UCOSP

Undergraduate Capstone Open Source Projects

Archive for the ‘Sprint’ Category

“What we learnt” + “Where we’ll go”

Posted by clgoh on 2010/01/19

The day before the sprint, I knew of nothing about what I should do and what I can do in this project. I woke up at 5 on Friday morning, took a taxi to the bus terminal, got onto Greyhound to Toronto, got lost on our way to Bahen Centre (we were on University and College, 5 minutes away from Bahen, lol)  To my surprise, I arrived, and joined a team of people as lost as I was. We found out throughout the sprint, many differences between universities, the courses, the cultures, etc. But all of us being in our last year, I believe there’s one thing in common that we’ve learnt before we get there: Well, no one’s going to help us if we don’t start helping ourselves. Before long, the whole team started the sprint ourselves. We asked lots and lots of questions (Thank you very much Andrew for your detailed presentation and all your help even by having to whisper), we read wiki pages, downloaded source code and snorkeled in it. No one sat still waiting for an answer, there was no awkward silent moment as the team of “professionals” blended from the very first day and moved forward as one. The three days went by very quickly.

We learnt the following during the sprint:

– OpenLayers draws map using Web Mapping Service provided by Geoserver, which in turns uses Geotools to obtain GeoSpatial Data from databases. Geotools can currently talk to many databases but not Ingres.

– In Geotools there are connectivity code under modules/plugin/jdbc and modules/plugin that uses various jdbc implementations to connect to various databases. There is also a deprecated version of the framework scheduled to be removed soon. Ingres team needs to make a similar jdbc implementation under Geotools to add support for Ingres.

– Sara and Eva will be working on the Ingres server side to make sure GeoSpatial queries are handled properly, while Anthony, Henry, Xiaoxiao and Lim work on Geotools to add Ingres support.

We decided the following during the sprint:

The project lead will communicate with the team and with the community at large to keep everyone informed about the status of the team, and make sure the team is moving in the right direction. Lim will take on this role.

The high-level view of the project is:

– Study the code and play with Geotools in action

– Design and plan potential changes to Geotools source tree

– Implement the designs and also implement support for GeoSpatial data in Ingres

– Unit test the Ingres connectivity code in Geotools, test GeoSpatial data support in Ingres, and system test Geotools actually using Ingres.

Now everyone’s back to where they are from, but technology is constantly connecting us together. We are all a phone call/email/IM message away from each other, and we are preparing for the next get-together online every Monday with more progress to make.

Posted in Ingres, Sprint | Tagged: , , | Leave a Comment »

The pony-build Express

Posted by mlaite on 2010/01/19

As the dust settles from this years Spring Code Sprint, I am left with some interesting reflections. One, coding for 9 hours straight sure gets hard without proper caffeine intake, maybe an IV drip will be a good investment for the next one. Two, ordering lots of food when someone else is paying is lots of fun, even if you end up walking (waddling?) out with some of it in a bag. Three, you can never have a shortage of computer related jokes or quips with rooms full of CSC majors.

The only sad part of my Sprint experience was that my group was unable to make it up to Toronto. Luckily for us Al Gore was able to invent the internet, so no matter how far apart people are we can get stuff going. With Titus, our fearless leader, at the helm we quickly got the ball rolling.

  • We where able to get the remote group members setup and running pony-build.
  • We cleaned up and implemented the virtualenv support for build recipes.
  • We began work on some other improvements and bug fixes.

My work at the Code sprint was on getting virtualenv support working properly and working on this issue. All and all it was a fairly successful two days. Since pony-build is in such an early alpha state, our team will pretty much be able to mold it however we want. I am looking forward to working on this project throughout the semester and seeing where we can take it.

Posted in Sprint, Status | Tagged: | Leave a Comment »

And A Good Time Was Had By All

Posted by Greg Wilson on 2010/01/18

On Friday morning, 45 strangers sat down to work on this term’s UCOSP projects. On Sunday, after three solid days of coding, talking, and learning, seven software development teams headed home. It was a long haul (especially for our visitors from the west coast and the Caribbean), but a lot of fun. I’d like to thank everyone who took part, especially Alan Rosenthal, Carolyn Ursabia, Mike Conley, and Zuzel Vera Pacheco, who helped set everything up and run smoothly. I’m looking forward to the next eleven weeks a lot.

Posted in Sprint | Leave a Comment »

Mercurial – Getting Into the Groove

Posted by Anton Markov on 2010/01/17

The coding sprint is winding down here in Toronto, but the Mercurial UCOSP project is just getting into the groove. As people are making their way home – across the street or across the country – I want to take a few minutes to expand on Alex’s post and summarize this weekend.

