Greetings Jan,

Can you describe the feedback you'd be looking to gather for this test day? What sort of tests (or activities) would you want participants to run?

Hello!

== Testcases ==

I have repared 2 test-cases:

=== Look for MD5 functions calls ===

Use systemtap scritp to track functions calls of relevant binaries (according to [https://fedoraproject.org/wiki/Hash_algorithm_migration_status#Done migrated packages]). We are interested in functions with names as md5 vs. MD5 vs. md_5 vs. MD_5 vs. md_five.

• http://sourceware.org/systemtap/examples/keyword-index.html#PROFILING - see profiling/functioncallcount.stp
• TODO: come up with script modified for us

=== Try to configure to not use MD5 ===

Use configuration suggested in config files to disable MD5 and try to use configured programs. These sugestions should be as a comments in config files and we are looking for cases where suggested (not default) config do not work as expected. Maybe we could use systemtap script from the previous test-case to ensure that MD5 function calls are only present with default config and not with this modified config.

== Output from test-day ==

We want to answer questions: are there any remaining calls to the '''md5''' functions in the binaries we have fixed? And: Does programs configurable to the state where they do not use '''md5''' works correctly?

I'm pretty open for any other ideas :)

Greetings Jan,

Thanks for the feedback. Adam Williamson and I discussed this proposed test day topic. If I'm understanding correctly, the idea is that participants will run some scripts against binaries to determine what calls remain to md5 functions. We're both under the impression that this might not be the best fit for a ''distributed'' test day model. Perhaps this might make more sense as a transparent/public test effort. Where, I don't think there would be lots of value in distributed participation, but maybe you increase awareness of this testing effort by being available on #fedora-qa ... and communicate the activity plan and results on fedora-test-list ... and planet.fedoraproject.org?

If it turns out there is a lot of work ... and you need people to ''process'' results somehow. That might be a good way to document and demonstrate how to get started with the work (with blog and test-list communications). From there, we could solicit volunteers to walk down the remaining TODO's?

Hello,
you are definetedly right - thanks for all the feedback. The goal here was that everybody is using the different configuration, so binaries are using different code paths. But I like the idea you have mentioned, it would be probably better to do it this way. I'll remove the test-day from the calendar and test-day page as well. And I'll probably email you and ask you for some propagation of these testing efforts. Does that seem like a sane plan to you?

Also I would like to do one more test-case which can be connected with this but originally comes from the "Yum--use libcurl / enhanced SSL for NSS/FIPS" feature and yum developers:

=== Yum works through various Proxy configurations ===
We want to ensure, taht after yum uses pycurl, it still works through various odd proxy configs in the wild.

1. Point yum to this ... repository
2. Try to do these ... basic yum commands
3. Send us the proxy name and maybe configuration you are using

Early in the F-11 test cycle, I tried a test event (not really a test day) where chris lumens (pykickstart developer) and I generated a framework, some examples and a TODO list. We invited folks to join and provided guidance on mailing list and IRC for participants. This '''might''' be something to consider? See several blog posts regarding the event ... http://jlaska.livejournal.com/tag/pykickstart (note, these posts are in reverse chronological order).

This has the benefit of not really requiring distributed testing, but perhaps carves out work and defines how participants can engage ... and isn't time sensative ... it can be worked in a day or over a month. In the end, if no participates join, other than you and the maintainer(s), it's still a success in that you've scoped a project and started to work down the list.

The benefit might be in doing this in a transparent open manner as to solicit ideas/feedback/interest from the fedora community.

So summary, I don't mean to discourage in not using a test day for this event. I think it would be an interesting test 'event/project' to blog about and discuss over email. What do you think?

That seems as a great way how to do this, thank you. You can close this ticket.

Thank you,
Jan

Thanks Jan, keep me in the loop if you don't mind. I'm interested in helping find a solution that fits your needs.