as possible. I have to deal with many an X Window student who is bewildered by the nearly total lack of similarity between what the world largely regards as a "typical" X environment and what OW sticks them with. This, to me, is a valid component of "portability". Consider also that no Athena toolkit example programs can be demonstrated in OW. I'll grant you that OLIT is there for doing intrinsics programming, but still I feel that the Open Windows is hardly "open". Quality: The server core dumps often, both on Sun 3's and Sparcs. Dbxtool constantly crashes with no good reason. Scrollbars erratically jump you in the wrong direction and have virtually worthless visual feedback. Guide is bug ridden and breaks data consistently. Cursor warping is inconsistent, which is worse that having it always or never. Is this enough, or do I have to have more? The MIT port is better. SImple as that. >I suspect that you just don't like the fact that xnews source is not >free and that the MIT server source is. This seems to offend you deeply >and thus xnews is trash and MIT X is God's gift to windowdom. Nonsense. Don't put words in my mouth. What I have said is what I have said. I think OpenWindows is a very poor product and I prefer MIT X11R4 by a very wide margin. You don't have to agree. But don't include me in your straw-man burnings.
bgeer@javelin.sim.es.com (Bob Geer) (06/06/91)
Literally all of the responses I have received in e-mail or postings that have made it to our node are favorable to X11R4. Several correspondents are neutral or barely favorable to Openwin, many are downright derrogatory towards it. Prevailing attitude is, if you have enough disk space, keep both around, run X11R4, start Openwin only if you need it. Without enough disk space, off Openwin. A few of the Openwin tools are nice, but the hassles of dealing with Sun's bogus X directory layout is a large pain. A couple of considerate individuals included brief instructions for using Openwin programs under X11R4. When doing this I ran into "font not found", "lib___ not current" & "couldn't find xv_get" type error statements. These are probably resolvable, but I'm expected to be productive on my assigned projects, & resolving these things looks like a short-term career commitment by itself. So, we're going to off Openwin & load X11R4 & Motif. EDITORIAL RANTING: I had occasion to hear McNeeley (Sun bigwig) give a keynote address to the Novelle conference held here in Ootah recently. His presentation was primarily a bashing of the rest of the Unix/Networking community for not cowtowing to Sun's attempts to dominate/direct development of standards. He was especially derrogatory towards OSF participants. He didn't say why Sun's approach was better; his comments were essentially snide & facetious & without informational content. Another comment he made was that you should "throw away" any workstation you have that you replace with one of better performance -- a rather arrogant attitude towards those folks still effectively using Sun3's, & consistent with their stated intention of not supporting Sun3 SunOS past 4.1 or so. It looks like Openwin's idiosyncratic implimentation of X is consistent with the otherwise selfish & arrogant attitude reflected in the content of McNeeley's address -- any standard, as long as it's Sun's. & besides...I don't like their stupid keyboards! -- <> Bob `Bear' Geer <> bgeer%javelin.sim.es.com (this *should* work) <> <> cola-zombie <> speaking only for myself, one of my many tricks <> <> Salt Lake City, <> "We must strive to be more than we are, Lal." <> <> Ootah <> -- Cmdr. Data, learning schmaltz <>
jon@choctaw.b23b.ingr.com (Jon Stone) (06/07/91)
In article <1991Jun06.060208.181@shawn.uucp> root@shawn.uucp (0000-Admin(0000)) writes:
.....lots of rhetorical flaming deleted...
In answer: I did not elaborate because I have limited time to read news
and even less time to write long dissertations (sp?). I still hold to
my opinion that the MIT release is superior in performance, portability,
and quality. And it is *not*, as you put it, a "crock". It is simply
my opinion, which you need not accept.
Performance: I have used both. The MIT stuff is faster by my perception.
You obviously have not compared xnews to Xsun on a machine running a
GX accelerator. MIT's server treats it as a dumb frame buffer (read
the source), whereas Sun's xnews seems to take full advantage of the
GX's capabilities. BTW, all our tests have been on SPARCstation 2's.
