<mbj>
snusnu: BTW the concept of terminated expressions rocks unparser.
<snusnu>
mbj: re the code, it can all be done without any instance, but it feels so ugly to stick everything inside class methods :)
<mbj>
snusnu: I support far more edgecases at a lower complexity level :D
<snusnu>
nice!
<mbj>
snusnu: After switching to the "termianted expression" model.
<mbj>
snusnu: Add support for comments, some spacing (empty line between methods) and %w() style literals. And unparser will soon be a nice prettifier.
<snusnu>
mbj: btw, i wanted to mention that quite a few times already, i strongly prefer %w[] over %w()
<mbj>
snusnu: finewithme
<snusnu>
mbj: makes it more obvious that we're dealing with an array
<mbj>
snusnu: never thought about it and %w[] does make more sense
<snusnu>
mbj: original suggestion is from, you guess who
<mbj>
snusnu: haha, yeah
<snusnu>
;)
<mbj>
snusnu: And unparser must emit "single quoted strings" if possible.
<snusnu>
mbj: yeah!
<snusnu>
mbj: fwiw, slightly improved version of the crypto code (less object creation): http://pastie.org/8301929
<mbj>
snusnu: Unparser now has "zero open issues", so its a good moment to hit bed.
<snusnu>
mbj: awesome, you've earned it :)
<mbj>
snusnu: I dont like #to_s here
<mbj>
snusnu: It is not clear if you access the input or the output of your method object.
<mbj>
snusnu: I have more thoughts, but they are to unstrucutred currently
<mbj>
snusnu: hitting bed, cu
mbj has quit [Quit: leaving]
<snusnu>
!memo mbj: decided to get rid of instances and renamed api methods: http://pastie.org/8302005
<snusnu>
lol
postmodern has quit [Quit: Leaving]
jfredett has joined #rom-rb
cored has joined #rom-rb
cored has quit [Ping timeout: 264 seconds]
jfredett has quit [Quit: Leaving.]
knowtheory has joined #rom-rb
snusnu has quit [Quit: Leaving.]
knowtheory has quit [Ping timeout: 245 seconds]
knowtheory has joined #rom-rb
knowtheory has quit [Quit: Computer has gone to sleep]
knowtheory has joined #rom-rb
knowtheory has quit [Ping timeout: 245 seconds]
postmodern has joined #rom-rb
lorenzo_off is now known as lorenzo_
V8Energy has quit [Ping timeout: 264 seconds]
mbj has joined #rom-rb
postmodern has quit [Quit: Leaving]
mbj has quit [Read error: Connection reset by peer]
knowtheory has joined #rom-rb
knowtheory has quit [Client Quit]
mbj has joined #rom-rb
knowtheory has joined #rom-rb
<mbj>
solnic: I think we should release devtools as a gem, and ship Gemfile.devtools via a gist.
<mbj>
solnic: You see any reason NOT to do so?
snusnu has joined #rom-rb
cored has joined #rom-rb
<solnic>
mbj: just one, it’s probably against github policy
<solnic>
we are not allowed to use github to distribute files
<solnic>
we should check it
<mbj>
solnic: we could add a project "gemfile-devtools" and use their download feature.
breakingthings has joined #rom-rb
snusnu has quit [Quit: Leaving.]
mbj has quit [Quit: Lost terminal]
jfredett has joined #rom-rb
snusnu has joined #rom-rb
indrek has quit [Ping timeout: 276 seconds]
indrek has joined #rom-rb
knowtheory has quit [Quit: Computer has gone to sleep]
V8Energy has joined #rom-rb
mkristian has joined #rom-rb
mbj has joined #rom-rb
mkristian has quit [Quit: Leaving]
<mbj>
snusnu: solnic also likes the idea of shipping Gemfile.devtools via another mechanism
<mbj>
snusnu: But we are unsure about githubs policy
<mbj>
snusnu: Im not sure if they like we use them as a "file host"
<snusnu>
mbj: if their policy doesn't allow us fetching it from a gist, i guess one simple "please let us do that for the sake of simplicity and oss" would probably be enough
<mbj>
snusnu: Yeah I expect
<mbj>
snusnu: You know someone to ask?
<snusnu>
mbj: but i agree, we should ask them
<snusnu>
mbj: not particularly, no
<snusnu>
mbj: i met schacon once and he was a fan of datamapper, he removed the api throttle for alfred back then
<snusnu>
mbj: but that's years ago
<mbj>
snusnu: hehe
<mbj>
snusnu: about to write an email to support@github.com
<snusnu>
mbj: sweet!
