Re: [opendx-dev] [PATCH] delete of something allocated with new[]
These have already been rolled into 4.3.2. David Original Message From: [EMAIL PROTECTED] Date: Tue 12/9/03 6:03 To: opendx2-dev@lists.berlios.de Subject:Re: [opendx-dev] [PATCH] delete of something allocated with new[] I'll try out these patches unless someone else wants the job. This patch should take care of the delete of something allocated with new[] part of http://opendx.watson.ibm.com/dx/mailArchives/mails.html/opendx-dev.0310/msg5 .html
Re: [opendx-dev] ImageMagick5.2.2 and xlc
I can't even find this code in my version of ImageMagick-5.2.2. That line looks like it has been mucked up. Are you sure something hasn't happened to it? David Hello- I know this might now be the correct place to ask the following question, but I also know there are may skilled people reading this list. My goal is to compile ImageMagick5.2.2 with AIX4.2 and compiler xlc. In the IM file draw.c the xlc compiler crashes at the following line(line 1439) p-pixel=p == primitive_info ? start : (p-1)-pixel; with the error message: draw.c, line 1439.49: 1506-127 (S) The second and third operands of the condit ional operator must have compatible struct or union types. make[1]: *** [draw.lo] Error 1 gcc compiles and builds the libraries and utilities correctly but adds the symbol __eprintf. This symbol hangs up the dx configure script when using CC=xlc when checking for GetImageInfo() in the ImageMagick library. Any sugestions would be greatly appreciated. Thanks, Randy -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Problem in graph.c with ArrangeMember
Greg, Randall, With the latest updates on the cache graph stuff, there is a problem with the ArrangeMember macro. With the SuperviseState and SuperviseWindow in the macro, a core dump occurs. I've traced it to the new function Greg built, but it looks like it is a problem with the Async modules being in a macro. As the function is called recursively, it seems to have problems. Any clue as to the easy fix? If not, I will start digging into the gory details. David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] _tiff.c????
Richard, Do you know if the libtiff can handle these? If so, it probably wouldn't take much to write a module around their function calls. David While you guys are looking at tiff code, I have a question. Ever encounter a multi-tiff? This seems to be the standard in confocal microscopy and I'm starting to see them more often. ReadImage will treat one like an ordinary tiff, displaying only the first frame in the stack. What I really need is a module that will detect the appropriate tags and output a series or other group that can then be stacked for volumetric work. Richard -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] _tiff.c????
Corrected, I think? The reason is, they are never used on a bigendian system and since they are static they use space in the program stack. Please let me know if you find any others, but I believe this was the only one like this. I thought that it might be a problem, but didn't have a LE system to test on first. I've tested all my commits with xlC on AIX and gcc on a Sun and LinuxPPC. The new fix, ifdefs it out on bigendian systems and ifdefs it in on little endian systems. David David, A couple routines in _tiff.c got ifdeffed out - write_rev_uint32 and write_rev_uint16. They're needed - why did they go away? Greg -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
RE: [opendx-dev] dx/include/dx arch.h
Yes, it should fix the (intel) as well. It was a problem where the defines from arch.h would be substituted into the actual functions and then get all fowled up for certain files in libdx. It's due to the fact that Solaris 8 now includes these fuctions natively. Try it and see. It should fix it. David, Are these fixes for Solaris8 x86 (Intel)? I could not get OpenDX 4.1.1 compiled on Solaris8 x86(intel), using GCC 2.95.2, Mesa 3.2.1, CDF, NetCDF, HDF and ImageMagick 5.2.2. It throws the compiler out right after compiling a very few files in scr/exec/libdx. I never got time to investigate it properly. Suhaib -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of David Thompson Sent: Thursday, August 10, 2000 7:19 PM To: opendx2-dev@lists.berlios.de Subject: [opendx-dev] dx/include/dx arch.h Update of /src/master/dx/include/dx In directory opendx.watson.ibm.com:/tmp/cvs-serv51804 Modified Files: arch.h Log Message: Included is a fix for Solaris 8 which now includes these functions. The defines were throwing off compilation. This fix may cause problems with Solaris 2.6, but I don't believe so. -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Any objections to byte def.
For the DX_CHECK_TYPE(byte, char) would there be any objections to changing this to DX_CHECK_TYPE(byte, signed char)? For systems that already use signed, this wouldn't make any difference and for systems that use char as unsigned char it would correct the error. Please speak up!! David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] build on linux-alpha machines?
We've only had one other person raise this question and he never responded back. I wonder if he didn't give up? Please let us know if you have any success. If you are successful, we'd be willing to put the binaries up on OpenDX.org. David I've poked around in the list archives and www.opendx.org, but haven't found much in the way of information about opendx running on linux-alpha machines (redhat). Does anyone out there know of efforts to build opendx on such machines, and if they exist how successful they've been? Thanks much, Nathan Barton -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Quiet on the Western Front
It sure has been quiet as of late. Everyone off enjoying their summer vacations? I'm just wondering who all will be attending VIS this year in Salt Lake City? We're planning on going and possibly setting up a booth but wanted to know if anyone else is planning an OpenDX type of booth (possibly to share costs). How about BOF meeting, any plans? David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
RE: [opendx-dev] SGI Image Magick 5.2 - FreeMemory
If you use the netcdf 2.7f from OpenDX.org, I have renamed the FreeMemory in netcdf to something else. I figure that netcdf changes a lot less frequently than ImageMagick does. David I think it is a god idea. But ImageMagick FreeMemory is DIRTY. Its causes conflicts not only with DX, but also CDF. Therefore I had been compiling ImageMagick after chaning FreeMemory to _FreeMemory in ImageMagick code. Suhaib -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] IRIX dxexec build errors
Sounds like your configure goof'd. Look at config.cache for ushort and tell us if it says yes or no. Are you saying that your /usr/include/sys/bsd_types.h has it declared twice? If so, there isn't much we can do about bad header files not provided by us. David I've been out of the loop for a while but am finally getting around to building the latest cvs version of dx on my IRIX system (using sgi's compilers) I get the following errors right on the first compile: cc: ERROR File = /usr/include/sys/bsd_types.h, Line = 31 The declaration contains a duplicate specifier. typedef unsigned short ushort; This seems to happen because /usr/include/sys/bsd_types.h defines: #define ushort unsighed short perhaps I'm missing a build option. Richard -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Fonts corrected--read to update
The fonts should now all be binary marked so we shouldn't run into problems between platforms. Make sure to update your cvs tree with the -A flag; for example cvs update -A dx This will ensure that you get the binary sticky bit set in your local cvs tree. David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] CVS - fonts (Greg, Pete, others dealing with CVS)
Thanks Mike, I see what you're saying. If I do a cvs status dx/fonts/area.dx, I see that the Sticky Options are set to -ko. And from what I read, -ko will convert the line feeds from platform to platform (this is really bad for binaries which may have \n in the binary portion and get converted to \n\r when checked out on NT or vice-versa). So yes, it seems the correct thing to do is change the options to -kb on all binary dx files. So Greg, I was looking through the CVS documentation and they have an example how to fix this and it follows below: EXAMPLE: - # Add the file so its like our font files area.dx, etc. $ echo '$Id$' kotest $ cvs add -mA test file kotest $ cvs ci -mFirst checkin; contains a keyword kotest # Now since the file kotest is already in CVS, it can be changed by $ cvs admin -kb kotest $ cvs update -A kotest # Copy in a good copy of the file from outside CVS $ cvs commit -m make it binary kotest -- Since I believe we don't have admin access to the cvs, could you perform this? Pete, what about the dxsamples tree? Shouldn't this be done as well for some of the binary files? David For my binary data under CVS I use the 'kb' flag, as in: cvs add -kb ... cvs import -I ! -W * -k 'b' ... The fonts are using 'ko' (from the sticky options). I would think 'kb' is what we want for any binary data. Mike -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] import_hdf.c???
