checking httpd syntax from within a shell script

couldn't find this anywhere including on this boards. seems strange that nobody has run into the same need ever before. this should have a very simple answer.

trying to put the httpd -t output into a shell script variable. have tried all possible options that i know. stuff like:

r=`httpd -t` or r=$(httpd -t)

even tried:

for i in `httpd -t`

any other "normal" shell command like date or ls, etc works but i can't get that simple output to be put in a variable - doesn't matter if the output is an error, warning or just "Syntax OK".

essentially i just want to have a shell script that makes sure any programmatical edits to the httpd configuration have not screwed up the conf before restarting the server. i refuse to believe that this is not possible with a shell script...

thanks.

If I recall correctly (I don't have a machine with Apache installed in front of me right now), the output text from that command is going to stderr, not stdout. You need to redirect the output in order to capture it, like this:

r=$(https -t 2>&1)

The above will collect both stdout and stderr output. If you want only stderr without stdout, you need to do something like this:

r=$(https -t 2>&1 > /dev/null)

However, you don't actually need the text output in order to do what you want. If all you need to know is whether there is an error or not, you should just check the return code. If the configuration is error free it will return 0, if there is an error it will return some other number. From the man page for httpd:

Run syntax tests for configuration files only. The program immediately exits after these syntax parsing tests with either a return code of 0 (Syntax OK) or return code not equal to 0 (Syntax Error).