Re: [tor-dev] Help with rebasing arma's n23-5 tor branch to current master

Hi Karsten,

I probably can't make the meeting since it is during my work day, but I'll try. I might be able to just leave myself logged in to the chat from home so I can read the transcript later.

You can grab my commits here: http://www.charliebelmer.com/tor.git:

git clone http://www.charliebelmer.com/tor.git

Basically, I took Arma's branch, and rebased the current master onto it. There were a number of merge issues, including the connection to channel changes that Andrea developed. I believe that I converted N23 to work with channels correctly, though I am still trying to figure out how to actually validate it properly (aside from basic run testing - advice / thoughts / instructions welcome!). I expect it will need some tweaks.

I started playing around with Chutney, but I don't fully understand what it does yet, so more to come there.

I would be happy to help take this section forward however you need - it sounds interesting learning how to test this to see what performance gains it might yield, if any.

Charlie

On Tue, Jul 30, 2013 at 4:05 AM, Karsten Loesing <karsten@xxxxxxxxxxxxxx> wrote:

On 7/27/13 5:55 PM, Charlie Belmer wrote:
> Hey Everyone,
>
> I wanted to give quick update on N23 rebasing progress:
>

> - Rebase N23 onto Master (Complete)
> - Refactor N23 to use channels (Complete)
> - Make clean w/ check-spaces (50%. ETA 7/28)
> - Testing with Chutney (To-Do)

Sounds great!

We'll have an IRC meeting on Wed July 31, 18:00 to 19:00 UTC in #tor-dev
to discuss deliverables including this one.

Can you give me more details what you achieved and maybe push your
rebased branch somewhere, so that I can report your progress at the meeting?

Or do you want to join the meeting? If so, see my earlier mail to this
list for details.

Thanks!

Karsten

> On Thu, Jul 18, 2013 at 11:43 AM, Karsten Loesing <karsten@xxxxxxxxxxxxxx>wrote:
>
>> On 7/18/13 5:30 PM, Charlie Belmer wrote:
>>> Karsten, I'll take a shot at it. Is there a specific timeline I should be
>>> aware of? My C is a little rusty and I am working to learn the codebase
>>> (This sounds like a good opportunity to learn more). I don't want to bite
>>> off more than I can chew if there is a tight release timeline.
>>
>> Neat!
>>
>> We have until end of October to "Write a proposal for N23, and write a
>> patch for N23 that functions correctly. Then simulate it to see how it
>> performs for typical and for really slow client connections."
>>
>> We're sorta doing this in opposite direction, because there's already a
>> patch, and we want to find out if performance improvements are worth
>> writing a proposal and cleaning up the patch enough to actually merge it
>> into 0.2.5.x.
>>
>> So, we have 3.5 months for everything. I'd think that 1 month for
>> rebasing the existing branch would leave us enough time to run
>> simulations and possibly write the proposal. Is mid-August a reasonable
>> time frame for you?
>>
>> Of course, if you realize you can't work on this, please let us know as
>> early as possible, so that somebody else can give it a try.
>>
>> Thanks for helping!
>>
>> Best,
>> Karsten
>>
>>
>>> On Thu, Jul 18, 2013 at 8:50 AM, Karsten Loesing <karsten@xxxxxxxxxxxxxx
>>> wrote:
>>>
>>>> Dear tor-devs,
>>>>
>>>> is anyone here up for a coding task that could help us research
>>>> performance improvements of the N23 design more?
>>>>
>>>> The situation is that we already have a branch (n23-5 in arma's public
>>>> repository), but it's based on 0.2.4.3-alpha-dev and needs to be rebased
>>>> to current master.
>>>>
>>>> In theory, it's as simple as the following steps:
>>>>
>>>> $ git clone https://git.torproject.org/tor.git
>>>> $ cd tor/
>>>> $ git remote add arma https://git.torproject.org/arma/tor.git
>>>> $ git fetch arma
>>>> $ git checkout -b n23-5 arma/n23-5
>>>> $ git fetch origin
>>>> $ git rebase origin/master
>>>> (clean up the mess)
>>>> $ git add
>>>> $ git commit
>>>> $ git rebase --continue
>>>> (back to clean-up-the-mess step until git is happy)
>>>> $ git push public n23-5
>>>>
>>>> Bonus points if you make sure the branch compiles with gcc warnings
>>>> enabled, appeases make check-spaces, and runs peacefully in a private
>>>> Chutney network.
>>>>
>>>> Unfortunately, the n23-5 branch touches a few places in the tor code
>>>> that have been refactored in current master, including Andrea's
>>>> connection/channel rewrite. It might be necessary to dive into the
>>>> channel thing in order to get this rebase right.
>>>>
>>>> Once we have a refactored n23-5 branch, I'll try to simulate it in
>> Shadow.
>>>>
>>>> For a tiny bit of context, this is for our sponsor F item 13:
>>>>
>>>>
>> https://trac.torproject.org/projects/tor/wiki/org/sponsors/SponsorF/Year3
>>>>
>>>> I'm asking here, because the usual suspects are already overloaded with
>>>> other stuff. As usual, I guess.
>>>>
>>>> Thanks,
>>>> Karsten
>>>> _______________________________________________
>>>> tor-dev mailing list
>>>> tor-dev@xxxxxxxxxxxxxxxxxxxx
>>>> https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-dev
>>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> tor-dev mailing list
>>> tor-dev@xxxxxxxxxxxxxxxxxxxx
>>> https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-dev
>>>
>>
>>
>