The xnews server does seem to use more memory than Xsun.
As for the other xnews bugs you mention, I don't know much about them.
These are just my opinions.
Jon
--
----------------------------------------------------------------------
Jon Stone jon@ingr.com
Intergraph Corp., Huntsville, AL
"He who wants by the yard, but tries by the inch should be kicked by
the foot." - a cookbook
----------------------------------------------------------------------
sarima@tdatirv.UUCP (Stanley Friesen) (06/07/91)
In article <1991Jun06.060759.420@shawn.uucp> root@shawn.uucp (Admin) writes: >Performance: I have used both. The MIT stuff is faster by my perception. On what machine? Using what version of OpenWindows? On a Sparc IPC with 12 megabytes of memory OpenWindows 2.0 runs so fast I do not see how it is *possible* to percieve a faster speed. [I.e. it runs at the speeds that correspond to the speed of human perception]. [Colloquially, it updates the screen 'in the blink of an eye']. Now of course on my *8* megabyte IPC there is a difference. >Portability: OpenWin has little or none of the standard utilities of > X-Windows. ... > ... This, to me, is > a valid component of "portability". Consider also that no > Athena toolkit example programs can be demonstrated in OW. I admit that this is a problem with OpenWindows as distributed. But there is nothing keeping you system administrator from installing all of the remaining X libraries along with the one that came with OpenWindows. [I would prefer it if Sun distributed all of the MIT libraries with OpenWindows, but it is by no means a fatal flaw]. We did that here, and I am running several Athena Widget based programs just fine under OpenWindows. (I use xman regularly). Also, I find that the XView library is easier to program than any of the widget sets. (And I am *not* an old Sunview programmer, my first window programming was X based). >Quality: The server core dumps often, both on Sun 3's and Sparcs. Which version of OpenWindows *were* you running! I have *never* had the Sparc OpenWindows 2.0 server crash on me. In fact the only servers I have ever crashed were the Sun386i OpenWindows 1.0 and the MIT X11r4 server on a Sun386i. Yes, that's right, I have had the *MIT* server crash more often than my Sun server. > Dbxtool > constantly crashes with no good reason. Yep, dbxtool is a crock. That's why I use xdbx! > Scrollbars erratically > jump you in the wrong direction and have virtually worthless > visual feedback. Hmm, this sounds like an old version again. The only scrollbars I have ever had significant trouble with lately are the *Motif* scrollbars in Framemaker. Just today I jumped many pages forward in a document when I *thought* I was *dragging* the scrollbar. (I had missed the elevator by a few pixels and it was treating my button press as repeated clicks below the elevator). > Guide is bug ridden and breaks data consistently. I have never used Guide, and I hope I never do. If I am going to use an application generator, I would prefer one that is GUI independent. [I think apps should tailor themselves to the individual user at run time]. > Cursor warping is inconsistent, which is worse that having it > always or never. Is this enough, or do I have to have more? This is not an OpenWindows vs MIT issue - this is a GUI issue. It applies equally well to the OpenLook GUI support that comes with the MIT distribution. [Both olwm and an old version of Xview are on the MIT tapes]. > The MIT port is better. SImple as that. I have failed to find any truly consistant differences. Especially after making sure I had all the libraries. In summary I think my main complaint about OpenWindows is that it doesn't contain all of the libraries by default. And i consider this a *minor* problem with a trivial fix. -- --------------- uunet!tdatirv!sarima (Stanley Friesen)
dhingra@wsl.dec.com (Satyendra Dhingra) (06/07/91)
