morreale@laplata.scd.ucar.edu (01/13/90)
I have two problems concerning R4 on a Sun 3/60 running OS 4.0.3 This is pure R4 with the Xmu TWO to ONE patch installed. (If it makes a difference, I *only* rebuilt that library) Ths following is part of the make output using gcc 1.35, (I read this newsgroup :-) making ddx/sun rm -f sunInit.o gcc -DNOSTDHDRS -fstrength-reduce -fpcc-struct-return -c -O -I../../../. -I. -I../mfb -I../mi -I../../include -I../../.././X11 -DXDMCP -DSHAPE -DMITSHM -DMULTIBUFFER -DMITMISC -DINCLUDE_ALLOCA_H sunInit.c /usr/include/sundev/vuid_event.h:42: unterminated character constant *** Error code 1 make: Fatal error: Command failed for target `sunInit.o' Current working directory /xw/X11R4/src/mit/server/ddx/sun *** Error code 1 make: Fatal error: Command failed for target `ddx/sun' I checked the file: vuid_event.hh and line 43 is in the middle of an ifdef COMMENT. :-) Hep! Note that the server was sucessfully compiled on a different Sun3 using a somewhat older version of the OS (same compiler) and all is well. We are very confused. Unrelated problem...... In R3, we were able to use the xterm Tek window to view plots generated with our local graphics package. With R4, we get garbage on the screen. Note that, for our translator, we only have a Tek 4010 device driver. What changed in R4? This used to work. Any specific help would be GREATLY appreciated as *lots* of our X users depend on the Tek stuff for viewing plots. Note that I have not rebuilt xterm since I installed the Xmu patch.... (I am rebuilding it right now.) -PWM -- ------------------------------------------------------------------ Peter W. Morreale email: morreale@ncar.ucar.edu Nat'l Center for Atmos Research voice: (303) 497-1293 Scientific Computing Division
morreale@bierstadt.ucar.edu (Peter Morreale) (01/13/90)
In article <6000@ncar.ucar.edu>, morreale@laplata.scd.ucar.edu writes: > > I have two problems concerning R4 on a Sun 3/60 running OS 4.0.3 > This is pure R4 with the Xmu TWO to ONE patch installed. > (If it makes a difference, I *only* rebuilt that library) > > Ths following is part of the make output using gcc 1.35, (I read this > newsgroup :-) > > making ddx/sun > rm -f sunInit.o > gcc -DNOSTDHDRS -fstrength-reduce -fpcc-struct-return -c -O -I../../../. -I. -I../mfb -I../mi -I../../include -I../../.././X11 -DXDMCP -DSHAPE > -DMITSHM -DMULTIBUFFER -DMITMISC -DINCLUDE_ALLOCA_H sunInit.c > /usr/include/sundev/vuid_event.h:42: unterminated character constant > *** Error code 1 > make: Fatal error: Command failed for target `sunInit.o' > Current working directory /xw/X11R4/src/mit/server/ddx/sun > *** Error code 1 > make: Fatal error: Command failed for target `ddx/sun' > > I checked the file: vuid_event.hh and line 43 is in the middle of an > ifdef COMMENT. :-) > "And the Bonehead of the Year award goes to ...." Thanks to all who responded! I forgot to run "fixincludes" after building gcc 1.35 on the new system. "One more time Alice,... One more time!!" Ralph C. :-) BTW, Our Visualization people tell me that we just got a copy of a Tek 4014 graphcap to test out. This may solve the second problem that was in my first posting. -PWM ------------------------------------------------------------------ Peter W. Morreale email: morreale@ncar.ucar.edu Nat'l Center for Atmos Research voice: (303) 497-1293 Scientific Computing Division Consulting Office ------------------------------------------------------------------
keith@EXPO.LCS.MIT.EDU (Keith Packard) (01/13/90)
Peter W. Morreale at the Nat'l Center for Atmos Research writes: > making ddx/sun > rm -f sunInit.o > gcc -DNOSTDHDRS -fstrength-reduce -fpcc-struct-return -c -O -I../../../. -I. -I../mfb -I../mi -I../../include -I../../.././X11 -DXDMCP -DSHAPE > -DMITSHM -DMULTIBUFFER -DMITMISC -DINCLUDE_ALLOCA_H sunInit.c > /usr/include/sundev/vuid_event.h:42: unterminated character constant > *** Error code 1 > > I checked the file: vuid_event.hh and line 43 is in the middle of an > ifdef COMMENT. :-) ANSI C requires that even the portions of a source file which are eliminated with CPP conditionals be separatable into CPP tokens. This is because the '#' sign is now recognised even if it is not the first character of the line, unless it is inside a string constant. I think this is a terrible "feature", but we'll probably have to live with it. It looks like I fixed this problem on our machine by making a copy of vuid_event.h in /usr/local/lib/gcc-include/sundev/vuid_event.h which gcc looks for first and replacing the single quote mark on line 42 with two quote marks: have to be concerned. In fact, the vuid interface doesn''t Looks like another candidate for the ERRATA file. Keith Packard MIT X Consortum
rws@EXPO.LCS.MIT.EDU (Bob Scheifler) (01/13/90)
In R3, we were able to use the xterm Tek window to view plots generated
with our local graphics package. With R4, we get garbage on the screen.
Please send some (simple!) examples to xbugs, so we can try to track this down.