On Aug 31, 2005, at 4:19 AM, Ralf Bokelberg wrote:
Interesting question, i think.
You could achieve it by the convention only to commit code, which
passes all the tests. If you update your codebase in the morning, and
run the tests too, and they break, you can assume, that the last
commit wasn't ok. Additionally you could arrange it so people have to
commit a testprotocol too.
Yeah, but that's one more thing I have to remember to do, and my brain
is in need of a RAM upgrade as it is ;)
To achieve the same using technology, we would need to automatically
compile all tests on the fly, run them and write the results back into
some file. This would take some executable, which wraps the
flashplayer and enables the test.swf to write to the filesystem.
Additionally you need a new TestRunner class, which does just that,
collect the results and writes them to a file.
Is this something that would be feasible with an SWF2EXE program like
Zinc or SW?
The main problem i can see with the technological approach, is to
create and keep up to date all this different tools for your server's
os. I'd rather go with convention.
If I can get it working I'll think about the logistics of maintaining
it, but for me even an out-of-date automated test solution would be
better than none at all.
Benjamin Jackson wrote:
Well...
If something breaks, I want to know about it. So yeah, I guess
postprocessing is a part of it. So is running all the regression
tests after each commit so that I know immediately which commit broke
the build (and more importantly, who will be paying for takeout the
next day) . And being able to integrate running the tests with a
package like ant or rake (my personal fav) seems like it would
require this as well. What do you think Ralf?
___________________
Ben Jackson
Diretor de Desenvolvimento
+55 (21) 9997-0593
[EMAIL PROTECTED]
http://www.incomumdesign.com
_______________________________________________
osflash mailing list
[email protected]
http://osflash.org/mailman/listinfo/osflash_osflash.org