>.....lots of rhetorical flaming deleted... > >In answer: I did not elaborate because I have limited time to read news >and even less time to write long dissertations (sp?). I still hold to >my opinion that the MIT release is superior in performance, portability, >and quality. And it is *not*, as you put it, a "crock". It is simply >my opinion, which you need not accept. > >Performance: I have used both. The MIT stuff is faster by my perception. > >Portability: OpenWin has little or none of the standard utilities of > X-Windows. It is largely a politically motivated product, > meant to continue looking as much like Sun's old technology > as possible. I have to deal with many an X Window student Above commecnts are obviously coming out of complete ignorance, openlook is totally different from sunview interface. Check out your facts before posting > who is bewildered by the nearly total lack of similarity > between what the world largely regards as a "typical" X and what is typical X, I only understand it to mean that the server should understand the right protocol and nothing more. Every vendor has political/financial reasons for pushing their own products > environment and what OW sticks them with. This, to me, is > a valid component of "portability". Consider also that no > Athena toolkit example programs can be demonstrated in OW. Can you explain which parts don't work, and are the parts that do not work completely compliant with ICCCM. because if not then the problem is with your examples (in other words they are not portable). > I'll grant you that OLIT is there for doing intrinsics > programming, but still I feel that the Open Windows is > hardly "open". > Why not, XView source is available free of charge, specs for openlook are available in bookstores that carry such material. >Quality: The server core dumps often, both on Sun 3's and Sparcs. do you have any cases that back up your claims or are you just sitting on your terminal and dreaming up these "facts". I have exercised the xnews server to the bone and found it very robust. > Dbxtool > constantly crashes with no good reason. Scrollbars erratically > jump you in the wrong direction and have virtually worthless > visual feedback. These problems were only in the demo version, but that is all it was, a "demo version". Have you thought of switching to decaf > always or never. Is this enough, or do I have to have more? No, in all of your "impartial" comments, you have not given one specific example except for making blanket statements like "this does not work" or "this crashes". satyendra
mrd@ecs.soton.ac.uk (Mark Dobie) (06/07/91)
In <1991Jun06.060208.181@shawn.uucp> root@shawn.uucp (0000-Admin(0000)) writes: [I don't want to join in the flaming - these are just my comments] >Performance: I have used both. The MIT stuff is faster by my perception. I have seen several sets of benchmarks that support this. >Portability: OpenWin has little or none of the standard utilities of > X-Windows. It is largely a politically motivated product, > meant to continue looking as much like Sun's old technology > as possible. I have to deal with many an X Window student > who is bewildered by the nearly total lack of similarity > between what the world largely regards as a "typical" X > environment and what OW sticks them with. But you don't have to use the openwindows programs. I use xterm and xdbx with OW all the time because I hate cmdtool, shelltool and I never had success with dbxtool. This , to me, is > a valid component of "portability". Consider also that no > Athena toolkit example programs can be demonstrated in OW. Its portable in as much as it is an X server which will serve any X application. You can use what you like (bugs permitting). > I'll grant you that OLIT is there for doing intrinsics > programming, but still I feel that the Open Windows is > hardly "open". I think that olwm is a better window manager than [tg]wm. The pin-up menus and drag and drop capability are real improvements. I use olwm on sun3s running the MIT server. I like the OpenLook scrollbars too (but not enough to use shelltool). File Manager is nice, and you don't need OW to run that (just olwm for the drag and drop). >Quality: The server core dumps often, both on Sun 3's and Sparcs. Dbxtool > constantly crashes with no good reason. Scrollbars erratically > jump you in the wrong direction and have virtually worthless > visual feedback. Guide is bug ridden and breaks data consistently. > Cursor warping is inconsistent, which is worse that having it > always or never. Is this enough, or do I have to have more? > The MIT port is better. SImple as that. Well I have never had problems with the server or scrollbars. Dbxtool leaves something to be desired. The MIT server may be better, but I use OW to preview postscript occaisionally. Also, I find the xview toolkit *much* easier to develop with than anything else. Its horses for courses and you can mix and match. You don't need to use the OW server to use the window manager, you don't need olwm to run xview programs and you needn't use the deskset programs. Take your pick and use the best of all worlds. Mark. -- Mark Dobie M.Dobie@uk.ac.soton.ecs (JANET) University of Southampton M.Dobie@ecs.soton.ac.uk (Bitnet)
stumpf@Informatik.TU-Muenchen.DE (Markus Stumpf) (06/07/91)
In article <JON.91Jun6145247@choctaw.b23b.ingr.com>, jon@choctaw.b23b.ingr.com (Jon Stone) writes: |> |> You obviously have not compared xnews to Xsun on a machine running a |> GX accelerator. MIT's server treats it as a dumb frame buffer (read |> the source), whereas Sun's xnews seems to take full advantage of the |> GX's capabilities. BTW, all our tests have been on SPARCstation 2's. |> Dunno, who good/bad it is, but there has been a patch in comp.windows.x in the early days of X.V11R4. I haven't actually tested it, as we don't have suns with GX here, but maybe it's useful for some of you out there! It's rather short, so I included the whole thing after the sig. (No flames please!) -- +- Markus Stumpf Technische Universitaet Muenchen -+ | Institut fuer Informatik, Rechnerbetriebsgruppe | | stumpf@informatik.tu-muenchen.de Postfach 202420 | +- ...@{unido.uucp,relay.cs.net} D-8000 Muenchen 2, West Germany -+ Article 5311 of comp.windows.x: Path: lan!tumuc!fauern!unido!ira.uka.de!sun1.ruf.uni-freiburg.de!root From: root@sun1.ruf.uni-freiburg.de (Martin Walter) Newsgroups: comp.windows.x Subject: R4 Xsun, GX support Keywords: R4 Xsun GX cgsix Message-ID: <1990Jan16.140753.6017@sun1.ruf.uni-freiburg.de> Date: 16 Jan 90 14:07:53 GMT Organization: Rechenzentrum der Universitaet Freiburg, Deutschland Lines: 75 Unfortunately R4 does not support the GX. Until Sun will adapt Xsun to there hardware (hopefully!) I use following little 'quick and dirty' patch. It accelerates only simple bitblock moves within the screen, but is very useful in xterm-scrolling and opaque-moves. Martin. ######################################################################### *** server/Imakefile.orig Sun Dec 17 01:09:30 1989 --- server/Imakefile Tue Jan 16 13:30:27 1990 *************** *** 241,245 **** SUNOBJS = ddx/sun/sunInit.o $(FONTUTIL) SUNLIBS = $(SUN) $(CFB) $(DIX) $(BSD) $(SNF) $(MFB) $(MI) $(EXTENSIONS) ! SUNSYSLIBS = $(SYSLIBS) $(SUNWINDOWSLIBS) XsunDIRS = $(SUNDIRS) --- 241,245 ---- SUNOBJS = ddx/sun/sunInit.o $(FONTUTIL) SUNLIBS = $(SUN) $(CFB) $(DIX) $(BSD) $(SNF) $(MFB) $(MI) $(EXTENSIONS) ! SUNSYSLIBS = $(SYSLIBS) $(SUNWINDOWSLIBS) -lpixrect XsunDIRS = $(SUNDIRS) *** server/ddx/cfb/cfbbitblt.c.orig Fri Dec 8 02:35:42 1989 --- server/ddx/cfb/cfbbitblt.c Tue Jan 16 13:01:03 1990 *************** *** 217,220 **** --- 217,255 ---- } + /*---------------------------------------------------------------------*/ + + if ((pSrc->type == DRAWABLE_WINDOW) + && (pDst->type == DRAWABLE_WINDOW) + && (alu == GXcopy) + ) + { + #include <pixrect/pixrect_hs.h> + + static Pixrect *screen = 0; + static planemaskSav = -999999; + if (!screen) screen = pr_open("/dev/fb"); + + if (planemask != planemaskSav) { + planemaskSav = planemask; + pr_putattributes(screen,&planemaskSav); + } + + while(nbox--) + { + pr_rop( screen, + pbox->x1, pbox->y1, + pbox->x2 - pbox->x1, pbox->y2 - pbox->y1, + PIX_SRC, screen, + pptSrc->x, pptSrc->y); + pbox++; + pptSrc++; + } + /* Important: Wait for Completion! */ + pr_get(screen, 0, 0); + + } else + + /*---------------------------------------------------------------------*/ + /* special case copy */ if (alu == GXcopy && (planemask & PMSK) == PMSK) ######################################################################### -- _____________________________________________________________________________ Internet: mawa@sun1.ruf.uni-freiburg.de (132.230.1.1) | Rechenzentrum Uni X.400: G=martin;S=walter;OU=ruf;P=uni-freiburg;A=dbp;C=de | Freiburg, Germany
