GSOC 2015 Documentation

All technical documentation has been posted to the GitHub repositories, either in the form of the README, or in form of comments.

The repositories are located here:

https://github.com/canihavesomecoffee/ccx_testsuite (Project from GSoC 2014, not that much changes this year, except for some fixes and improvements)
https://github.com/canihavesomecoffee/ccx_gitbot
https://github.com/canihavesomecoffee/ccx_submissionplatform

Test Suite

No changes from last year, still as described in the first section of CCExtractor regresssion testing / GitHub bot.

GitHub bot

Refer to the bot section in the CCExtractor regresssion testing / GitHub bot page.

Submission platform

To view samples, test results and download samples there is no need for an account. To upload samples, registration is required. Samples (<1GB) can be uploaded through HTTP; FTP allows all sizes. Special care has been taken to keep all samples anonymous (except to the user itself and the admin), and if required, a user can anonymize his/her account.

The bot can be evaluated (user-side) by issuing commands to it on GitHub (see CCExtractor regresssion testing / GitHub bot for instructions). It's been used already by Anshul, Kisselef and Ruslan for the last PR they made.

The sample submission platform can be tested (user-side) by either browsing around on the site and/or creating an account (the latter is required to upload samples). At this moment, a limited section of regression samples was already added, and this will be expanded very soon.

To test the admin parts of both the bot & the sample platform I refer to the mail I wrote to Carlos.

The submission of new samples has been quite chaotic in the past of CCExtractor, with users submitting samples through FTP, GitHub, Google Drive, upload websites, email, … The sample submission platform aims to create a single place to retrieve and upload new samples, as well as showing the current state (fixed, broken, …) for those samples. Besides that, a GitHub bot was introduced to allow other contributors run the regression tests without having to download all the samples. This should improve the stability of CCExtractor and hopefully lower the barrier for new contributions.