On Thu, Jul 21, 2016 at 2:38 AM, Meinert Schwartau <m.schwartau(a)gmail.com>
wrote:
---------- Weitergeleitete Nachricht ----------
From: Tudor Girba <tudor(a)tudorgirba.com>
To: Moose-related development <moose-dev(a)list.inf.unibe.ch>
Cc:
Date: Tue, 19 Jul 2016 07:41:05 -0600
Subject: [Moose-dev] [ann] jdt2famix standalone - alpha version
improve the error handling. As Java has a ton of
edge cases, it would be
great if you would use it to parse your systems and report possible errors.
We would then manufacture test cases out of those.
I'll definitely give it a try. Because the number of people who'll try
this out will be limited, to be sure that there aren't any bugs in this
parser I suggest that we let it parse all java github projects. Then we
could group the resulting errors (if there exists any ;-)) per error type,
write tests for these edge cases and then fix them. If all of the java git
hub projects can per parsed without errors then we could be really sure
this parser can is at least to parse the edge cases without exceptions. For
sure, we don't know if the result makes sense but at least it produces a
result.
Getting the download url for all java github projects is very easy:
https://api.github.com/search/repositories?q=language:Java
Further optional parameters for this url for paging:
{&page,per_page,sort,order}. Documentation for this request:
https://developer.github.com/v3/search/#search-repositories
How do you think about it?
I wonder if you could automate this with...