datri@convex.com (Anthony A. Datri) (06/08/91)
>I admit that this is a problem with OpenWindows as distributed. >But there is nothing keeping you system administrator from installing >all of the remaining X libraries along with the one that came with >OpenWindows. Sun gives libX11 a version number of 4.3, which causes ld.so warnings if you make the mistake of compiling against them and running on a machine with MIT stuff. My users have also had problems of strange behavior up to and including hanging when using the MIT libX11 with the xnews server. >Also, I find that the XView library is easier to program than any of >the widget sets. XView isn't strictly an OW-only thing. >Which version of OpenWindows *were* you running! I have *never* >had the Sparc OpenWindows 2.0 server crash on me. I had both xnews 2.0 and x11r4.pl9 crash (only on sun 4's) with certain invocations of xloadimage -slideshow. That hasn't happened to me since 3.0.1 came out. >> The MIT port is better. SImple as that. >I have failed to find any truly consistant differences. >Especially after making sure I had all the libraries. The xnews server lacks the shape extension, and the reservation of colormap cells for the NeWS side of things is annoying. I find xnews to use more memory and run more slowly than the X11R4 server. Not everyone has an SS2. Run both on a 3/60 and see if you change your mind a little. The xnews server also loves to LOSE icons -- they disappear *completely*. The inability to use normal fonts is also a real pain. In short, I tell my users to use the MIT server unless they have to run one of several tools from vendors who still cling to Suntools. -- Fly to the sky on GI-GI____________ and shout to datri@convex.com
jarober@aplcen.apl.jhu.edu (DE Robertson james an 740-9172) (06/08/91)
bgeer@javelin.sim.es.com (Bob Geer) writes: >It looks like Openwin's idiosyncratic implimentation of X is >consistent with the otherwise selfish & arrogant attitude reflected in >the content of McNeeley's address -- any standard, as long as it's >Sun's. No more idiosyncratic than Motif's implementation, and possibly more consistent due to the tighter 'look and feel' spec. My experience is that users prefer whatever they are exposed to first - whether it be OpenLook, Motif, or Twm. BTW, to say you prefer X11R4 is to say nothing - the Window Manager gives you the look and feel, not the server. James A Robertson jarober@aplcen.apl.jhu.edu
johnt@unipalm.uucp (John Taylor) (06/10/91)
stumpf@Informatik.TU-Muenchen.DE (Markus Stumpf) writes: >In article <JON.91Jun6145247@choctaw.b23b.ingr.com>, jon@choctaw.b23b.ingr.com (Jon Stone) writes: >|> >|> You obviously have not compared xnews to Xsun on a machine running a >|> GX accelerator. MIT's server treats it as a dumb frame buffer (read >|> the source), whereas Sun's xnews seems to take full advantage of the >|> GX's capabilities. BTW, all our tests have been on SPARCstation 2's. >|> >Dunno, who good/bad it is, but there has been a patch in comp.windows.x >in the early days of X.V11R4. I haven't actually tested it, as we don't have >suns with GX here, but maybe it's useful for some of you out there! If you need a standard MIT server, but with GX performace, you need to use the XTech X Server from Unipalm. This is a standard MIT X11R4 server that has been modified to use the full performance benefits of the GX. You can contact Unipalm on +44 954 211797, or email xtech@unipalm.co.uk. JohnT johnt@unipalm.co.uk
sarima@tdatirv.UUCP (Stanley Friesen) (06/12/91)
