#dat

ogd: should probably add jbenet/transformer jbenet/data jbenet/datadex to dat-git-bot?

jbenet: transformer looks like its growing up fast!

jbenet: you might wanna change require(path.join(__dirname, 'filename.js')) to require('./filename') to make it browserifyable

@mafintosh thanks :D

@mafintosh: won't that break windows compatibility?

fff. got used slack. mafintosh ^

jbenet: no node is really *really* clever about paths

jbenet: stuff like fs.readFile('./foo/bar/baz') also works

jbenet: and path.join('c:\\foo', 'foo', 'bar/baz') also returns kinda what you would expect

mafintosh: ah, nice. so, why is path used all over the place?

jbenet: Its to support custom backends i would assume. since you need to require them somewhere !== __dirname

ogd: o/

mafintosh: oh i thought only ./foo.js worked but not ./foo/bar/baz, but i dont relaly know

jbenet: dat-git-bot is just the github irc hook, you can configure it from your repo settings

jbenet: you can only do it 1 repo at a time

mafintosh: hey whats the status on the appveyor integration? did you want me to take a stab at it?

jbenet: according to https://www.npmjs.org/doc/misc/npm-disputes.html i should wait another couple of weeks for an answer before I can claim the unused module

ogd: feel weird about taking someone else's name. another reason i'm all for namespacing :) -- it may be ok to use `dat-transformer`

jbenet: nah it's normal to take unused modules, but only if they are inactive (this one is)

it has no activity in 2 years and no dependents

ogd: I'm only got so far as finding out that it should be easy to do. fell free to start out - i'll can join in and contribute.

mafintosh: im thinking we should postpone the fully automated builds for now

mafintosh: and instead just do them by hand

mafintosh: (on windows)

ogd: good idea. we can just build them in the vms

actually probably should just do them by hand on all machines

I'm doing that all weekend. Finally got some free time :)

we should probably just have a shell script that dumps the build somewhere and we can hand upload them to gh-releases or whatever

*dat

haha

ogd mafintosh: that's what i was doing for data. I have scripts that fire up the vms, compile, and copy the files. Look at https://github.com/jbenet/data/tree/master/plat... it doesn't have windows yet, but the makefiles do it all automatically.

later we can get travis/appveyor to publish releases for us but i'd like to deprioritize that for now

My brain does auto correct for me.

Ah!!! ogd: that's where that link was! I couldn't find it in my email. :]

(todo part has link to windows vms)

http://xdissent.github.io/ievms/ worked well for me to installign the modern.ie stuff

but havent seen anything to fully automate booting + building

jbenet: those vagrantfiles look useful

it had a more general repo here: https://github.com/jbenet/platform-vms the idea is clone `platform-vms` into your project (like into `data/platforms`) and then modify as needed. any general stuff can be pushed back upstream.

ogd: the makefile automatically (a) builds in every platform, (b) extracts the binaries, (c) packages the tarballs with a readme, ready to drag + drop to github.

mafintosh o/

whoa perfect

:D

`tar.README.md` is what's put into the tarball. it's a template, has a simple script there that filled in the values. can just port that to js.

ogd: happy to help with this, if you show me an example of how the builds work right now, i can `vm-platformize` dat.

we only need it for leveldown + leveldown-hyper actually as the rest of dat is just JS

i think mafintosh has an example of building leveldown

Sorry. A bit slow. Danish culture dictates that we spend the entire friday drinking beer :)

lol

jbenet: check out my node-leveldown fork

ogd: ah cool. https://github.com/mafintosh/node-leveldown/ i assume?

mafintosh: are you before, at or after the ballmer peak

hey did you guys see that I got github to change javascript to yellow? muahahah