dinah@WILKINS.BCM.TMC.EDU (Dinah Anderson) (06/28/90)
I applied this patch and did a make world and installed the new software and include files. Most of the clients that use the Xt library would die and dump core. For example, xterm would startup, but the second you tried to type anything in the window, the process would die and dump core. I backed out the changes and recompiled and everything now works ok. Did anyone else have this problem? Dinah Anderson Manager of Systems Integration Baylor College of Medicine Houston, Texas internet: dinah@bcm.tmc.edu uucp: {rutgers,mailrus}!bcm!dinah
envbvs@epb2.lbl.gov (Brian V. Smith) (06/29/90)
In article <9006271916.AA00839@NICOLLE.IAIMS.BCM.TMC.EDU>, dinah@WILKINS.BCM.TMC.EDU (Dinah Anderson) writes: |> |> I applied this patch and did a make world and installed the |> new software and include files. Most of the clients that use the |> Xt library would die and dump core. For example, xterm would startup, |> but the second you tried to type anything in the window, the process would |> die and dump core. |> |> I backed out the changes and recompiled and everything now works ok. Did |> anyone else have this problem? |> |> Dinah Anderson Manager of Systems Integration |> Baylor College of Medicine Houston, Texas |> internet: dinah@bcm.tmc.edu uucp: {rutgers,mailrus}!bcm!dinah Yes! Here is a core dump of xterm: You get it when you type any character in it. dbx xterm dbx version 2.0 of 10/19/88 2:02. Type 'help' for help. reading symbolic information ... (dbx) run -e bea Bus error in _XtMatchUsingDontCareMods at 0x28c74 00028c74 movl (r11),(r1) (dbx) where _XtMatchUsingDontCareMods(0x86484, 0x7fffda6c) at 0x28c74 MatchEvent(0x7eb84, 0x7fffda6c) at 0x28ec4 _XtTranslateEvent(0x84404, 0x84434, 0x7fffdf08, 0x7fffdad0) at 0x290e1 DispatchEvent(0x7fffdf08, 0x84404, 0x1, 0x66908) at 0x19981 DecideToDispatch(0x7fffdf08) at 0x19d2f XtDispatchEvent(0x7fffdf08) at 0x19d9e xevents() at 0xc1c1 in_put() at 0x36c9 VTparse() at 0x2866 VTRun() at 0x4987 main.main(0x2, 0x7fffe0ac, 0x7fffe0bc) at 0x892 _____________________________________ Brian V. Smith (bvsmith@lbl.gov) Lawrence Berkeley Laboratory I don't speak for LBL, these non-opinions are all mine.
envbvs@epb2.lbl.gov (Brian V. Smith) (06/29/90)
Sorry, cardinal sin time. I forgot to say: gcc1.36, Vaxstation II Xqvss, Ultrix 3.0, X11R4 patches 1-12 gcc options: -fwritable-strings -fstrength-reduce -fpcc-struct-return _____________________________________ Brian V. Smith (bvsmith@lbl.gov) Lawrence Berkeley Laboratory I don't speak for LBL, these non-opinions are all mine.
rws@EXPO.LCS.MIT.EDU (Bob Scheifler) (06/29/90)
For example, xterm would startup, but the second you tried to type anything in the window, the process would die and dump core. Hmm, I rebuilt the public patched Xt on both a DECstation 3100 (cc) and a VAXstation 3100 (gcc), and linked xterms with them, and both xterms seemd to work just fine. Beats me. Has anyone gotten fix-12 to work? Is anyone else having problems? (If others are, I'll probably withdraw it until something gets resolved.)
evans@decvaxdec.com (Marc Evans) (06/29/90)
In article <9006282345.AA22953@expire.lcs.mit.edu>, rws@EXPO.LCS.MIT.EDU (Bob Scheifler) writes: |> Has anyone gotten fix-12 to |> work? Is anyone else having problems? (If others are, I'll probably |> withdraw it until something gets resolved.) I have applied the patch without error top the sources, built using 'make World' (I didn't have objects around still), and installed the result onto a PMAX. It has been running for about 3 days now without any problems. Also, I have rebuilt many clients not included in the *statdard* MIT distribution which also work fine. Therefore, at least from the PMAX perspective, I don't think the patch should be withdrawn... - Marc =========================================================================== Marc Evans - WB1GRH - evans@decvax.DEC.COM | Synergytics (603)635-8876 Unix and X Software Contractor | 21 Hinds Ln, Pelham, NH 03076 ===========================================================================
jody@shell.COM (Jody Winston) (06/29/90)
I've gotten fix-12 to work on Sun4s OS 4.0.3 using cc.
