mark@comp.vuw.ac.nz (Mark Davies) (03/02/90)
Has anyone build R4 under HP-UX 7.0 on a series 300 machine? Everything built and worked fine under 6.5. Under 7.0 everything builds without complaint (except for warnings because SIGCHLD is now defined in signal.h), the server works, some of the clients work, but other clients (notably xterm, xman, xcalc, xdm) fail immediately with the message: Cannot perform realloc (from lib/Xt/Alloc.c) There should be plenty of swap space (the 6.5 built clients work in the same circumstances). The only thing I changed (apart from the operating system :-) was in config/hp.cf change the OSName etc to: #define OSName HP-UX 7.0 #define OSMajorVersion 7 #define OSMinorVersion 0 Any ideas what is going on? cheers mark
bobm@hpcvlx.cv.hp.com (Bob May) (03/03/90)
> Has anyone build R4 under HP-UX 7.0 on a series 300 machine? > Everything built and worked fine under 6.5. > Under 7.0 everything builds without complaint (except for warnings because > SIGCHLD is now defined in signal.h), the server works, some of the clients > work, but other clients (notably xterm, xman, xcalc, xdm) fail immediately > with the message: > > Cannot perform realloc > > Any ideas what is going on? There's a bug in the hp.cf file on the standard distribution bits. The compiler flag MALLOC_0_RETURNS_NULL must be defined if you are using the standard libc malloc package. If you don't, you see the realloc errors. Try adding -DMALLOC_0_RETURNS_NULL to the compile line for the Xt code, and recompile Alloc.c. This should help you. Yes, the bug's been submitted to MIT... > cheers > mark Bob May Hewlett-Packard Interface Technology Operation bobm@hpcvlx.cv.hp.com
dgreen@monet.watson.ibm.com (Dan Greening) (03/03/90)
mark@comp.vuw.ac.nz (Mark Davies) writes in comp.windows.x: > Has anyone build R4 under HP-UX 7.0 on a series 300 machine? > > Everything built and worked fine under 6.5. > Under 7.0 everything builds without complaint (except for warnings because > SIGCHLD is now defined in signal.h), the server works, some of the clients > work, but other clients (notably xterm, xman, xcalc, xdm) fail immediately > with the message: > > Cannot perform realloc > > (from lib/Xt/Alloc.c) > There should be plenty of swap space (the 6.5 built clients work in the > same circumstances). > > The only thing I changed (apart from the operating system :-) was in > config/hp.cf change the OSName etc to: > #define OSName HP-UX 7.0 > #define OSMajorVersion 7 > #define OSMinorVersion 0 > > Any ideas what is going on? > > cheers > mark This is exactly the problem I had. A similar problem was reported to me by a colleague at Washington State University. I posted a request for help to comp.sys.hp about three weeks ago. I received this response from Kris Olander, who runs on a 9000/340: > I've compiled X11R4 (with patched through fix-2) with very few > difficulties. I wasn't able to startup X fresh from compiling until > I made /usr/bin/X a soft link to /usr/bin/Xhp, chmod 2755 on Xhp,xterm, > and xload, and finally, chgrp sys Xhp,xterm,xload. It was the only helpful response I've received. HP techs told me "gosh, I thought everything worked in HP 7.0 :-)" and that was about it. I have not been able to test this solution, since I am now physically distant from the HP machines I help administer. IF THE ABOVE METHOD SOLVES THE "Cannot perform realloc" PROBLEM, COULD SOMEONE PLEASE LET ME KNOW BY E-MAIL? Thanks. Dan R. Greening 12 Foster Court NY (914) 789-7860 dgreen@cs.ucla.edu Croton-on-Hudson, NY 10520 CA (213) 825-2266
mark@comp.vuw.ac.nz (Mark Davies) (03/04/90)
I write about X clients failing under HP-UX 7.0 with: Cannot perform realloc In article <1990Mar2.185840.24495@arnor.uucp> dgreen@cs.ucla.edu writes: >This is exactly the problem I had. A similar problem was reported to me >by a colleague at Washington State University. I posted a request for >help to comp.sys.hp about three weeks ago. I received this response from >Kris Olander, who runs on a 9000/340: >> I've compiled X11R4 (with patched through fix-2) with very few >> difficulties. I wasn't able to startup X fresh from compiling until >> I made /usr/bin/X a soft link to /usr/bin/Xhp, chmod 2755 on Xhp,xterm, >> and xload, and finally, chgrp sys Xhp,xterm,xload. >IF THE ABOVE METHOD SOLVES THE "Cannot perform realloc" PROBLEM, COULD >SOMEONE PLEASE LET ME KNOW BY E-MAIL? Thanks. No it doesn't. The correct fix came from Bob May <bobm@hpcvlx.cv.hp.com> in article <100920183@hpcvlx.cv.hp.com> in comp.windows.x. Seems under 7.0 you need to define MALLOC_0_RETURNS_NULL. I did this and everything now works smoothly. cheers mark
ben@hpcvlx.cv.hp.com (Benjamin Ellsworth) (03/06/90)
> Seems under 7.0 you need to define MALLOC_0_RETURNS_NULL. > I did this and everything now works smoothly. > > mark Yup. Seems that the config file in the MIT distribution missed that define. --- Ben