Greg, As Suhaib noted the other day, you changed the import_hdf.c files with the #include dfsd.h to #include hdf/dfsd.h. Why is this? Do you want to document why and how we should compile hdf, etc.? From what I can see in a normal installation of HDF it would not put the include files in a hdf directory. David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Fonts??
I noticed the other day that the fonts were updated in CVS. Well, now they don't work--at least on my Unix box. What were the changes for and what is the fix. The error trying to import variable is: Invalid data: error 0 reading binary data; 3 items successfully read (33 expected)/array staring at data offset 1280/file 'usr/local/dx/fonts/variable.dx' line 82. David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Fonts??
Does this happen often with binary files in CVS? The could cause problems with the dxsamples in the future (possibly). David Suhaib M. Siddiqi wrote: Invalid data: error 0 reading binary data; 3 items successfully read (33 expected)/array staring at data offset 1280/file 'usr/local/dx/fonts/variable.dx' line 82. Oh this problem I had too, but soon discovered the files somehow were corrupted during CVS download. Doing a CVS checkout in a fresh directory Is it possible that CVS tries to 'patch' the binary files upon 'cvs update' thereby introducing nonsense characters? Christian -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
RE: [opendx-dev] Fonts??
Are you sure this works. I just tried several different ways to re-checkout the fonts (into a new folder, delete fonts from my dx and re-checkout, etc.) and they still are showing up unusable. If I do a diff on them from the ones in 4.1.0, there is definitely some differences in the binary data. Greg, the history shows that you changed these on 5/16--if I do a cvs diff between the two versions, there is a difference in the binary. Could you please let us in on what changed (is it an error?) David I noticed the other day that the fonts were updated in CVS. Well, now they don't work--at least on my Unix box. What were the changes for and what is the fix. The error trying to import variable is: Invalid data: error 0 reading binary data; 3 items successfully read (33 expected)/array staring at data offset 1280/file 'usr/local/dx/fonts/variable.dx' line 82. Oh this problem I had too, but soon discovered the files somehow were corrupted during CVS download. Doing a CVS checkout in a fresh directory mkdir temp cd temp cvs -z9 :prser. co dx/fonts fixed the problem. David, I got hold of X-Win32 5.0.2 for Windows. The OpenDX binaries for Windows works without problems with X-Win32. I am talking about Windows2000 and WindowsNT. Did not get time to check Win9x. Suhaib David -- .. ... David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530 -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Sun Solaris 8 Binary...
Let me see if I understand your problem. You have gotten the source to compile and you installed your compiled version. Then you type dx -uionly and dx -execonly . Connect the ui to the exec and it works. But if you type dx -edit, it never connects to the exec. How about if you just do dx -uionly, go to the connect menu and then try and start a dx server? What is the error message? Are you willing to try and debug this with a debugger? David I tried the Solaris 2.6 binary on Solaris 8 on Ultra10 and 5 without success. Then, I got and compiled the source. After some fiddling I got it to compile. However, when I run the program, the dxexec server cannot be started. The only way to get the program running is to start dxexec first, then start the gui and then connect to the existing server. so far, nobody on the list could help me figure out what the problem here is. I would sure like to get this to run without having to start the server manually. Michael Boyles wrote: I downloaded the binary that was compiled for Sun Solaris 2.6. But, I am running Solaris 8 and am having problems getting the server to start. The error message is as follows: Starting DX executive ... cannot get shared memory segment / failed on segment 1 of 1 needed segments / last call to shmget returned errno 22 cannot initialize DX library. Any suggestions? Someone mentioned that Sun might have changed the shared memory calls for Solaris 8. Anyone developing for Solaris 8? Thanks, - Mike - | Michael Boyles - Lead Research Programmer | [EMAIL PROTECTED] - Advanced Visualization Lab | www.avl.iu.edu - Indiana University Content-Type: text/x-vcard; charset=us-ascii; name=sucher.vcf Content-Transfer-Encoding: 7bit Content-Description: Card for Nikolaus J. Sucher Content-Disposition: attachment; filename=sucher.vcf Attachment converted: Powder1:sucher.vcf (TEXT/R*ch) (00015DCB) -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Re: dxexec: 20% perf speedup
I would definetly be willing to commit something as long as I can verify it on a few systems first. David Randall Hopper: | From profiling dxexec, I've found that 20-25% of it's time in the |single-process case is spent locking and unlocking mutex locks (on SGI). |AFAICT these locks add unnecessary, avoidable overhead to the |single-process case. | | I can cook a patch so that mutex locks are only used in dxexec when |nprocs 1. But before I do, I'd like to find out whether the DX |committers would be willing to commit this patch. Perhaps worth mentioning that I already have a patch and the 20-25% savings is real savings, not estimated from prof numbers. However the patch needs refined to be ready for a commit. -- Randall Hopper [EMAIL PROTECTED] -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] ADMIN: CVS commits on opendx-dev list???