Working on a mature project like Mercurial has its advantages and disadvantages. On the one hand, we have a real successful project to learn from, but at the same time, it is hard to find work that is both accessible to a newbie and has real impact on the project. With that in mind, I think we’ve all made great progress this weekend.

I can’t comment on everyone’s work individually because I have not tried out all the changes, but a quick look through Kiln (our source control system) shows that everyone has spent some time today on code cleanup and formatting. There are many references to PEP-08, the Python coding conventions. This is great news because it means people are close to getting their code ready for wider-scale review. It also means that we all have lots of code-review work coming up.

An important part of this code sprint was learning to work together and establishing the conventions for communication. This includes learning to use FogBugz, our bug tracker and wiki, Kiln, the source control and code review system, and setting up weekly meetings. Our supervisors have been emphasizing the need for more communications every day, and these are the tools will will be using.

FogBugz and Kiln (user: Guest, password: anonymous):

Blog: part of the UCOSP Blog under the Mercurial category – you are reading it.

Weekly meetings: IRC-style; details and archive will be available

General Mercurial resources:

General UCOSP resources:

Alex wrote a great summary for Friday and Saturday at https://ucosp.wordpress.com/2010/01/17/first-push-mercurial/.

Posted in Mercurial, Sprint, Status | Leave a Comment »

Fingers of Fury

Posted by evanstratford on 2010/01/17

It’s 2:30 pm on Sunday; a thick pallor persists over an unseasonably warm Toronto. For roughly 40 students from various Canadian and American universities, this marks the end of the UCOSP code sprint. Those of us working on Thunderbird have made progress on a number of bugs tagged with student-project:

  • Zach Church is tackling Thunderbird’s handling of recently closed tabs as reported here and here.
  • Tim Miller is working on better integration with Windows 7 as reported here and here.
  • Marcel Guzman is improving the Activity Manager as reported here and here, and is also adding a keyboard shortcut for switching between message and header panes as reported here.
  • Kefu Zhao is automatically inferring From: addresses for replies to mailing lists as reported here.
  • Lindauson Hazell is resolving a bug where message-related menu items are enabled in non-message-related tabs, as reported here.
  • Evan Stratford is adding support for vCard import/export in Address Book as reported here.
  • Wei Xian Woo is implementing session restore for Thunderbird as reported here.

Not a bad start at all! As we continue to make progress this week, more detailed updates will surface over on the Mozilla wiki. Some lessons from the weekend:

  • Thunderbird is massive, and the codebase can often feel like C written in C++: it’s peppered with gems like return codes and extra output parameters and pointers to C-style strings. Put briefly – this is not the code you saw in class. Deal with it.
  • It’s incredibly important to discard hubris in open-source development. Check your ego at the door; don’t be afraid to ask questions; understand that you will be wrong many times before you are right. On the other hand, IRC places hundreds of experts at your fingertips, and is therefore more powerful than almost any tool you could use.
  • Practical lessons: pbranchack, and MXR are your friends; ctags is also quite useful. If you don’t understand what to do, look at code that does something similar. If you still don’t understand what to do, ask. When using Mercurial, never forget: commit early, commit often.
  • Remember: have fun! As in any creative endeavour, passion keeps you going.

As a final note: I’d like to thank gvwilson for organizing the code sprint, which has proven to be an amazing opportunity to meet other talented and motivated developers. Additional shout-outs are due humph and bwinton for remaining diplomatic throughout the torrent of confused questions, and for providing helpful answers in return.

Happy hacking!

Posted in Sprint, Thunderbird | 1 Comment »

Day 3 of the Sprint

Posted by Brenda Sadoway on 2010/01/17

This weekend, we focused on finding the main aspects of the program that we think need working on and deciding where to go with each one.  Specifically, there was the simplification of toolbars and menus to cater to beginner programmers, comparing the Java versus Scheme parts of the project and approaching the open bugs that have already been identified.  Also we experimented with the program from a first year student standpoint to determine what is and isn’t necessary and went from there.

On Saturday, Dwight Deugo met with us and discussed the structure and aims of the project, a bit of the history of Eclipse, plus where we could go with it over the next few months.

We found and reported some bugs/tasks that are possibilities to look into this semester, as well as spent a lot of time looking through the code to become more familiar with the extensive framework of Eclipse.  This was a major task but will make things easier in the long run.

We intend to have weekly meetings on Fridays at 2:00 Eastern Time, using Skype as our main communication method.

