The core reads ~/.radare2rc while starting. You can add e commands to this file to tune radare configuration to your taste.
To prevent radare from parsing this file at start, pass it -n option. To have a less verbose output for batch mode runs, it is also better to decrease verbosity level with -v command-line option.
All the configuration of radare is done with the eval commands. A typical startup configuration file looks like this:
$ cat ~/.radare2rc
e scr.color = true
e dbg.bep = loader
Configuration can also be changed with -e <config=value> command-line option. This way you can adjust configuration from the command line, keeping the .radare2rc file intact. For example, to start with empty configuration and then adjust scr.color and asm.syntax the following line may be used:
$ radare2 -n -e scr.color=true -e asm.syntax=intel -d /bin/ls
Internally, the configuration is stored in a hash table. The variables are grouped in namespaces: cfg., file., dbg., scr. and so on.
To get a list of all configuration variables just type e in the command line prompt. To limit output to a selected namespace, pass it with an ending dot to e. For example, e file. will display all variables defined inside "file" namespace.
To get help about e command type e?:
Usage: e[?] [var[=value]]
e? show this help
e?asm.bytes show description
e?? list config vars with description
e list config vars
e- reset config vars
e* dump config vars in r commands
e!a invert the boolean value of 'a' var
er [key] set config key as readonly. no way back
ec [k] [color] set color for given key (prompt, offset, ...)
e a get value of var 'a'
e a=b set var 'a' the 'b' value
env [k[=v]] get/set environment variable
A simpler alternative to e command is accessible from the visual mode. Type Ve to enter it, use arrows (up, down, left, right) to navigate the configuration, and q to exit it. The start screen for the visual cofiguration edit looks like this:
Eval spaces:
> anal
asm
scr
asm
bin
cfg
diff
dir
dbg
cmd
fs
hex
http
graph
hud
scr
search
io