This is the way it is set up. David CVS commit log messages are being sent on the opendx-dev list. Its this supposed to be happening or can it be fixed? Isn't there a separate list for CVS-commits? Thanks Cliff -- VIM is my GUI -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] JNI headers path
The problem is, that not all jni headers are in the same place for the jdk. On Suns, Sun automatically installs this in /usr/include/... not in the jdk. It does try to find the jni headers within your jdk if it is not specified. If you are getting that it can't find it, then it is not in a default area that we are checking. If you let me know where your jni headers are, we can determine whether checking this location should be added. Pete's code is still being used--it was added back in around 4.0.9. So it is not possible to find the jni headers which are required for compiling some of the javadx code. David Why do we have to check for JNI Headers. When JDK is installed, the JNI headers are definitely present in jdk/include. Checking the JNI headers extra seems to me adding uncessary comand to configure like ./conifgure --with-jni--header-path=balh --with-jdk--path=blah --this-is-year-2000=blah and blah blah add 100 more here checking for javah... javah checking for java... /usr/bin/java checking for valid jni headers path... no configure: warning: JavaDX will be skipped during compilation due to limitations. checking for dirent.h that defines DIR... yes This seems to me idiosyncratic way of supplying 100 commands to ./configure. In the begining when Pete added JAVADX it was much simplere, that to check for the presence of only JDK. I syggest these extra stuff should be removed, it maked life more compilcated. Thanks Suhaib -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] JNI headers path
This is not the case. The jni headers are used with gcc to compile not javac. Thus the -I flag needs to have the paths set for the jni headers. So we are searching for them to include them in the compile lines. David David, I had JDK in /usr/local/jdk1.2.2 installed. I tried 10 times to specify the JNI headers path, also tried made symbolic links from /usr/include... The configure will keep complaining no valid JNI headers in PATH. The only way to get it working, is to edit the configure.in and through out all the JNI header checking stuff, then run CVSMake to regenrate configure script which does not check JNI headers and needed only to know the PATH of JDK directory. In my opinion telling PATH of JDK to configure script should be sufficient instead of telling him where my javac, is installed, where my classesa re installed and where my jni headers are installed. Once it knows where JAVA compiler is installed then it is java compiler (javac) job to find the headers, not configure script. It make life compilcated for no good reasons. Regards Suhaib David L. Thompson wrote: The problem is, that not all jni headers are in the same place for the jdk. On Suns, Sun automatically installs this in /usr/include/... not in the jdk. It does try to find the jni headers within your jdk if it is not specified. If you are getting that it can't find it, then it is not in a default area that we are checking. If you let me know where your jni headers are, we can determine whether checking this location should be added. Pete's code is still being used--it was added back in around 4.0.9. So it is not possible to find the jni headers which are required for compiling some of the javadx code. David Why do we have to check for JNI Headers. When JDK is installed, the JNI headers are definitely present in jdk/include. Checking the JNI headers extra seems to me adding uncessary comand to configure like ./conifgure --with-jni--header-path=balh --with-jdk--path=blah --this-is-year-2000=blah and blah blah add 100 more here checking for javah... javah checking for java... /usr/bin/java checking for valid jni headers path... no configure: warning: JavaDX will be skipped during compilation due to limitations. checking for dirent.h that defines DIR... yes This seems to me idiosyncratic way of supplying 100 commands to ./configure. In the begining when Pete added JAVADX it was much simplere, that to check for the presence of only JDK. I syggest these extra stuff should be removed, it maked life more compilcated. Thanks Suhaib -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530 -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
RE: [opendx-dev] rpm spec file
I agree. I think it is fair to just put the MSVC files directly in the cvs tree. Since they are not going to affect anything else. David You may add MSVC projects files in CVS. Maybe under Windows directory? Best would be to add these file in the respective driectory and the dx.dsw in dx (root). Sicne these are text files they should not affect other platforms. I would prefer MSVC project files. Once users start using your MSVC wrappers, I am affraid mailing list might get flooded with questions. Suhaib -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of [EMAIL PROTECTED] Sent: Saturday, May 20, 2000 9:22 PM To: opendx2-dev@lists.berlios.de Subject: Re: [opendx-dev] rpm spec file Sounds reasonable to me. What do y'all think about MSVC project files? It requires a file in the root and separate files in each C/C++ source leaf. Isn't there a way to handle it as a separate module in CVS. Should they be distributed in the source tarball, or some other way? Greg -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Solaris 8 compiled (at last), dxexec problem
Did you try connecting to an already running server started with dx -execonly? If you did, what was the error message? David Hi there, I got dx4,1 compiled on Solaris 8 for sparc after renaming queue in qmessage.c to queuex and adding isystem/usr/openwin/include as x-include path. When | run dx, I cannot connect to the server. I can start dxexec by itslelf . Can someone point me to a possible solution or even where to check for problems??? When i start ds -uionly and then try to connect to the server it always fails. I tried localhost or my hostname to no avail. Wher are the possible options described??? I really want this to get to work. Nikolaus -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] DX requires GNU make
My feeling is that in order to use JavaDX, gnumake is a requirement. There are a few pieces of the compile process for the Java code that will not work unless there is a way to export symbols. I don't know what others feelings are. David What is the policy on generated makefiles working with the system make? Should they, or should GNU make be used? The DX README indicates the former. Building on Solaris 2.6, the build runs well with Sun make until it dips into dx-4.1.0/src/uipp/java. This Makefile fails with Sun make and succeeds with GNU make. Randall -- CAUSE OF ERROR: src/uipp/java/Makefile:446 -- export CLASSPATH -- BUILD OUTPUT - SUN MAKE -- make Making all in src Making all in exec Making all in libdx Making all in dxmods Making all in hwrender ... Making all in java make: Fatal error in reader: Makefile, line 446: Unexpected end of line seen Current working directory /home/rhh/t/dx-4.1.0/src/uipp/java *** Error code 1 make: Fatal error: Command failed for target `all-recursive' Current working directory /home/rhh/t/dx-4.1.0/src/uipp *** Error code 1 make: Fatal error: Command failed for target `all-recursive' Current working directory /home/rhh/t/dx-4.1.0/src *** Error code 1 make: Fatal error: Command failed for target `all-recursive' -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] How'd VDE 2000 go?
I was wondering if anybody wants to give a recap of how VDE2000 went for those of us that couldn't be there. David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Lesstif problem.
Rick, et. al., I've been keeping current with the cvs version of Lesstif and now that 0.90.0 is out, I just wanted to make a comment. There is a problem with the XmText with the Annotation capabilities in OpenDX. It seems that the first line of any annotation tends to disappear and a XtWarning: XmStringGetNextComponent: unknown type error is listed. For example, create an annotation on the canvas, save the network, and then re-open the file. The first line is gone. This can also be shown by creating an annotation, then drag copy it. The first line will also disappear. Any ideas? David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Lesstif problem.
