Harish Krishnaswamy

• Running Federico’s oneliner for unshared mappings for Evolution and EDS (unstable series) gave me this :

  evolution-2.6: 38144
  evolution-data-server-1.6: 36056
  evolution-alarm-notify: 11848
  evolution-exchange-storage: 2108

• Some earlier efforts on studying and profiling Evolution memory and performance is available here and here too.
• I am personally excited by the rejuvenated focus on Evolution’s performance woes. I look forward to spending much of my spare cycles during the year-end ‘slow-down’ towards putting Evolution on a memory diet and clearing off unreviewed patches towards memory and performance.
• And for those who were eagerly awaiting, some early results on DBUS Vs. Bonobo. No conclusions yet – there are a lot more ifs and buts to be sorted out.

>

Tigert,

The Memos component in Evolution is actually about VJOURNAL data in EDS that can be sync’d with your Pilot and also be recognized by Groupware servers. What I need now is a right-click option on Tomboy that exports the data to EDS – so my desktop notes can flow into anything that interacts with my evolution data server. Considering how addicted I am to Tomboy now, it would be awesome to be able to access it anywhere and share my notes as ‘memos’ with my contact-list buddies.

The thousand and odd test cases written and used by the Novell Evolution QA team in the 2.4 development cycle are now ‘open’ and have been published here. Nagappan, the LDTP wizard, is working on making this available as a TestLink repository where the community (and other distros that do their own Evolution testing) would be able to add new test cases, look up reports of test execution and so on.d so on.

I was in GUADEC last week. Mapped faces to nicks. Met old friends and made a lot of new ones. Learnt a bit on a wide range of things. In all, had a fabulous time.

While almost everyone gave me a personal crib sheet on evo – their pet peeves etc., no doubts on how much they use it and how much they care. Drew up a Ten by Ten plan – clear 10 unreviewed patches a week till October – to reduce the backlog (160) and climb down a few notches – actually, the first step is all that matters 😉 – on the various reports on bugzilla.gnome.

Lots of action ahead !!!

With Evolution 2.3.2, you can now compile plugins (based on the mailer) independent of the Evolution sources, thanks to NotZed. Building cool and nifty stuff so quickly with Evolution was never so easy as it is now. To confirm this, the Evolution community is planning a hackfest today at #evolution. More at the evolution wiki . Drop in and have fun.

And yes, I agree with Garnacho. We need to find some place to store all the wonderful plugins that are going to flood us 🙂

On Wednesday, six interns under the NOSIP programme joined us for an Evolution hackfest at office and walked out as
proud Evolution hackers by the end of the day. Working on what we have heard from the folks who have participated in NOSIP (it is just over an year old now), the hackfest was designed to show the newbies and the wannabes – what open source hacking is all about – fun and learning. The idea was to let them get a feel of ‘Open source development’ by watching me and chen working on some live bugs – so they can have a crack at it in the second half.

(Left to Right – Harish, Sharath, Archana, Chen, Darshini, Shashidhar, Vivek, Praveen)

After exchanging names and introductions, we got down to business. None of them had used Evolution before (for all but two – it was the first encounter with GNOME/Evolution, the other two were a week old). So, Chen began by showing off Evolution. Then it was build time. Contrary to the approach of building 46 modules to get evolution up and running, we piloted a new approach, (hope this helps the students who are on low-bandwidth, unreliable internet connections) – Install the Novell Linux Desktop (evaluation copy) with the development packages for most of the basic libraries and build just 5 core modules.. so you can start looking at the source and get hacking, rather than spending too much of time on trying to get myriad version of autotools and libtools to work with each other and finish off with 46/46 on jhbuild.

While jhbuild bootstrapped, I briefed them about the project, the evolution hackers and the individual modules they work on, the blogs, bounties, planet.gnome.org, mailing lists, the bugzilla, bug days, the irc channels and some popular nicks. Also shared some of the useful environment settings, scripts and nifties to make work light while hacking.

We had started off with the source (a few months old as found on the cds that we distribute) and updated it to the HEAD (that had just seen a release a couple of days and hence had version bumps). Then began a wave of errors and we got the ‘now you know what we went through’ looks from those who had attempted this before.
The guys watched me and chen solve them ‘live’ and thinking aloud – probing for causes, discussing and arguing on what is the best way to get rid of the block – pull it from rug, change the aclocal flags, uninstall the autoconf that seems to interfere, make clean to get rid of er, some stray binaries that had made their way into the source cds.