In article <1991Jun07.233354.2660@convex.com> datri@convex.com (Anthony A. Datri) writes: >Sun gives libX11 a version number of 4.3, which causes ld.so warnings if you >make the mistake of compiling against them and running on a machine with MIT >stuff. My users have also had problems of strange behavior up to and >including hanging when using the MIT libX11 with the xnews server. Well, we did not install the MIT libX11, just the libraries that were not part of OpenWindows (libXaw, libXmu, and libXext). >I had both xnews 2.0 and x11r4.pl9 crash (only on sun 4's) with certain >invocations of xloadimage -slideshow. That hasn't happened to me since 3.0.1 >came out. Aaah! I have never had occasion to use -slideshow. (I mostly use -onroot). >The xnews server lacks the shape extension, and the reservation of colormap >cells for the NeWS side of things is annoying. I find xnews to use more >memory and run more slowly than the X11R4 server. Not everyone has an SS2. >Run both on a 3/60 and see if you change your mind a little. The xnews >server also loves to LOSE icons -- they disappear *completely*. The >inability to use normal fonts is also a real pain. Yep, it does indeed use more memory, and on smaller, slower machines that becomes noticable (it is quite slow on 8 meg Sun386's as well). It depends alot on what machine you are running. Also, from what I hear, OW 3.0 is going to be a major improvement. -- --------------- uunet!tdatirv!sarima (Stanley Friesen)
bill@polygen.uucp (Bill Poitras) (06/12/91)
In article <JON.91Jun6145247@choctaw.b23b.ingr.com> jon@choctaw.b23b.ingr.com (Jon Stone) writes: > Performance: I have used both. The MIT stuff is faster by my perception. > >As for the other xnews bugs you mention, I don't know much about them. Could some one post a summary of xnews/Openwindows gripes. I would be interested in seeing them. The only problem I see with Openwindows is the fact that the X server doesn't do a good job of storing font properties. I ported a large system from X11R3 -> X11R4 -> xnews. The xnews port was a pain because I didn't know that we relied so heavily on the font properties of the server fonts, which xnews doesn't store. Also it doesn't have all the MIT extensions (SHAPE, Multi-Buffering, ...) As far as the speed and size of the server I think some people forget a couple of facts about xnews 1) It runs X. 2) It runs NeWS, which also makes it a nice postscript interpreter 3) It runs SunView applications. For the first person who says "Who cares", I can already say: "Some of my customers" MIT X does only 1. +-----------------+---------------------------+-----------------------------+ | Bill Poitras | Polygen Corporation | {princeton bu}!polygen!bill | | (bill) | Waltham, MA USA | - This space for rent - | | | FAX (617)890-8694 | bill@polygen.com | +-----------------+---------------------------+-----------------------------+
dan@scooter.rosemount.com (Dan Messinger) (06/13/91)
In article <1991Jun07.233354.2660@convex.com>, datri@convex.com (Anthony A. Datri) writes: |> The xnews |> server also loves to LOSE icons -- they disappear *completely*. This is my primary reason for NOT using OL. I would loose an icon at least a half dozen times a day. I reported it to Sun, and they claimed I had something configured wrong, or there was some MIT software contaminating the system. I don't buy it. I've lost icons while running several different configurations. I most often would run the xnews server (because there are a couple of 3rd party SunView programs that I occasionally needed to run), but all the clients are from the MIT tapes. I would be willing to believe that I was creating my own problems if I was using OL libraries with MIT clients, or vice versa. But the icons still vanish when I use 100% MIT clients on the xnews server. Same thing goes for the server crashes. I have never been able to reproduce the crash at will. Xnews would simply die every now and then. All these problems have completely dissappeared after switching back to the MIT Xsun server. (no other changes) |> In short, I tell my users to use the MIT server unless they have to run one |> of several tools from vendors who still cling to Suntools. |> Same here.