Sorry, You select Annotation from the Edit menu then click on the canvas. This allows you to annotate on the canvas. It will show up okay the first time. Double click it to edit it. Type something in (2 lines or more). Then try to drag copy it (with the middle button), or save then re-open the network. David I'm going to need more detail. If I start from dx -edit, what next?? I can select Annotation, and then one of the tools, but this simply results in one of the little boxes with the tabs. I don't seem to be able to make any XmText's appear. I feel like one of those helpless newbies :) On 16-Apr-00 at 21:17, David L. Thompson ([EMAIL PROTECTED]) wrote: Rick, et. al., I've been keeping current with the cvs version of Lesstif and now that 0.90.0 is out, I just wanted to make a comment. There is a problem with the XmText with the Annotation capabilities in OpenDX. It seems that the first line of any annotation tends to disappear and a XtWarning: XmStringGetNextComponent: unknown type error is listed. For example, create an annotation on the canvas, save the network, and then re-open the file. The first line is gone. This can also be shown by creating an annotation, then drag copy it. The first line will also disappear. Any ideas? David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530 -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Anybody know how to get a cvs log between dates
I would like to get a full log of all changes to cvs between 4.0.6 and 4.1.0 on the dx branch. I have the dates between the two so I tried cvs log -d 1999-07-072000-04-15 However, this will only log the current part of the tree I'm in and none of the subdirectories. If I don't put the date flag on, then it recursively provides all information on the entire subtree as expected. Anybody? David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] 4.1.0
That isn't until the 27th-28th. David Unless I'm mistaken, the conference in NY is going down about now On 12-Apr-00 at 18:10, Suhaib Siddiqi ([EMAIL PROTECTED]) wrote: There seems to be no activity on OpenDX list. Is 4.1.0 rock solid for all the users? Is not time now to check in the new patches? Any news about Win32 patches from Greg? Thanks Suhaib -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] stripping libraries?
What is the effect of stripping libraries? Are they still usable and only the debugging info is removed? This could really help cutting down the size of the distributions, I've just never done it before and didn't know it would do. David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Can we roll a release of the samples?
Is it possible to go ahead and bump the samples to 4.1.0 and roll a tar ball of it as well? I'd like to see these stay in sync for releases. David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Which version number to use
I question this patch before the bump. We've seen stable builds with what we have now. I started to perform a lot of the same stuff they did at Redhat, and found that at times things destabilized. I'd rather get 4.1.0 out with what we have, get the readmes, etc that we're supposed to add and then do the bump. Past that, then apply all these patches and start working on an odd number revision again. I have a lot more to add than just the changing the return NULLs to return ERRORs. Basically they just cleaned up some minor compiler errors. David A couple of weeks ago I asked for tests of a large patch submitted from redhat. We heard back from Jeff with his partial OK, and that's it. I think these patches should go in and we should fix obvious breakage prior to 4.1 . Pete [EMAIL PROTECTED] wrote: I'm in. Any objections? BTW. I have a version that uses MS tools under the standard gnu build structure ready for checkin. In addition to making lots of code changes for MS, I did wrappers for the MS compilers etc. that make them compatible with the gnu tools, and have MSVC6 projects for everything. I think its a necessary step before checking in the true Windows version of the exec that doesn't require Exceed and supports the ActiveDX component stuff. So. Are we go for 4.1.0? Greg Suhaib M. Siddiqi [EMAIL PROTECTED]@opendx.watson.ibm.com on 03/14/2000 07:43:30 AM Please respond to opendx2-dev@lists.berlios.de Sent by: [EMAIL PROTECTED] To: opendx2-dev@lists.berlios.de cc: Subject: [opendx-dev] Which version number to use Are we going to declare 4.0.10 are 4.1.0? I am trying to get OpenDx binaries compiled by using Exceed XDK 6.2 and MSVC 6.0 out of door. MSVC compilation requires a lot of manual editings, if we are set on declaring 4.0.10 as 4.1.0 then I might bump the version number now, instead of redoing it again. After the release of Windows 2000 Microsoft SDK has a lot of new headers and it required a good amount of OpenDx 4.0.10 source patching. M$ is about to release MSVC 7.0 which would have a lot of Win64 releated library and header changes again. We would need to do a lot of patching again after MSVC 7.0 is released. It may make binaries unstable at that time. I prefer to see the DX version bumped now before we get into another cycle of unstable binaries due to changes in compilers and libraries. Suhaib -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] 4.0.10 - 4.1.0
I did a check on Solaris 2.6 (sparc), Solaris 2.7 (intel), LinuxPPC, AIX 4.1 (xlc) and all worked fine. This is with today's cvs. David I agree, too. OK, I'd like to suggest that we do a little sanity checking just to make sure we are not missing anything obvious. It looks like Jiten's nightly builds are running smoothly on DEC 4.0, AIX 4.3.2 w/ xlc, and Linux/RH6.1. I'll run a clean build on SGI/IRIX6.2, and run a couple tests on each of these. Sounds like Suhaib has cygwin under control. Can anyone else report on a recent clean cvs build? Anyone willing to build and run on other platforms over the next couple days? I'd be happy if we could just run a couple samples on as wide an array of platforms as we can, but let's set a limit so this doesn't drag on too much. Friday? Greg Suhaib M. Siddiqi [EMAIL PROTECTED]@opendx.watson.ibm.com on 03/14/2000 12:56:34 PM Please respond to opendx2-dev@lists.berlios.de Sent by: [EMAIL PROTECTED] To: opendx2-dev@lists.berlios.de cc: Subject: RE: [opendx-dev] Which version number to use I abolsutely agree with David. I think it is better to declare 4.0.10 from yesterday CVS as 4.1.0. It is stable, then introduce other changes like MSVC and the patch from RedHat and work to fix new bugs. Suhaib -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of David L. Thompson Sent: Tuesday, March 14, 2000 12:41 PM To: opendx2-dev@lists.berlios.de Subject: Re: [opendx-dev] Which version number to use I question this patch before the bump. We've seen stable builds with what we have now. I started to perform a lot of the same stuff they did at Redhat, and found that at times things destabilized. I'd rather get 4.1.0 out with what we have, get the readmes, etc that we're supposed to add and then do the bump. Past that, then apply all these patches and start working on an odd number revision again. I have a lot more to add than just the changing the return NULLs to return ERRORs. Basically they just cleaned up some minor compiler errors. David A couple of weeks ago I asked for tests of a large patch submitted from redhat. We heard back from Jeff with his partial OK, and that's it. I think these patches should go in and we should fix obvious breakage prior to 4.1 . Pete [EMAIL PROTECTED] wrote: I'm in. Any objections? BTW. I have a version that uses MS tools under the standard gnu build structure ready for checkin. In addition to making lots of code changes for MS, I did wrappers for the MS compilers etc. that make them compatible with the gnu tools, and have MSVC6 projects for everything. I think its a necessary step before checking in the true Windows version of the exec that doesn't require Exceed and supports the ActiveDX component stuff. So. Are we go for 4.1.0? Greg Suhaib M. Siddiqi [EMAIL PROTECTED]@opendx.watson.ibm.com on 03/14/2000 07:43:30 AM Please respond to opendx2-dev@lists.berlios.de Sent by: [EMAIL PROTECTED] To: opendx2-dev@lists.berlios.de cc: Subject: [opendx-dev] Which version number to use Are we going to declare 4.0.10 are 4.1.0? I am trying to get OpenDx binaries compiled by using Exceed XDK 6.2 and MSVC 6.0 out of door. MSVC compilation requires a lot of manual editings, if we are set on declaring 4.0.10 as 4.1.0 then I might bump the version number now, instead of redoing it again. After the release of Windows 2000 Microsoft SDK has a lot of new headers and it required a good amount of OpenDx 4.0.10 source patching. M$ is about to release MSVC 7.0 which would have a lot of Win64 releated library and header changes again. We would need to do a lot of patching again after MSVC 7.0 is released. It may make binaries unstable at that time. I prefer to see the DX version bumped now before we get into another cycle of unstable binaries due to changes in compilers and libraries. Suhaib -- .. ... David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530 -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] newly updated tree, can't run dx anymore
I was wondering the same thing? David Will this also be a problem for the rest of us??? I, for one, am not subscribed with the same address as my commits come from. BTW, I'm compiling from scratch as I type this..stand by Mail from the commit won't be forthcoming (pdk wasn't a subscriber to opendx-dev so our recent spam filter intercepted it). -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] newly updated tree, can't run dx anymore
Do you have an directory named DXROOT/samples/macros with nothing in it? I know what the error is, is the command is getting -macros with nothing following it. Why I don't know. David For the first time in a long time I cvs up'ed my entire tree then did a make install. To my surprize I now get $ dx -edit /opt/dx/dx/bin/dxworker : -macros: missing directory list (use -help to get usage information) $ $ dx -help /opt/dx/dx/bin/dxworker : -macros: missing directory list (use -help to get usage information) $ Have I missed something important in the commits -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] When to start submitting a large number of fixes?