After about an hour, the guys got into the groove and started cracking the issues on their own. We were joined by the rest of our evolution team in bangalore – setting off another round of intros and a mini debate on open source Vs free source. By lunch, we were still building. As we were joined by the evo and oo team over the lunch table, chen chose to be a lone warrior ignoring my calls over his mobile to grab some food.

Post lunch, while the machines churned out the binaries for the 5 modules, we walked them through a complete bug fixing cycle – picking up the bug from bugzilla (how to select, choose and query), understanding the problem, checking if hte bug still exists, locating the right module and grepping your way to the source of the problem. understanding the context, sticking to the code conventions while making the changes, and the fix->compile->test cycles till the bug was solved to satisfaction. Then of course, writing a Changelog and generating the unified patch. This was followed by a second demo – solving a crash to demonstrate the use of gdb – attaching to a process, analyzing the stack traces, breaking at the right spot etc.

Then it was show time – we tossed some of the bugs we had handpicked for them and the guys split into teams of two and kicked off. By then, just two machines had all the 5 modules built while the rest were ambling their way towards the mark. So chen graciously shared his source tree and i had to follow suit.

[A short digression… me and chen then discovered that we had mistakenly ‘used’ one of the bugs we had earmarked for the day during the demo. Evolution to the rescue.. as one of the teams ran evo they had just compiled and tried to create an account – they ran to a crash. they decided to fix that. cool.. some luck]

The next hour, they spent glued to their monitors – trying out stuff, occasionally asking a question or two.
Neeti dropped in to say hi – and a few exchanges on why each had chosen to do opensource and how they can benefit from the programme followed. Neeti, then handed over a prized gift – a gnomebangalore T shirt to each of them.

Team one had a false start, thought they had just cracked it only to find that they had fixed the crash but introduced a different problem. They were back at it again after realizing why their fix would not work.

The other bug, 58786 that had proved quite popular was being attempted by both the other teams. – a sheer coincidence, it had been filed by another NOSIP intern (hpnadig). This was the problem of the ‘Mark completed’ menu item not being disabled even when invoked on a completed task item. We dropped in a few hints and some nudges, careful not to give the solution away – for eg. the need to search for a particular function in evolution-data-server and not under evolution – or where to look for the callbacks – areas where brute find and grep would not help.

We overshot the planned time but our young hackers would not leave until they have seen the end of it. Then came the first fix, the second and the third. On pointing a possible optimization, Team 3 had a go at it again.
With the joy on their faces seeing their fix working, we knew the Open source bug had bitten them. They are looped in for life. It was pure ecstacy as they composed the Changelog and generated the patches and posted them on the mailing list. Fun, relief and pride written all over their faces.

At the end of a long, tired but a day well spent – more feedback on what they would like us to do while i stressed on the need for them to contribute not just by submitting patches, but also by filing bugs, triaging, adding documentation to the wiki sites as well as helping other newbies in gnomebangalore – a number that seems to be rising everyday.

Just a request to fellow gnome hackers who stopped by to read this – an increasing number of such hacker newbies work during the nights – (the connectivity is better in off-peak hours and also cheaper for the students who do make sacrifices to afford the luxury of being online) – while most of us here do try to lurk around during midnights – you can make a big difference to these people and GNOME in India.

It would be a great opportunity for these guys to interact with the gnome community all over the world over the irc – getting some answers and guidance, participating in bug days, learning the ropes and sharing thoughts..
Drop in to #nosip on irc.gimp.org, do spend some time with these students and welcome these guys into our wonderful community.

Friday – Some of my fellow cit(y)zens decided that

1. some of the vehicles plying on the Hosur road needed newer windscreens and

2. we should get a ‘forced break’ from work.

So, on a Friday afternoon, I found myself staring at the television watching one of our leaders on national television – hinting to the anchor that the media and the literati has never got it right in getting the pulse of his electorate – he knows them better, he has been getting it right for the past two decades, even though some of them paint him as a clown, at best.
Glanced through a nice article from Richard Koch – hypothesizing that we in the 21st century are still largely driven by Neanderthal genes and it is better to wake up to this reality than fight it. Helped me view both the previous events of the day in a new light.

BTW, I had promised some starter material on evolution to some of the NOSIP hackers. I had intended to post it on the gno-bang wiki, but it is finally up on a more appropriate place, http://live.gnome.org/Evolution

And it is a wiki, so pl. feel free to edit and improve the material. Contribute..