jeffreyhorner

You could add that code using the REvalOnStartup or RSourceOnStartup directive. That way every time a new apache process is created it gets set. http://rapache.net/manual.html

What's happening is that the linux service script(unity, init, system, etc.) ought to be amended to update the LD_LIBRARY_PATH (And other side effects of R CMD sh) prior to starting...

Here's a version of the init script that ultimately get's placed into /etc/init.d/apache2 on ubuntu: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/vivid/apache2/vivid/view/head:/debian/apache2.2-common.apache2.init It references a file that's typically located at /etc/apache2/envvars Maybe we need an additional...

You'll see in the /etc/init.d/apache2 script that the environment has to be set up at the invocation of apachectl. The facility for setting up that enviroment seems to be defined...

Something like this would populate the R environment to /etc/apache2/envvars: ``` #!/bin/bash # save this to /tmp/exportenv.sh for i in `env`; do echo export $i done ```` Then run `/usr/bin/env...

Yes, that is inconsistent. How about an R level option, say options(rapache.formvars.default.value='') help?

Or, it could just default to using empty strings for Rook apps.

Can you send me the test script or how to go about replicating? On Dec 23, 2013 9:52 PM, "Jeroen Ooms" [email protected] wrote: > After updating to rApache 1.2.5, almost...

Sure! On Feb 12, 2015 5:03 AM, "Ruben C. Arslan" [email protected] wrote: > I'm gettting about two of these notices per second, but with a colon in > front of...

ee4f2bb46087d305c45cc64a72fbc2215284cef7 fixes the first warning, but I don't have a Fedora box to test the others.