Figuring out a nice workflow for unittesting with Emacs

Figuring out a nice workflow for unittesting with Emacs

[Rob Renaud]

[-- Attachment #1: Type: text/plain, Size: 658 bytes --]

I am used to slightly abusing compile command to run unittests, because i
get the nice go to file:line behavior on failures.  But I am now unittesting
3 different files at different levels of abstraction and constantly
going back and forth between the tests.  When I am unittesting a single file
at a time, compile-command works beautifully.  But now I need to constantly
re-type the compile command when I want to change the test I am running.

Is there a way to add history to the compile command function?
 Alternatively, can I augment use some shell mode (which obviously has
command history) with the go to file:line behavior when I get a test
failure?

[-- Attachment #2: Type: text/html, Size: 711 bytes --]

Re: Figuring out a nice workflow for unittesting with Emacs

[Eli Zaretskii]

> Date: Sun, 17 Apr 2011 18:55:08 -0400
> From: Rob Renaud <rrenaud@gmail.com>
> 
> Is there a way to add history to the compile command function?

"M-x compile" has its own history, granted.  You can move through it
with M-p and M-n (resp up-arrow and down-arrow), and you can serach
for commands that match a regexp with M-r.

In addition, there's "M-x recompile" that will automatically reuse the
last compile command.

Finally, you can use file-local variables to define the compile
command for each file (unit test in your case).