- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
It does seem like a DYLD_LIBRARY_PATH related error.
Maybe the 11.1 entries are just old leftovers. Can you try compiling with -V and confirm the paths reported during the compilation match the expected paths and those in the DYLD_LIBRARY_PATH setting too?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thank you, Kevin
Of course, you are right. So I examined several things. First, I reinsalled composerxe, but the situation does not changed.
Then, I did a few things.
1) At the terminal mode, I typed
%printenv > printenv.txt
The output printenv.txt does contain DYLD_LIBRARY_PATH .
2) I wrote a little shell scrip "test.csh"
#!/bin/csh -f
##!/bin/tcsh
#-f
printenv > printenv_in_csh.txt
The output printenv_in_csh,txt does not have a line of DYLD_LIBRARY_PATH at all.
So this is the reason why my code failed to run.
But, I don't know why this happened. Could you figure it out?
I am a retired professor working at home, so that I do not have any students to help me.
I have just posted this trouble at Apple supporting community already. But I do not have any response yet.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
This is puzzling.
I guess your program is being run from a .csh style script and that is the reason you created the test script?
With -f you would expect underlying shell settings to pass to the new shell but it seems like something in the initialization of the new shell though is clearing or not-initializing the same from your description.
In your test script, would the -V or -X (along with -f) show any output that would be helpful to you?
Despite -f, you might inspect any of your own private .cshrc/.bashrc style files and also any system-wide under /etc/*shrc or /etc/*csh* files.
Maybe another test is to spawn the new shell just using csh at the prompt and then use printenv to see whether that replicates your findings when using the shell test script.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
This appears to be related to the El Capitan System Integrity Protection feature.
In your test.csh file, for the first line, add an extra # to de-activate the line and then run the script and see whether you can see DYLD_LIBRARY_PATH in the output file. If it is present then you would need to make the same sort of change to any script that is running your program.
Here is a relevant thread from the Apple Developer forum, https://forums.developer.apple.com/thread/9233 (see the post by JGD on 6/16/2015 and khoyt on 10/9/2015).

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
I am a user of OSX 10.11.1 (but the following happened while I was using OSX 10.10).
It seems to me since I started to use /opt/intel/composer_xe_2015.3.187/compiler, I have been encountering the following troubles.
When my code run, I have the message
dyld: Library not loaded: libmkl_intel_lp64.dylib
Referenced from: /Users/iwatasue/bin/Molyx/debug/two.x
Reason: image not found
I examined the environment as
KyoRyu:/<3>Molyx/Ne_n%printenv |grep DYLD
65:DYLD_LIBRARY_PATH=/opt/intel/composer_xe_2015.3.187/compiler/lib:/opt/intel/composer_xe_2015.3.187/compiler/lib/intel64:/opt/intel/composer_xe_2015.3.187/compiler/lib:/opt/intel/composer_xe_2015.3.187/mkl/lib:/opt/intel/Compiler/11.1/058/lib:/opt/intel/Compiler/11.1/058/lib/intel64
Obviously the link was correctly set.
Could somebody help me?