<mbj>
snusnu: I'd love to release devtools as gem
<mbj>
snusnu: And it would allow it to have deps!
<snusnu>
how so?
<snusnu>
i mean, how would it be any different?
<mbj>
snusnu: We dont have deps in devtools because it would complexify setting up a new project even more.
<snusnu>
mbj: no, we have no deps to avoid dep hell when bumping devtools deps
<mbj>
snusnu: It is no hell if devtools depends on "> 0.0.0"
<mbj>
Im addicted to coffeine, and dont have any beans here :(
<snusnu>
mbj: which would obviously be "wrong"
<snusnu>
mbj: or to be more precise, not inline with our policy of testing against latest releases only
<snusnu>
mbj: not saying it's no option tho, just stating our regular rules
<snusnu>
bbiab
snusnu has quit [Quit: Leaving.]
snusnu has joined #rom-rb
snusnu1 has joined #rom-rb
snusnu has quit [Ping timeout: 245 seconds]
mbj has quit [Ping timeout: 245 seconds]
mbj has joined #rom-rb
<mbj>
snusnu1: while writing the email I realized: Maybe we should release devtools containing Gemfile.devtools
<mbj>
snusnu1: My point is, we can do releases at any "consistend" snapshot.
<mbj>
snusnu1: It should be very easy for us to do "patchlevel" releases.
<snusnu1>
mbj: hmm i dunno, strictly speaking, devtools version doesn't change, as it will be rare that actual devtools code changes .. it'd leave us with literally shitloads of releases, a crazy version number eventually, no semantic version whatever … it's just not the right thing to do, well imo at least
<snusnu1>
mbj: if we release new devtools when Gemfile.devtools changes, devtools version number becomes essentially completely meaningless .. i see no reason why should go down that route
snusnu1 has quit [Quit: Leaving.]
snusnu has joined #rom-rb
<mbj>
snusnu: Because we'd have less moving parts.
<snusnu>
not a good reason imo
<mbj>
snusnu: I'm totally okay with having 0.1.$huge_number_here over time
<snusnu>
like at all, i don't like that at all
<mbj>
haha
<snusnu>
it will be worse once/if we pimp devtools
<snusnu>
it's just no good option
<mbj>
snusnu: As an aditional benefit, you can easily "downgrade" devtools in case you run in problems.
<mbj>
snusnu: Just use a more early gem and done.
<snusnu>
still not convinced, like, at all
<snusnu>
sorry
<mbj>
snusnu: With externalized Gemfile.devtools you maybe have to downgrade devtools and search an older version of Gemfile.devtools.
<snusnu>
so what, are we talking regular problems here? like, did that *ever* happen in the past?
<mbj>
snusnu: One could say, the dependency informaton *is* code :D
<snusnu>
to be clear, i have no problem versioning the Gemfile.devtools on it's own, and then maybe set a dep on that from devtools the gem, a wide one
<mbj>
snusnu: nice idea!
<snusnu>
but imo it's wrong to make devtools version number useless by itself
<mbj>
snusnu: I like your idea
<mbj>
snusnu: two gems
<mbj>
snusnu: one "devtools"
<mbj>
snusnu: and the other "devtools-gemfile" or whatever :D
<snusnu>
yeah, that might work
<mbj>
snusnu: You see, it is good I think loud.
<snusnu>
i just want to be able to follow a regular version scheme (semantic versioning) within the parts where we actually have *code*
<snusnu>
always! :)
<mbj>
snusnu: And that "devtools-gemfile" could just be a $format.$revision version number.
postmodern has joined #rom-rb
<snusnu>
yeah, stating clearly that there's no semantic versioning happening
<mbj>
solnic: you have a minute?
<mbj>
solnic: snusnu had the idea to use two gems, one "stable" devtools gem. And another maybe "devtools-gemfile"
<snusnu>
you know, my main problems with stuffing it all together bumping version numbers like crazy, would sorta prevent any kind of cool additions we'll have over time .. well, it wouldn't prevent them, but basically all advantages of (semantic) versioning would be lost for that code
<mbj>
snusnu: You already convinced me :D
<snusnu>
still thinking out loud tho :)
<mbj>
solnic: So we dont have to use any "new" infrastructure (a gist)
<mbj>
solnic: You see any problem with this approach?
<mbj>
solnic, snusnu: also the upgrade path is "dead simple".
<mbj>
Current users of devtools dont have to change anything.
<snusnu>
yeah, giving it a quick thought, i don't see any problems
<mbj>
Just for the ones like me, we just dont use the :git source.
<mbj>
bundle update && devtools sync will *still* work.