envbvs@EPB2.LBL.GOV (Brian V. Smith) (06/29/90)
me: For example, xterm would startup, but the second you tried to type anything in the window, the process would die and dump core. Bob: Hmm, I rebuilt the public patched Xt on both a DECstation 3100 (cc) and a VAXstation 3100 (gcc), and linked xterms with them, and both xterms seemd to work just fine. Beats me. Has anyone gotten fix-12 to work? Is anyone else having problems? (If others are, I'll probably withdraw it until something gets resolved.) I'll try it on our 4/280 with cc and see what happens. Brian
envbvs@epb2.lbl.gov (Brian V. Smith) (06/29/90)
In article <9006282345.AA22953@expire.lcs.mit.edu>, rws@EXPO.LCS.MIT.EDU (Bob Scheifler) writes: |> |> For example, xterm would startup, |> but the second you tried to type anything in the window, the process would |> die and dump core. |> |> Hmm, I rebuilt the public patched Xt on both a DECstation 3100 (cc) |> and a VAXstation 3100 (gcc), and linked xterms with them, and both |> xterms seemd to work just fine. Beats me. Has anyone gotten fix-12 to |> work? Is anyone else having problems? (If others are, I'll probably |> withdraw it until something gets resolved.) It seems to be related to typing into a text widget. In xterm it dies, and in xfig it dies when typing into the change-object menu text widgets. Again, Vaxstation II Xqvss, gcc1.36, Ultrix 3.0, X11R4-PL12. It works fine on a 4/280 under SunOs 4.1 with cc and, as several other people have noted, on other suns. Please help. I will have to back out patch 12 until it is determined what the problem is. Thanks for any help. Here again is the procedure where the death occurs (bus error): _XtMatchUsingDontCareMods(0xbfaa0, 0x7fffd820) at 0x3f6b0 MatchEvent(0xbd4c4, 0x7fffd820) at 0x3f900 _XtTranslateEvent(0xbb604, 0xbb634, 0x7fffdcbc, 0x7fffd884) at 0x3fb1d DispatchEvent(0x7fffdcbc, 0xbb604, 0x1, 0x89508) at 0x2feb9 DecideToDispatch(0x7fffdcbc) at 0x30267 XtDispatchEvent(0x7fffdcbc) at 0x302d6 XtAppMainLoop(0x89404) at 0x30589 XtMainLoop() at 0x30565 main.main(0x0, 0x7fffddf4, 0x7fffddfc) at 0xd818 _____________________________________ Brian V. Smith (bvsmith@lbl.gov) Lawrence Berkeley Laboratory I don't speak for LBL, these non-opinions are all mine.
kaleb@mars.jpl.nasa.gov (Kaleb Keithley) (06/30/90)
In article <9006282345.AA22953@expire.lcs.mit.edu> rws@EXPO.LCS.MIT.EDU (Bob Scheifler) writes: > > For example, xterm would startup, > but the second you tried to type anything in the window, the process would > die and dump core. > >Hmm, I rebuilt the public patched Xt on both a DECstation 3100 (cc) >and a VAXstation 3100 (gcc), and linked xterms with them, and both >xterms seemd to work just fine. Beats me. Has anyone gotten fix-12 to >work? Is anyone else having problems? (If others are, I'll probably >withdraw it until something gets resolved.) I had this problem with fix-12 also on our Sun 4s. Then I went back and did: make clean. make depend. make. make install. ldconfig. and then it started working. kaleb@thyme.jpl.nasa.gov Jet Propeller Labs Kaleb Keithley "So that's what an invisible barrier looks like"
hosking@cs.umass.edu (Tony &) (06/30/90)
I had similar problems as described in previous postings with installing fix 12 but I think I know what's going on. After applying the patch, I did a make (not a "make World"!!!), in each of the 3 versions of the X hierarchy we have around (one for the DECstation 3100, one for the VAXstations, and one for a Mac II). As it turns out, make on the VAXstation rebuilt the *entire* Xt library, whereas make on the DS3100 and the Mac only selectively rebuilt files in mit/lib/Xt. When I tried to get X going on the VAXstation I had no problems, however the DS3100 and the Mac II exhibited the problems that were described in earlier postings. So, all I did was to "make clean" in mit/lib/Xt and do a top-level make (not "make World") for each of the DS3100 and Mac II mit hierarchies, and everything turned out OK. I guess there are some weird differences in make between VAXen and the DECstations and Mac IIs. Hope this helps. -- Tony Hosking Dept. of Computer and Information Science _--_|\ University of Massachusetts / \ Amherst, MA 01003 \_.--._/ ) (413) 545-0256; hosking@cs.umass.edu v /
envbvs@epb2.lbl.gov (Brian V. Smith) (06/30/90)
In article <HOSKING.90Jun29140113@ibis.cs.umass.edu>, hosking@cs.umass.edu (Tony &) writes: |> I had similar problems as described in previous postings with installing fix |> 12 but I think I know what's going on. After applying the patch, I did a make |> (not a "make World"!!!), in each of the 3 versions of the X hierarchy we have |> around (one for the DECstation 3100, one for the VAXstations, and one for a |> Mac II). As it turns out, make on the VAXstation rebuilt the *entire* Xt |> library, whereas make on the DS3100 and the Mac only selectively rebuilt files |> in mit/lib/Xt. When I tried to get X going on the VAXstation I had no |> problems, however the DS3100 and the Mac II exhibited the problems that were |> described in earlier postings. So, all I did was to "make clean" in mit/lib/Xt |> and do a top-level make (not "make World") for each of the DS3100 and Mac II |> mit hierarchies, and everything turned out OK. |> |> I guess there are some weird differences in make between VAXen and the |> DECstations and Mac IIs. |> Actually, on our Vaxstation (Ultrix 3.0) it did not make the entire Xt library, just the files that changed - and that is where the problem was for our case. When I did a "make clean", as you did for your DS3100 then everything worked ok. >>> So, the answer appears to be "do a make clean before rebuilding Xt". _____________________________________ Brian V. Smith (bvsmith@lbl.gov) Lawrence Berkeley Laboratory I don't speak for LBL, these non-opinions are all mine.