jc@raven.bu.edu (James Cameron) (06/14/91)
>>>>> On 13 Jun 91 14:06:56 GMT, dan@scooter.rosemount.com (Dan Messinger) said: |> In article <1991Jun07.233354.2660@convex.com>, datri@convex.com (Anthony |> A. Datri) writes: |> |> The xnews |> |> server also loves to LOSE icons -- they disappear *completely*. |> This is my primary reason for NOT using OL. I would loose an icon at least |> a half dozen times a day. I reported it to Sun, and they claimed I had I have NOT ONCE seen this problem. I have used xnews on both the console of a 4/380, Sparc2's, and an slc. I have NEVER simply lost icons. The reason *I* don't use xnews as the general server and DO use Xsun is that xnews does NOT work well with XDM and there are apparently some clients which like Xsun better than xnews. (* grin *) So, until I have figure out how to bypass the problems, or until OW 3.0 comes out (whichever comes first) I will continue to run it as the general X server and use xnews when I want to use its special capabilities. jc -- -- James Cameron (jc@raven.bu.edu) Signal Processing and Interpretation Lab. Boston, Mass (617) 353-2879 ------------------------------------------------------------------------------ "But to risk we must, for the greatest hazard in life is to risk nothing. For the man or woman who risks nothing, has nothing, does nothing, is nothing." (Quote from the eulogy for the late Christa McAuliffe.)
jarober@aplcen.apl.jhu.edu (DE Robertson james an 740-9172) (06/14/91)
dan@scooter.rosemount.com (Dan Messinger) writes: >In article <1991Jun07.233354.2660@convex.com>, datri@convex.com (Anthony >A. Datri) writes: >|> The xnews >|> server also loves to LOSE icons -- they disappear *completely*. >This is my primary reason for NOT using OL. I would loose an icon at least >a half dozen times a day. I reported it to Sun, and they claimed I had >something configured wrong, or there was some MIT software contaminating >the system. I don't buy it. I've lost icons while running several different I've heard this, but to me it sounds like an urban legend. In 2 years of using Openwin (from 1.x --> 3.0 beta) I have NEVER had this happen. And my experience with OL Vs vanilla MIT is that MIT is faster on a SUN 3, but that OL is just as fast on a Sparc. jarober@aplcen.apl.jhu.edu
barnett@grymoire.crd.ge.com (Bruce Barnett) (06/14/91)
In article <1991Jun14.040542.8286@aplcen.apl.jhu.edu> jarober@aplcen.apl.jhu.edu (DE Robertson james an 740-9172) writes: >|> The xnews >|> server also loves to LOSE icons -- they disappear *completely*. The only thing I have seen is an XView application that uses a icon and mask (so that the background is transparent) disappears when using TWM as a window manager. My program has an #IFDEF to work around this. -- Bruce G. Barnett barnett@crdgw1.ge.com uunet!crdgw1!barnett
fgreco@fis1026.govt.shearson.com (Frank Greco) (06/14/91)
> >The xnews server lacks the shape extension, and the reservation of colormap > >cells for the NeWS side of things is annoying. I find xnews to use more > >memory and run more slowly than the X11R4 server. Not everyone has an SS2. > >Run both on a 3/60 and see if you change your mind a little. The xnews > >server also loves to LOSE icons -- they disappear *completely*. The > >inability to use normal fonts is also a real pain. You can tell the server to favor NeWS apps with the -favorstatic option to xnews. Or, fyi, you can reduce the colormap contention with X apps with the '-cubesize small' flag. Frank G. I use "normal" fonts with xnews all the time...