I know things are trying to solidify for 4.0.10, so can we start a new branch in the CVS? I'm ready to start submitting compiler warning fixes, etc. This would be a whole bunch of files. Thanks, David -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] [PATCH] Makefiles, warnings
These are some of the same things that I had started to do, but ran into problems when I fixed them. We have to be careful when qualifying them to be correct. There are lots of different operating systems and different compilers involved here. I agree a lot of these are straight forward such as return ERROR instead of return NULL, but after I went through and cleaned up just one file (memory.c), the number of core dumps I got increased by ten fold. Your suggestion on builddir!=srcdir has been already submitted and I believe Pete is working on this. David Against latest CVS, http://people.redhat.com/sopwith/opendx-misc.patch If I'm supposed to be sending this somewhere else, please let me know. What the patch does: . I build with builddir != srcdir as much as possible, so the Makefile.am's and some other files needed changing to work with this. . On the way to finding a bug (that turned out not to be in opendx anyways :), I fixed tons of warnings. Hope this helps, -- Elliot Red Hat, a Red Hat company -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Some build problems and fixes
Pete, These were my feelings as well. I would volunteer to apply the fixes, but I just don't have time to do it right at the moment. David Thanks for the submission. I like i iii and can input the changes (unless there is another volunteer). I have some misgivings about ii --adding a gl-specific customization for libs and includes. Do you object to CFLAGS, CPPFLAGS and LDFLAGS for this purpose (and which also accommodate many other customizations)? Regards, Pete Jose Gomes wrote: Hello, I have dx built, installed and running for linux and solaris and would like to report 3 problems. The fixes can be found in the file in attachment (which is the output of cvs diff -C3). Please let me know if something is wrong with this. I hope this will help. Best regards, Jose Gomes --- Following are the three bugs description, labeled i), ii) and iii). i) The main problem is that the Makefile.am's are not written to support both builddir==srcdir and builtdir!=srcdir which was required for me since I wanted to compile it for two architectures and which is (as far as I know) the standard way to use the gnu tools. So I had to change in almost every Makefile.am's lines like: INCLUDES = -I../../../include [EMAIL PROTECTED]@ by lines like: INCLUDES = -I$(srcdir)/../../../include [EMAIL PROTECTED]@ Another modification is due to the fact that c++ sources are generated automatically by a script called class which asssumed that srcdir==builddir. So this script is slightly modified so that it can be passed the srcdir directory as argument. See below for a list of the modified files. ii) My opengl headers and libraries are not is a standard place. I added the options --with-gl-includes and --with-gl-libs to the configure script so that these locations can be passed to all Makefile.am through the variables GL_INCLUDES and GL_LIBS. See bellow for a list of modified files. iii) I have flex installed for solaris and the scripts think that linux implies flex and solaris implies lex. The problem is that flex and lex are not compatible. The file src/uipp/dxuilib/Network.C uses the yylineno variable which does not exists in flex but does in lex. This was taken into account in this C file by: #if defined(linux) || defined(cygwin) || defined(freebsd) int yylineno; #else extern int yylineno;/* lexer line number */ #endif I think it is not the correct test since the fact yylineno exists is just a matter of lex or flex. So, instead of adding the test || defined(solaris), I hadded the autoconf variable HAVE_YYLINENO (initialized by configure) and replace the previous code by: #ifndef HAVE_YYLINENO int yylineno; #else extern int yylineno;/* lexer line number */ #endif See below for a list of modified files. Following are the list of modified files for fixing the 3 bugs. builddir==srcdir --- M acconfig.h M configure.in M src/exec/dpexec/Makefile.am M src/exec/dpexec/local.mk M src/exec/dxexec/Makefile.am M src/exec/dxmods/Makefile.am M src/exec/hwrender/Makefile.am M src/exec/hwrender/opengl/Makefile.am M src/exec/libdx/Makefile.am M src/exec/libdx/class M src/exec/libdx/local.mk M src/misc/Makefile.am M src/uipp/base/Makefile.am M src/uipp/dxl/Makefile.am M src/uipp/dxui/Makefile.am M src/uipp/dxuilib/Makefile.am M src/uipp/dxuilib/Network.C M src/uipp/java/Makefile.am M src/uipp/mb/Makefile.am Non standard gl location. - M configure.in M src/exec/dxexec/Makefile.am M src/exec/hwrender/Makefile.am M src/exec/hwrender/opengl/Makefile.am No yylineno in flex. M acconfig.h M configure.in M src/uipp/dxuilib/Network.C -- [EMAIL PROTECTED] Tel. +33 4 92 38 76 48 http://www.inria.fr/robotvis/personnel/jgomes Projet RobotVis, INRIA, 2004, route des Lucioles - B.P. 93 F-06902 Sophia Antipolis Cedex, FRANCE -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] multiple image difficulties
Randy, I do have one more question here. Is this with a compile done with gcc? The latest compiles I've been using xlc and xlC and noticed that a lot of my problems are gone. David One additional comment: I tried running dx on one machine, and displaying on another. Disconnect from server at the same line 4775 in displayx.c with cmapsize=i=256. On Mon, 14 Feb 2000, David L. Thompson wrote: A couple of questions. 1) Are you on an 8 bit display system? Which version are you compiling (cvs current)? David Hello- First, before I forget, thanks Pete and Daivd for you tips for getting the java part to compile. I changed the line in /src/uipp/java/Makefile from cc=gcc to cc=xlc an everything build OK. I changed the jni_mdh.h file, and next build I'll tell you how it goes. Next, I am having difficulties displaying multiple images within the same net for AIX 4.2. For an example, I chose AutoColor2d from the tutorial. It runs just fine. If I add a second image module, it disconnects from server. The output from dbx is below. -Randy Segmentation fault in getOneMapTranslation at line 4775 in file /u/res3/rpolson /OpenDX/dx-4.0.10/dx/src/exec/libdx//displayx.c 4775 for (i = 0; i cmapsize; i++) (dbx) where getOneMapTranslation(dpy = 0x2010c328, d = 0x206126b0, force = 0), line 4775 in displayx.c createTranslation(couldn't resolve reference (dbx) (dbx) print cmapsize 256 (dbx) print i 256 (dbx) -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530 -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] javadx and AIX4.2
Pete, I can tell you why I don't move to 4.3.2. For us at the U, we purchased 4.2 off of the long standing agreement with IBM a few years ago. Then IBM cancelled that agreement and created a new one that cost $15,000 to join (as a U). Several departments then decided to pay for this and we weren't one of them. For us to get in with them now and get a valid license, it will cost us $600.00 for one license. How important is it? 4.2 will have to do--unless we could get a free license with media, etc. David Randy- There's a *space* following the continuation-backslash on the #define that starts just above the line that offends the gcc compiler in jni_md.h . remove it and you are home free. For my jdk1.1.8 it is line 91, wonder why it is 89 for you!?!?! Regards, Pete PS why don't you move to aix 4.3.2? -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] multiple image difficulties