kaleb@mars.jpl.nasa.gov (Kaleb Keithley) (06/30/90)
In article I wrote: -In article <9006282345.AA22953@expire.lcs.mit.edu> rws@EXPO.LCS.MIT.EDU (Bob Scheifler) writes: -> -I had this problem with fix-12 also on our Sun 4s. Then I went back and did: -make clean. -make depend. -make. -make install. -ldconfig. - I should have mentioned that I did this in src/mit/lib/Xt. kaleb@thyme.jpl.nasa.gov Jet Propeller Labs Kaleb Keithley "So that's what an invisible barrier looks like"
mike@tekgen.BV.TEK.COM (Mike Ewan) (06/30/90)
I've installed and run with fix-12 on Sun3 SunOS4.0 with no problems. I did a make Everything to install the fix-12. It's been running a couple of days. I've also installed it on a VAXstation II/GPX and it works there, but I haven't tested it much. Mike
john@acorn.co.uk (John Bowler) (07/02/90)
MIT public fix #12 changed InitialI.h; with the (sole) exception of Functions.c every lib/Xt .c file includes IntrinsicI.h, which includes InitialI.h. Therefore everything (except for Functions.o) should be rebuilt. The change in InitialI.h made a substantial change to the XtPerDisplay structure - so if you have a system which recompiled without rebuilding everything (possibly because of an omitted or defective make depend step in the distant past) you probably now have a defective libXt.a... John Bowler (jbowler@acorn.co.uk)
stumpf@tuminfo2.informatik.tu-muenchen.de (Markus Stumpf) (07/02/90)
In article <9006282345.AA22953@expire.lcs.mit.edu>, rws@EXPO.LCS.MIT.EDU (Bob Scheifler) writes: |> Hmm, I rebuilt the public patched Xt on both a DECstation 3100 (cc) |> and a VAXstation 3100 (gcc), and linked xterms with them, and both |> xterms seemd to work just fine. Beats me. Has anyone gotten fix-12 to |> work? Is anyone else having problems? (If others are, I'll probably |> withdraw it until something gets resolved.) Built it on DECstation 3100 (cc) and mVaxII (gcc) without any problems! Everything works fine! \Maex +- 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 -+
morreale@bierstadt.scd.ucar.edu (Peter Morreale) (07/03/90)
|> In article <9006282345.AA22953@expire.lcs.mit.edu>, rws@EXPO.LCS.MIT.EDU |> (Bob Scheifler) writes: |> |> Hmm, I rebuilt the public patched Xt on both a DECstation 3100 (cc) |> |> and a VAXstation 3100 (gcc), and linked xterms with them, and both |> |> xterms seemd to work just fine. Beats me. Has anyone gotten fix-12 to |> |> work? Is anyone else having problems? (If others are, I'll probably |> |> withdraw it until something gets resolved.) |> I've built on Sun3's, Sun4's with OS 4.0.3 and 4.1. All is fine. (including a brand new Sun4/490 running 4.1, the "World" built without a hitch...) -PWM ------------------------------------------------------------------ Peter W. Morreale email: morreale@ncar.ucar.edu Nat'l Center for Atmos Research voice: (303) 497-1293 Scientific Computing Division Consulting Office ------------------------------------------------------------------
peirce@gumby.cc.wmich.edu (Leonard Peirce) (07/04/90)
On an unrelated note, I had a small problem applying fix 12. The patches bombed on mit/lib/Xt/Varargs.c. I put them in by hand with no problem. But I'm surprised that it failed; I kept up with all of the patches (or so I thought). Did anyone else have this problem? -- Leonard Peirce Internet: peirce@gumby.cc.wmich.edu Western Michigan University peirce@gw.wmich.edu Academic Computer Center UUCP: ...!uunet!sharkey!wmichgw!peirce Kalamazoo, MI 49008 Phone: (616) 387-5469