Will’s Miro dev blog

Figured I’d send this out early because I’m not going to get anything more accomplished.

I spent Wednesday, Thursday and Friday:

• triaging old bugs
• continuing to clean up comment spam left over from when our bug system was managed by Trac
• fixing my Windows build environment problem (bug 9327)
• fixing the problem with Miro on Windows in the 1.0 branch (bug 9363)
• checked in a fix from a downstream reported problem where Miro on GTKX11 doesn’t honor the –xine-driver option (bug 9373, r5897, r5898)
• started making adjustments to our build scripts to allow us to tag with “Miro-1.1″ instead of “Democracy-Player-With-White-Wine-Sauce-1.1″ which is too long and I’m sick of typing it
• set up a pbuilder environment for testing Ubuntu/Debian packages and continuing to make our packages better

That about covers the last three days. pbuilder is really cool, but takes forever to run because it’s building a fresh environment in which ti build everything and it does that by downloading and installing all the packages each iteration. It’s definitely a good idea to do this, though, as it’ll increase the confidence in our Ubuntu/Debian packaging and potentially fix outstanding issues.

I’m doing a whirlwind tour of family visiting over the next few days, then going down to Louisiana on a service trip to rebuild houses. I’ll be back online January 1st.

I submitted a new patch for bug 303645 on the 12th. On the 13th, I started having problems seeing the feed preview page. I talked with the folks on #develop, but wasn’t able to get it working. No clue what the problem is. I did a fresh checkout on a different machine and that works fine, so I continued working on the second machine and produced a better version of the 303645 patch. I’m gone next week and between that and the time I’m going to have to put into the 1.1 release, I don’t think I will get a patch together for the rest of the changes for bug 400059.

Firefox takes me about an hour to compile from scratch. While trying to figure out what my feed preview page problem was, I spent some time using the reports feature in Bugzilla and doing bug triage on old bugs. I also wrote a script to remove comment spam from the comments we migrated from our old Trac bug-tracker. There’s still a lot of comment spam in there. Ick.

I merged the libtorrent changes from trunk to the 1.0 branch and tested out the GTKX11 platform. I went through Ubuntu Gutsy, Gutsy-64, Feisty and Dapper, made sure Miro in the 1.0 branch compiled and updated the requirements as listed in GTKX11BuildDocs.

I tried to test the 1.0 branch of Miro in Windows, but I’m hitting the LIBEAY32.dll problem in bug 9327 and the fix suggested in the comments isn’t working for me. So… my Windows build environment isn’t working again and I don’t know whether the 1.0 branch is stable or not after the merges I did.

That about covers me for this last week.

In the next few days, I plan to help out with the 1.1 release, make sure I have the Ubuntu platforms covered and continue working on the Firefox patch.

I’m gone from December 23rd through December 31st. After December 22nd, I won’t be online again until January 1st.

It’s been a good-ish week.

Nathan finished up his changes for bug 9077, they were merged into the trunk and 1.0 branches, and Janet did some testing and came up with bugs 9334 and 9335. I fixed 9334 but I think I’m going to skip 9335–it involves changing the padding for those items in the item view and I’m not wildly excited about doing that and I think the issue is cosmetic. We only implemented CC metadata at the item scope–not the feed scope, yet.

I worked my way through some bonehead issues I had caused, finished up the patch for bug 303645 and submitted it. I’m a little apprehensive about submitting a patch to Firefox, but … I’ll suck it up. The important thing is that this patch populates the enclosures array for each FeedEntry item. That was a pre-requisite for bug 400059. I’m working on that one now. As a side note, the folks on #developer on the Mozilla IRC channel have been really helpful.

On Friday, I went to lunch with Dean, Chris Blizzard and John Resig–that was really neat. A little hard to quell the star-struck feelings–hopefully I didn’t make a total ass of myself.

I created the “other packages” page for the download section of the getmiro web-site and changed around the download page, too.

I sent an email to Justin at Mozilla in response to his blog entry. I pointed him to the code for the timeline script and the script for migrating data from Trac to Bugzilla.

Looking forward to a Miro 1.1 release with all its libtorrent and CreativeCommons metadata goodness….

I’ve been super busy this past week.

I spent the brunt of my time on the Firefox patch. I’ve solved most of bug 303645, though the output isn’t pretty and it doesn’t support Yahoo’s MRSS or iTunes enclosures. I think I’ll have it figured out in the next couple of days and submitted to the Mozilla folks. Once I’ve populated the enclosures and they’re available in the FeedWriter, I can work on what we really want to do. I still need to figure out how to associate applications with different feed types. Mozilla froze the trunk today for the upcoming beta 2, so none of the changes I’ve done will be available until beta 3 at the earliest.

I spent several hours looking for enhancements or bugs that lend themselves to being small 1-5 day tasks for high school/college level people in the PSF section of the GHOP. I didn’t find any that I thought were promising. This is a bit unfortunate as it’d give us some good exposure, gets us some help, and would be good for the project. Still, there’s a certain amount of work that would need to be done to be part of the GHOP. The contest lasts until February, so I’m hoping I can figure something out before the end.

On Friday, Dean and I talked with Henri of CivicActions. They work with clients who produce content and we talked about various directions our groups can take to help each other. He’s particularly interested in how Miro could interact with mobile and embedded devices. I’m pretty interested in that, too. I plan on working on that when I get a Nokia n810. I’ve written about that in previous posts.

Also, I’ve been working with Nathan of Creative Commons to get Miro to understand and work with licensing metadata (bug 9077). He’s done most of the work so far; I’ve been providing feedback and working out the implementation issues. This is really useful since it allows content producers to embed licensing data in the feeds that Miro will display to users viewing the content.

And I’ve spent some time doing bug triage and talking with users about various issues, mostly related to packaging.

Current deadlines:

The CreativeCommons birthday is 12/15, so we need to have the 9077 work done by then.

The Mozilla folks told us that we need to get the patch done before the end of the month.

We were thinking of doing a Miro 1.1 release mid-month. I want to fix the packaging scripts so that we can name tags and branches Miro-x.y instead of Democracy-Player-x.y. I’ll probably look into that later this week.