A couple of questions. 1) Are you on an 8 bit display system? Which version are you compiling (cvs current)? David Hello- First, before I forget, thanks Pete and Daivd for you tips for getting the java part to compile. I changed the line in /src/uipp/java/Makefile from cc=gcc to cc=xlc an everything build OK. I changed the jni_mdh.h file, and next build I'll tell you how it goes. Next, I am having difficulties displaying multiple images within the same net for AIX 4.2. For an example, I chose AutoColor2d from the tutorial. It runs just fine. If I add a second image module, it disconnects from server. The output from dbx is below. -Randy Segmentation fault in getOneMapTranslation at line 4775 in file /u/res3/rpolson /OpenDX/dx-4.0.10/dx/src/exec/libdx//displayx.c 4775 for (i = 0; i cmapsize; i++) (dbx) where getOneMapTranslation(dpy = 0x2010c328, d = 0x206126b0, force = 0), line 4775 in displayx.c createTranslation(couldn't resolve reference (dbx) (dbx) print cmapsize 256 (dbx) print i 256 (dbx) -- . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] javadx shareable libraries
This is some of the stuff that libtool takes care of. It can take and put the appropriate information into the header of the library and binaries for linkages. In fact, this is how I noticed that the netcdf and other libraries need to have shareable versions installed when using javadx (libtool told me so.). If nothing else, I may add the libtool functionality just for the javadx branch of the tree. But I'd rather add full functionality right off the bat. One of the things that I'd like to have available while adding this functionality is a test set that I could build for checking validity of libs and executables. Do we have a module that would test a fair amount of the dxlibs when linked in as an inboard, outboard, and loadable and executed from a script? I already have the autoconf stuff set up for most platforms for a loadable and to add the functionality for outboard and inboard shouldn't take too much. Something that would require DXlite and libDX might be good. (On that thread, is there a chance that libDX could be broken into two libraries? DXlite and OtherDX? If they were shareable, then libDX wouldn't have to be so big.) Can someone send me a good dx-link program that could test the DXL library? (Why does the shareable code with javadx only link to a subset of DXL?) Opinions, suggestions? David Well, thanks, but again, LD_LIBRARY_PATH isn't working for me on aix 4.3.2. When it seems to work it is only a coincidence, the load succeeds when the libMagick.so shared lib is fortuitously in one of the directories in dxexec's header. Pete . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Generating export symbols from binary in AIX
All you AIX persons in the know--or who may know somebody. Is there anyway to generate an export symbol table file in AIX from an executable (binary)? The libtools guys currently do not support this but would love to add it if someone could provide them with the code. Thanks, David . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Sun fun
Okay, After 5 hours of uploading, I've put up a new version for Solaris. This is 4.0.10 and is quite huge 27.7MB due to the fact that it has all the bells and whistles (full ImageMagick support). David I'm working with a new user to DX who is having trouble getting things going. He's tried downloading both the available Solaris open source version (I think that was about 4.0.5?) and the 3.1.4b. He had trouble getting the OpenDX version to run at all, and now is having trouble running a net under 3.1.4b which runs quite successfully on SGI Extreme and Indy at two other locations, both using 3.1.4b. His machine is: Sun Ultra Enterprise 2, dual processor, SunOS 5.6/Solaris 2.6 512 MB memory I guess what I'm wondering is how soon one of you who has built the new 4.0.10 version for Sun might be posting it as a binary? This would be a good opportunity for yet another tester with a real-world problem to run on it. Thanks! I know nothing about Sun's and their various OSes. Does it seem likely that this is a later version than would reasonably likely run 3.1.4b? He can run some samples but not this program: yes, I developed it, but no, there are no custom modules or anything weird in it to speak of. All help appreciated. Chris Pelkie Vice President/Scientific Visualization Producer Conceptual Reality Presentations, Inc. 30 West Meadow Drive Ithaca, NY 14850 [EMAIL PROTECTED] (607) 257-8335 or (607) 254-8794 . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Re: LessTif Problems (bugs)
Rick, Thanks for all your help. David David, I think all of the problems that you mention in ftp://ftp.cs.umt.edu/pub/user/dthompsn/lesstif-probs/lesstif_probs.html have now been addressed, either with changes to LessTif or DX. The vertical/horizontal/vertical layout problem, which seemed pretty simple on the surface actually triggered a massive re-write of the Form code. Things are never what they seem. I have an annotated copy of your lesstif_probs.html file if you would like. I tried to upload it to where I got it from, but like I suspected couldn't. I'm currently cleaning up some work on the Options-Grid... dialog, it behaves _very_ badly. It will also result in some Form changes, but in the process I noticed that this thing was put together very badly (IMO), conflicting constraints and so on, so I re-built it. After I commit the changes someone had better test this out with Motif to make sure it still works. The problems show up while switching amonst the different grid types. Don't just click through the list in order once, different things happen depending on your starting and ending points. The GridDialog update will close out my list of known problems with DX and LessTif. Are there any more that I don't know about.Checking JitterBug shows only 1 outstanding with lesstif mentioned, bug 18, which I think has been addressed many times as a PATH related problem. Maybe now I'll sit down and actually try to use this thing :) . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Bugs-Fixed in Bug Database
Big Oops, I'm still linked to the old gnats bug database. I'll get this fixed and then re-look these over. David For the benevolent dictator. I just went through the bug list and believe the following bugs have been fixed over the last couple of weeks: Numbers: 11, 12, 13, 16, 17, 20, 21, 22, 23, 24, and 26. I will be going back through the mailing lists looking for others that haven't been fixed and get them added. But thought we should remove ones that have been fixed. Some of the others may also have been fixed, but I can't attest to them. David . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530 . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Bugs-Fixed in Bug Database
