When relevance is evaluated in the QnA GUI, it does not showtypes. If the -showtypes command line parameter is used when the qna binary is invoked to evaluate the relevance, then qna will return the type of the result, such as boolean or string or similar.

Examples:

• https://github.com/jgstew/tools/blob/master/bash/openQnA.sh
• https://github.com/jgstew/tools/blob/master/AppleScript/openQnA.scpt

Whenever QnA GUI is run as root, it opens, but any relevance entered does not evaluate. It seems like when run as root it can't find the qna binary or something like that. The app gives no indication as to why the evaluation is failing.

Sometimes testing relevance as the root user is required to more closely emulate the environment in which the BigFix client evaluates relevance.

For example, relevance evaluated as root would be able to read files that relevance evaluated as the current user cannot.

Related:

• https://bigfix.me/fixlet/details/20469#comments
• http://osxdaily.com/2013/02/06/how-to-run-gui-apps-as-root-in-mac-os-x/
• http://hints.macworld.com/article.php?story=20030330175504453
• https://github.com/jgstew/tools/blob/master/bash/openQnA.sh

This is a minor thing, not an immediate request.

It would be interesting to look for QnA in other locations like the Applications folder like it used to exist in the past, as well as the locations that Matt's AutoPkg install recipe puts it.

I think it makes sense to prefer the QnA bundled with the BigFix agent if it exists, otherwise it would probably check each location until it finds one. Bonus points would be if it could detect the newest version of QnA available on the system.

just before running a relevance query through QnAGUI, save a copy of the query to be run to /tmp/_QnA_GUI_relevance.txt or similar so that if I ask a relevance query that takes forever to return, that I can still recover my last query if I have to force quit the app.

Next level would be on start of the app, load the last query from that file if it exists.

I have a bad habit of accidentally asking unanswerable questions of relevance when testing.

other queries seem to work fine, but if I try to evaluate properties, which has a very large result, QnA GUI crashes.

Not sure if it always did this.