Posted in Eclipse4Edu, Sprint | Tagged: , | Leave a Comment »

Code Sprint Day 1 &2 : Data Server Flight Simulator

Posted by Michael Man on 2010/01/17

Day 1 (At IBM Canada):

Activities:

  1. Welcome, Introductions, OSS Development talk, CLA
  2. Introduction to the TE by Peter Kohlmann
  3. IBM DB2, Technology Explorer (TE) & Workload Multiuser Driver (WMD) – setup on your systems
  4. DB2 architecture overview
  5. Walk through – basic tutorial & view creation  – Matthew Vandenbussche

Issues:

  1. There is a bug in db2 driver on Macbook, 3 of the team members are using Mac and had to switch to Windows/Linux

Day 2: (At University of Toronton)

Activities:

  1. Code work through – Matthew Vandenbussche
  2. Implementation Q&A
  3. Repository setup
    • https://db2mc.svn.sourceforge.net/svnroot/db2mc
    • Login with your sourceforge id and password to commit
    • Our mainline code is in UCOSP/mainline
    • Personal workspaces are in UCOSP/individuals/<your first name>
    • Work in your personal workspaces, deliver to mainline periodically (TBD)
  4. communication plan work setup
    • We’re using skype for group chat, everyone is in the same group now
  5. scrum meeting scheduled
    • Weekly meeting with IBM: Tuesday and/or Thursday at 1pm EST(conferencing id TBD)
TODO list for Day 3:
  1. Decide work items to meet first deliverable date
  2. Split the work
  3. Assign task to each individual
  4. Come up with our own version naming
  5. Request hardware resources(a testing server)

Posted in FlightSim, Sprint | Tagged: , | Leave a Comment »

First Push: Mercurial

Posted by alexandru on 2010/01/17

During this Code Sprint we focused on a few improvements to Mercurial that make it easier and safer to use. Each of the tasks we worked on was scoped such that only one developer was needed, but this gave everyone a chance to dive head-first into the Mercurial codebase at their own pace. For some (myself included) this is also the first serious project in Python, and these two days were a great opportunity to ask lots of questions.

We’re well on our way implementing support for using git sub-repositories, interactively editing Mercurial’s config files (.hgrc), cache annotations, auto-merging .hgtags, matching wildcards when converting from a different source-control system to Mercurial, and preventing users from using files whose names differ only in case (i.e. filename and FileName).

We’ve also discussed the harder problems we’re attempting over the next few months, namely fixing the way Mercurial handles history for large file renames, improving the way very large files are being stored, and improving hg serve output. And we’re sure there will be plenty more to do.

If you want to take a look at what’s going on in our world, please check out the UCOSP FogBugz install and log in with Guest/anonymous.

Posted in Mercurial, Sprint, Status | Tagged: , , | 1 Comment »

Code Sprinting with MarkUs: Day 2

Posted by Farah Juma on 2010/01/16

It’s been a productive day for the MarkUs team. We had a meeting this morning to discuss our plans for the semester. Also, review requests have already started to come in! We’re ready for a great semester!

Posted in Sprint, Uncategorized | Leave a Comment »

Code Sprint: Day 1

Posted by Greg Wilson on 2010/01/16

The first day of this term’s code sprint has come and gone. Donuts and pizza were eaten, coffee was drunk, people who’d only met virtually introduced themselves in person, and a whole lotta software was installed, explained, and fixed. (“You want it to run on a Mac? I don’t think it does that…”) Blake Winton (Thunderbird), Andrew Ross (Ingres), Karen Reid (University of Toronto), and Steven Case (Minnesota State University) spent most or all of the day with us, and we had a few other visitors as well:

  • Andrew Overhold and Deepak Bhole (Red Hat) and Ben Newton (First Media Group) came down to introduce themselves. Without sponsorship from their companies and others, the event wouldn’t have been possible; we’re very grateful for their support.
  • Dave Bolter, an accessibility expert at Mozilla, came in to talk and give demos.
  • Adam Goucher, our tame testing guru, showed off Selenium and talked about SLIME.
  • Peter Kohlmann, Matthew Vandenbuss, and the rest of the IBM team hosted the Data Centre Flight Simulator students up at their lab. We got ’em all back at the end of the day, though, so that’s OK 🙂

Thanks as well to the staff at Bright Pearl, who fed us all dinner (in Mike’s case, twice). Thanks also to Carolyn Ursabia for handling arrangements, and to Alan Rosenthal, our indefatigable sys admin, for making everything run so smoothly. Our second day starts in two and a half hours; I’m looking forward to seeing what everyone accomplishes.

Posted in Sprint | Leave a Comment »