Today (29-10-2009) marked the Day 2 of SubConf 2009 in which there were plans for many talks by prominent subversion developers and others. Three parallel sessions on various Software Configuration Management related topics went on simultaneously, out of which this report captures only the English sessions that happened in Room:Madrid of the Conference Hotel.
10:00 am – 10:45 am
Keynote — by C Michael Pilato, one of the long term developers in the subversion project right from 2001 till date
The Keynote title read as “The Subversion Legacy (So Far) – Philosophical Musings of a grateful participant”. CMike walked through the past history of subversion, the development methodology, why CollabNet chose subversion to be run as an Open Source Project, how it all evolved, etc.
The following were some of the questions posted by audience at the end of the talk and responses from CMike:
Q1. Why do we have SubConf for Subversion?
Ans: SubConf is good for subversion because it helps users from all over the world to meet in such a great user conference in which we don’t sell anything (it is all there out for free), but get user’s feedback and experiences to make it a more happy ride in the future.
Q2. What does Subversion Community’s feeling about DVCS?
Ans: The Subversion Community is excited about Distributed Version Control Systems (DVCS), since we are part of advancing the “State of the Art” and we are happy that, ultimately we have competitors in the version control world :)
Q3. Is there any Subversion User Conference like this planned in the US?
Ans: There is none but you can get together with Subversion folks in other conferences such as the ApacheCon.
Q4. What is the number of developers working on Subversion?
Ans: Developers come and go, at any point of time we have from 10 to 20 developers actively looking at the Subversion code base.
Q5. Is there a dictator in the Subversion project?
Ans: No we don’t have any “Dictator”.
11:15 am – 12:00 pm
Subversion Release Process — Hyrum Wright (Release manager of Subversion project and President of Subversion Corporation) and Stefan Sperling (Subversion Developer)
This talk explained in detail about the various process involved in making a subversion release. Another focus of this talk was to get more people help with testing of subversion pre-releases which will help in catching and fixing bugs early, though we don’t want you to run trunk on your production data. Right now we have very limited buildbots, so we encourage people to donate boxes for the subversion project if they have any so that it can be used to run some tests remotely and collect results. We need to remove redundancy in tests and introduce more real time tests as the users face in BIG deployments.
The following are some of the questions posted by audience at the end of the talk and responses:
Q1. Subversion community can talk with the universities to get help from their Software Engineering and Software Testing departments to test the pre releases of subversion by the students, which will be very useful.
Ans: Yes that is an excellent idea and made a note.
Q2. How to keep track of bugs that are found and fixed between the releases of Subversion?
Ans: They are available from the CHANGES and branch based STATUS files.
Q3. How does a bug fix go from a release branch to trunk in subversion code base?
Ans: We do it the other way :) by backporting fixes from trunk to the release branches.