freenode/#clim - IRC Chatlog
Search
12:44:21
beach
The specification uses the term "command" for several different things, unfortunately.
12:44:56
beach
And DEFINE-COMMAND does not define a command according to the meaning(s) of the term.
12:46:08
scymtym
it says "/command/ is a command object, that is, a cons of a command name and a list of the command's arguments." (// indicating italic). so it refers to the parameter specifically
12:51:22
jackdaniel
scymtym: yes, allowing only a command name is more a syntactic sugar. the real problem was that restart report function assumed, it is a cons
12:52:02
jackdaniel
so if we have issues invalid command (i.e atom), error was caused in restart report function (when debugger was invoked)
12:54:22
scymtym
but won't the change enable the execution of invalid commands? that is, i would have expected (command-name (ensure-list …)) in the restart, not on the normal execution path
12:58:00
jackdaniel
ensuring that it is a list falls into robustness principle (be liberal with what you accept, be strict with what you issue)
12:58:38
jackdaniel
but I'm fine with any solution. I don't think there is a need for specialization here
13:05:59
scymtym
i mean specializing the default method on CONS so a no-applicable-method error will be signaled
15:16:24
jackdaniel
I also have some fun with visualisation right now: http://i.imgur.com/i8DvvST.png
15:17:10
jackdaniel
your hint about contrasting colors was right on a spot, I wonder if we could find more than just 8 which would look good
15:17:55
jackdaniel
beach: I'm writing personal accounting system, we plan to use it as a training program for testers
15:18:40
jackdaniel
empty part is remaining monthly resources, filled parts are different categories of expenses
15:20:49
beach
Well, I guess that depends on your functionality. I found CLIM very handy for the double-entry accounting system I wrote.
15:25:30
scymtym
the flamegraph thing is a bit challenging with respect to performance. that's why i applied it to itself :)