For the benevolent dictator. I just went through the bug list and believe the following bugs have been fixed over the last couple of weeks: Numbers: 11, 12, 13, 16, 17, 20, 21, 22, 23, 24, and 26. I will be going back through the mailing lists looking for others that haven't been fixed and get them added. But thought we should remove ones that have been fixed. Some of the others may also have been fixed, but I can't attest to them. David . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] new configure comments
Randy, If the configure can't figure out where the java is by doing the tests, you will HAVE to give both directories as you state. There is just no way around it. Your java compile problem is the same as what I had on AIX. This is actually the java compiler dying with a native threads crash. There isn't much we can do against a compiler crashing. I have compiled this (once) with gcj (the gnu java compiler) on AIX but it has been a while and I don't know if it will still work. Even if it did, I'm not sure it would run with the aix buggy java. David Hello. As downloaded, the Java headers jni.h and jni_md.h are in different directories. The to configure I had to type them both: --with-jni-path=/J1.1.8/include:/J1.1.8/include/aix. The make in /src/uipp/java still bombs for me. I get the error: (javac -d dxsc DXServer.java) awk: newline in character class [/ ]*$... input record number 2, file source line number 1 SIGSEGV received at d06db090 in /u/res3/rpolson/OpenDX/J1.1.8/lib/aix/native_thr eads/libjava.a. Processing terminated Writing stack trace to javacore.txt ... OK /u/res3/rpolson/OpenDX/J1.1.8/bin/aix/native_threads/javac[15]: 12564 Abort(core dump) make[1]: *** [server.jar] Error 134 make[1]: Leaving directory `/OpenDX/dx-4.0.9cvs/dx/src/uipp/java' make: *** [all-recursive] Error 1 Old HDF consideration: as a newbie I assumed that the latest version HDF5 was what the standard should be. The configure checks for hdf.h, the HDF5 has files HDF5.h and work differently. I personally don't care, except it confused me. . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Imagemagick notes
This may be the cause. Try adding the extension you want it to result as imagemagick2.tiff. It may be the case that we need to check for no extension. David I used the name imagemagick2 with no extension. On Wed, 8 Dec 1999, David L. Thompson wrote: What filename are you using when you choose ImageMagick supported format? This is how ImageMagick knows what to save the file as. Pete, If you don't add an extension, does ImageMagick have a default that it uses? When I go to save an image I can chose ImageMagick supported format. When I save the image I get an error message: Magick: no encode delegate for this image format(). Segmentation fault (core dumped) Two files are created : 'filename.' and 'filename.miff'. 'filename' has length 0, and 'filename.miff' has the image which can be brought up under imagemagick display. -Randy . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] A new tarball / revision?
I have also found that for the link for JavaDX to work, the flags for the dynamic load items on Solaris needs the flags -dy -G. So right now the compile cannot complete on Solaris without editting the Makefile in src/uipp/java. For now, I'm just editing this, since I don't want to spend too much time replicating work that could be done by libtool. I'm going to try and get to that sometime. David People have been asking me when they can get a stable version that includes JavaDX. I think it'd be a good idea to put out a new tarball sometime soon that incorporates bug fixes on top of 4.0.9, test it, and then anoint it golden at 4.1.0. What I'd like to do is to get everyone to check in the fixes they are sitting on, then freeze CVS and label it as 4.0.10. We'd then roll a new tarball that people could test in concert. If any serious problems show up, we'd very carefully fix them in the CVS and, when it seems to stabilize again, iterate. Then label it as 4.1.0, pronounce it golden, and open the CVS for development again. What do y'all think? Are there major unfixed problems that need dealing with first? If so, I've lost track - how about posting them? I'll check the bugs database, too.How about opinions on timing? By the way, I don't mean to imply actually locking out CVS commits - just ask people to hold off on things that can be held off. Bll tells me I don't misspell enough to be taken seriously as a computer guy. Greg . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530 . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Interactors
Okay, I didn't do any wacky builds. Pretty straight forward (CVSMake followed by configure). I test this, and the arrows are doing what they are supposed to Enter, Arm, Activate 000, Disarm, and Leave. So it looks like it must be in the Callback somewhere (right?) On Friday, I may have some time to devote to trying to get this and the Control Panel problem looked at. David David, I think I have addressed all of the usability concerns from your recent list that I can reproduce. Like I have said in the past, activating the arrow buttons on the Integer Interactor seems fine for me, so I'm going to need a little help with this one. I suspect if we find out what is going on for one of these the rest will fall into place. So here goes. Hopefully you haven't specified any of the wacky build flags for LessTif. I think the one that you must _not_ have specified is --enable-production. If by chance you did, re-configure and build _without_ the --enable-production flag. Okay, start dx with DEBUGSOURCES=ArrowB.c dx -edit Drop down an integer interactor and double click on it You should see a couple of lines of LessTif debug output on stderr. So far there should be a couple of lines indicating a set_values has been done on both of the arrows. Now activate one of the arrow buttons. You should get an indication of well, it's easier to show you XmArrowButton left: EnterWindow XmArrowButton left: LeaveWindow XmArrowButton left: EnterWindow XmArrowButton left: Arm XmArrowButton left: Activate 000 XmArrowButton left: Disarm XmArrowButton left: LeaveWindow The same sort of thing should show up for the right arrow, except it's name will be right not left. If we are getting this far we can move on to the next step. Now this is what you could call remote debugging :) If you've got some time on your hands try running DEBUGSOURCES=all for an indication of how much LessTif code this thing is actually ploughing through. . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] LessTif Problems (bugs)
No, I can't use real Motif on the system with Lesstif. This is on a LinuxPPC system and nobody has produced a purchasable version yet. I also can't justify spending $1000 for the source code since Lesstif is free and available. I'd appreciate anything you can do. Otherwise, I'll have to start digging into the UI myself. I've been learning a lot about the exec code, but this is a big beast to be messing around in. David Good list, should keep me out of trouble for a while :) What type of system are you running the LessTif tests on? Is it the same system that you run the Motif tests on? Maybe an ldd of dxui or something would be interesting. The reason I ask is that I don't see the problems. Okay, I've just started on the list, but I don't see any of the integer interactor problems. From your description they sound like typical XtGetValues with the wrong argument type, but this is pure speculation so far. On 06-Dec-99 at 16:15, David L. Thompson ([EMAIL PROTECTED]) wrote: Rick, et. al. I went through the interactors and a few other interface things and collected problems between Lesstif and Motif. There are a few severe bugs and the rest are cosmetic or inoperable. You can look at my write up at ftp://ftp.cs.umt.edu/pub/user/dthompsn/lesstif-probs/lesstif_probs.ht ml. David .. ... David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530 . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] LessTif Problems (bugs)
