Automated tests using dovecot
glasse at cs.rpi.edu
Wed Jun 8 16:58:10 BST 2011
On 06/08/2011 11:12 AM, Sebastian Spaeth wrote:
> first of all, let me say this is fantastic work and exactly what is
> needed to get me to really, really trust oflineimap with my mail. We
> should definitely accept this and build upon it. That having said, the
> bash-based testing syntax is certainly weird if you are not used to it :).
It is a little weird but I feel the same way about Python unit test cases :)
> +1 from me. It will require some more fiddling to avoid hard-coded paths
> in dovecot.conf.template, no idea how that would be possible.
Oh, yeah, forgot about those. That should be fine, we can just use the
same trick as the mail location.
> OK, I had to fiddle with some hard coded paths and the password file
> settings to get it up and running, but got it in the end.
What password file settings? The location of the file?
> My first improvement suggestion is that you attempt to locate the path
> of "dovecot" with "which" and attempt "/usr/sbin/dovecot" if that did
> not work out. I do not have that path in my regular search path as
> normal user and it failed to invoke dovecot (basically without any
> sensible error message).
Oh, yeah, that's an important one.
> FAIL test suite can mark a message as read
> --- basic.8.expected 2011-06-08 15:05:13.004041250 +0000
> +++ basic.8.output 2011-06-08 15:05:13.004041250 +0000
> @@ -1 +1 @@
> ./test-lib.sh: line 643: [: cur: integer expression expected
> mv: cannot stat `test-maildir/INBOX/cur/*,': No such file or directory
> FAIL read flags are synced
> --- basic.9.expected 2011-06-08 15:05:13.384041251 +0000
> +++ basic.9.output 2011-06-08 15:05:13.384041251 +0000
> @@ -1 +1 @@
Oops! It looks like I forgot to commit something here. Try a pull and
Assuming that everyone is in favor, once it gets cleaned up, what's the
way forward for getting something like this merged? I'm a big fan of
small, atomic commits, but I'm not really sure how to break something
like this down.
More information about the OfflineIMAP-project