thompson@uxe.cso.uiuc.edu (05/26/88)
I am having trouble getting X.V11R2 up and running on a Sun 3/160 which may come as a surprise to many people since it seems to be such an easy installation. It compiles and installs without any errors. When I start it up, the server puts up the gray X background, and places the "X" mouse in the center of the screen just like it should. Then it dies. I get back my unix prompt with keyboard locked. The initial xterm console window is never displayed on the screen. No error messages are printed out to the console. The X server dumps core in my current directory that dbx says it died in: cfbPadPixmap(0x3) at 0x90a0 cfbChangeWindowAttributes(0x1c333c, 0x0) at 0xb11b sunChangeWindowAttributes(0x1c333c, 0x4) at 0x3ff7 ChangeWindowAttributes(0x1c333c, 0x0, 0x1c16dc, 0x1c25c4) at 0x36e53 ProcChangeWindowAttributes(0x1c25c4) at 0x1f28b Dispatch() at 0x1f0a5 main(0x2, 0xefffe68, 0xefffe74) at 0x2d4fd I know that the server has gotten as far as invoking xterm by looking at lastcomm. But xterm doesn't make it to actually being displayed on the screen. I have tried substituting X.V11R1's xterm. X.V11R1 xterm is displayed on the screen, but it won't accept keyboard input. I've applied patches 1-10 to X.V11R2, but that didn't seem to make a difference. Has anyone else out there had these problems compiling it on a Sun 3/160C running SunOS Release 3.3? The installation has been made so simple, that it's really frustrating looking for a problem. The only modification that I make to the source is in ./util/imake.includes/Imake.tmpl to tell it where to put the binaries, libraries, etc. I do have some experience in installing these things from X.V10R4 and X.V11R1.93 but I'm really stumped with this one. Any help would be much appreciated. Thanks in advance, Dave Thompson UUCP: {ihnp4,convex,pur-ee}!uiucdcs!newton!davet INTERNET: davet@ncsa.uiuc.edu CSNET: thompson%uxe@uiuc.csnet BITNET: 14011@NCSAVMSA.BITNET
takagi@icot32.icot.junet (Takagi Shigeyuki) (05/28/88)
In article <61800003@uxe.cso.uiuc.edu> thompson@uxe.cso.uiuc.edu writes: >Has anyone else out there had these problems compiling it on a Sun 3/160C >running SunOS Release 3.3? The installation has been made so simple, ^^^^^^^^^^^^^^^^^ I had the same problem with SUN 3/260HM running OS 3.3. Similar installation problems occurred with both X11R1 and X11R2. When I changed my OS from 3.3 to 3.5, everything went fine. (Of course, you need to do "make Everything".) I think this is a bug of 3.3's C compiler or library. S.Takagi takagi%icot.jp@relay.cs.net
jw@obelisk.sics.se (Johan Widen) (06/01/88)
>I am having trouble getting X.V11R2 up and running on a Sun 3/160 >which may come as a surprise to many people since it seems to be such an easy >installation. It compiles and installs without any errors. When I start it >up, the server puts up the gray X background, and places the "X" mouse in >the center of the screen just like it should. Then it dies. I get back >Has anyone else out there had these problems compiling it on a Sun 3/160C >running SunOS Release 3.3? The installation has been made so simple, Yes! The culprit seems to be the compiler. Try the SunOS 3.4 compiler. Worked like a charm for me. -- Johan Widen SICS, PO Box 1263, S-164 28 KISTA, SWEDEN Tel: +46 8 752 15 32 Ttx: 812 61 54 SICS S Fax: +46 8 751 72 30 Internet: jw@sics.se or {mcvax,munnari,ukc,unido}!enea!sics.se!jw
wohler@joyce.UUCP (06/05/88)
while we're talking about strange installations of the sun server and clients, note that i recently discovered that the 4.3BSD netdb.h for the nameserver, if installed in /usr/include, will cause all clients to die with a segmentation fault in the XOpenDisplay code. it appears that #define h_addr h_addr_list[0] /* address, for backward compatiblity */ doesn't work as advertised. we've noticed this in other applications as well. can anyone explain? --bw
mic@ut-emx.UUCP (Mic (... K[a-z]+) Kaczmarczik) (06/06/88)
In article <12879@joyce.istc.sri.com> wohler@spam.istc.sri.com (Bill Wohler) writes: > > while we're talking about strange installations of the sun server > and clients, note that i recently discovered that the 4.3BSD netdb.h > for the nameserver, if installed in /usr/include, will cause all > clients to die with a segmentation fault in the XOpenDisplay code. > > it appears that > >#define h_addr h_addr_list[0] /* address, for backward compatiblity */ > > doesn't work as advertised. we've noticed this in other > applications as well. can anyone explain? I had the same problem when I rebuilt X11R2 on our cluster of Suns a few weeks after after we had installed the 4.3 netdb.h. I traced through the code and had it isolated to the line in XOpenDisplay() that calls gethostbyname(). The problem is that the SunOS libc versions of gethostbyname(), etc. still use the old definition of h_addr -- they return a (char *) instead of a (char **), so XOpenDisplay() was trying to use the host's Internet address as a character pointer! We thought of two solutions: 1) re-install the SunOS netdb.h and only use the 4.3 netdb.h when building the things we want to use the resolver library with, or 2) change the Sun makefiles for X to link everything against the resolver library. Either way makes the definitions of h_addr match up again; we opted for the first approach because it was simpler and required fewer changes to our systems. --mic-- -- Mic Kaczmarczik If you drink, don't drill. UT Austin Computation Center -- Matt Groening mic@emx.utexas.edu MIC@UTAIVC.BITNET