![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
TMI: Things that are not debugging
Jun. 19th, 2012 08:49 pmToday was full of things that took time away from debugging, but that's okay, because there sure has been a lot of debugging lately.
First, the weekly meeting, during which we agreed on a solution to issue 2633, yay! And quite a lot of type system discussion.
In my limited coding time, I did do something exciting, which was solve the "nested class" problem from yesterday. It turned out the solution to the "duplicate definition" error was that my fix affected one part of resolve that deals with scopes, but another part adds things to an index and that was already correct, so my fix caused entries for the same class constructor to be added multiple times. I fixed that and make check ran with my comm changes... all the way through. That branch hadn't compiled all the way through in quite a while. But then I needed a new snapshot (I'd made one locally, but to check in code I need ones for all the platforms built), and as an aside, despite over a year of Rust being self-hosting, I have to read these instructions every single time... it's bad. In any case, something went not-quite-right with the snapshot and I'm a bit hazy on the details now, since I ended up working on other stuff for the rest of the day.
I went to Brian (former and future Mozilla Research intern who was visiting)'s talk on Timelapse, which was excellent, and it was good to go to a talk since I don't usually go to them... something about how my brain doesn't work that way. But when it's a good speaker like Brian who also puts in a lot of effort into delivery, I can get something out of it. And after that, we went to lunch... for two hours. Hey, we don't have visitors every day.
After lunch I worked on the POPL paper a bit; I added an example that I think will tie the "intro to Rust" section together with some more work. Then there was a mixer for current interns and ex-interns-who-became-fulltimers, but there was a mix-up with the food so I left after a while and worked on the paper some more. This paper-writing thing is strange. I haven't worked on a conference paper in months and months (and if you don't count the typestate paper draft that didn't get very far, I haven't worked on one in almost two years), but working on a paper again seemed totally natural. But hopefully someone likes my contributions besides me.
First, the weekly meeting, during which we agreed on a solution to issue 2633, yay! And quite a lot of type system discussion.
In my limited coding time, I did do something exciting, which was solve the "nested class" problem from yesterday. It turned out the solution to the "duplicate definition" error was that my fix affected one part of resolve that deals with scopes, but another part adds things to an index and that was already correct, so my fix caused entries for the same class constructor to be added multiple times. I fixed that and make check ran with my comm changes... all the way through. That branch hadn't compiled all the way through in quite a while. But then I needed a new snapshot (I'd made one locally, but to check in code I need ones for all the platforms built), and as an aside, despite over a year of Rust being self-hosting, I have to read these instructions every single time... it's bad. In any case, something went not-quite-right with the snapshot and I'm a bit hazy on the details now, since I ended up working on other stuff for the rest of the day.
I went to Brian (former and future Mozilla Research intern who was visiting)'s talk on Timelapse, which was excellent, and it was good to go to a talk since I don't usually go to them... something about how my brain doesn't work that way. But when it's a good speaker like Brian who also puts in a lot of effort into delivery, I can get something out of it. And after that, we went to lunch... for two hours. Hey, we don't have visitors every day.
After lunch I worked on the POPL paper a bit; I added an example that I think will tie the "intro to Rust" section together with some more work. Then there was a mixer for current interns and ex-interns-who-became-fulltimers, but there was a mix-up with the food so I left after a while and worked on the paper some more. This paper-writing thing is strange. I haven't worked on a conference paper in months and months (and if you don't count the typestate paper draft that didn't get very far, I haven't worked on one in almost two years), but working on a paper again seemed totally natural. But hopefully someone likes my contributions besides me.