I've compiled all of said things myself with gcc 2.95.2. I have libc6, then compiled XFree86 3.3.5, and then LessTif, plus all the gnome and enlightment stuff. David Understood. Pretty much the same reason I got into LessTif. I don't know much about LinuxPPC. Do you have any idea how X, libc, and LessTif were compiled? The interactor problems sound like some of the strange reports we've had when X was compiled with libc5, LessTif with libc6, or X was compiled with gcc 2.7.2 and LessTif with egcs, and many other different combinations. I'm not sure of the exact details, but it seems that mixing and matching compilers/libs is not a good idea. Best if all of the components get compiled with the same tools. It's not always easy to tell, but if the LinuxPPC dist. comes with X, and you haven't upgraded the compilers and you compiled LessTif on this machine, everything is probably matched fairly well. I reproduced a couple of the dumps, so I will start with those. . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
[opendx-dev] Why DX_CHECK_HEADERS
Greg or Peter, I tried to compile the latest CVS on AIX 4.1 and found a little problem. I had to add the AC_CHECK_HEADERS( sys/socket.h ). This header is being checked in DX_CHECK_HEADERS; however when autoheader runs, it doesn't create symbols in dxconfig.h.in for those headers. Since AIX 4.1 needs HAVE_SYS_SOCKET_H defined then I will submit an update with AC_CHECK_HEADERS( sys/socket.h ). Is there going to be any more like this that you can think of? David . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] instance numbers in dxui
How do you turn on the DXDebug instances in the executive? For example, there is code called similar to DXDebug(X,error); David This change causes a strange change of the text in the modules on the VPE. For example Compute suddenly becomes Compute:2, Color becomes Color:1 etc. All modules get a colon plus a (random?) integer. just fyi - the numbers are part of trace mode. they are instance numbers which let you identify exactly which modules are which when there are multiple calls to the same module in a network. the default is not to display them, but if you want/need them you can enable them through one of the menus in the message window. in addition to the extra information on the module labels you get a trace of exactly which modules are executed in order if you watch the message window as you run. nancy [EMAIL PROTECTED] . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Latest gcc
In order to get around this, use cvs and get the latest version. David Hi Just wanted to mention that to be able to compile OpenDX under Debian potato, using gcc/g++ (gcc version 2.95.2 19991109) I had to use the -fpermissive option to be able to compile the C++ files. I don't know alot about autoconf/automake, but maybe someone can do something about it (I just set CXX=g++ -fpermissive). I also ran into a snag in src/ui++/base/Application.h line 55 (OpenDX 4.04, but it hadn't changed in 4.09), which reads friend void main(unsigned int argc, char** argv); but should, as far as I can see be friend int main(unsigned int argc, char** argv); /Samuel Meder . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530
Re: [opendx-dev] Memory Leak clarifications
In the release version of DX, all images are cached internally differently than what you can set with the modules cache pulldown options. What you must do is set an attribute to turn this off. At 400k per execution--I'd bet this is what you are seeing. Even if you run with -cache off, it would still cache the images in Image and Display. You can test to see if this is still the case by placing an Option before your image and set the attribute cache to 0. David In-Reply-To: [EMAIL PROTECTED] Hi! I'm the original poster about the possible memory leak issue, and I just wanted to clarify a few things in response to the other posts on this topic. I hope this message gets posted in the right thread...Problems with internal mail have kept me from receiving the list messages, so I can't actually reply to the thread. In response to Richard's post: When I run DX on a remote machine (SGI Irix 6.5 with OpenDX in the most recent case) through an X window on my local SGI (with setenv DISPLAY) I have noticed that the Xsgi process on my local machine (owned by root) grows and grows with every use of DX until memory is exceeded and very bad things happen (X dies and sometimes the whole machine dies). An interesting problem, to say the least, but I think what I'm seeing is more of an OpenDX memory-management issue than a system-level problem. When OpenDX crashes in this case, it only runs up against the dxexec memory limit (-memory 200) and doesn't free any of that up, so the network won't run. This is not a case of any single image being too large to run in memory, but a problem that compounds with each network execution. It's never crashed X or my machine...just bumped up against OpenDX's internal memory limit. Only takes a disconnect/reconnect server (not reset server) in OpenDX to free up the memory. In response to Ballard's post: I don't have the seed email but was this user trying to import a lengthy file series using the Sequencer? Nope. Just one file, using the sequencer to select data at different time-steps out of the file. In response to Suhaib's post: The OpenDx/Cygwin version Windows 98? That is a known problem. You have right to yell at M$ for selling the gloriously shitty product. Win98 itself has memory leak, which gets horrible when an X-application (X-server) is running because Win98 tries to be selfish and hat to share resources with any other server... With X-server trying to share Display causes Win98/95 to start eating CPU resources and may once in a while will force you to reboot the PC. because it gets horribly slow. Actually, I was able to create the problem under NT as well. The problem is not with CPU resources, but memory usage. It seems like OpenDx isn't releasing memory somewhere, so eventually it fills up its memory limit (-memory 200) and won't release any of it. Nothing effects the system, except the dx network won't run until you disconnect/reconnect the dxexec to force the release of memory. Thanks for the responses! Let's keep at it so we can figure out what's going on here... Jeremy Zoss Southwest Research Institute (210)522-3089 [EMAIL PROTECTED] . David L. Thompson The University of Montana mailto:[EMAIL PROTECTED] Computer Science Department http://www.cs.umt.edu/u/dthompsn Missoula, MT 59812 Work Phone : (406)257-8530