<travis-ci>
[travis-ci] dkubb/axiom-types#143 (master - c8a2f0b : Dan Kubb): The build was broken.
travis-ci has left #rom-rb [#rom-rb]
kleech has joined #rom-rb
kleech has quit [Remote host closed the connection]
kleech has joined #rom-rb
kleech has quit [Remote host closed the connection]
kleech has joined #rom-rb
snusnu1 has joined #rom-rb
snusnu1 has quit [Client Quit]
kleech has quit [Remote host closed the connection]
kleech has joined #rom-rb
snusnu1 has joined #rom-rb
snusnu1 has quit [Client Quit]
snusnu has joined #rom-rb
kleech has quit [Ping timeout: 248 seconds]
kapowaz has quit [Ping timeout: 260 seconds]
Gibheer has joined #rom-rb
mbj has joined #rom-rb
snusnu has quit [Quit: Leaving.]
zekefast has joined #rom-rb
kapowaz has joined #rom-rb
mbj has quit [Ping timeout: 248 seconds]
zekefast has quit [Read error: Operation timed out]
knowtheory has joined #rom-rb
kleech has joined #rom-rb
kleech has quit [Remote host closed the connection]
mbj has joined #rom-rb
mbj has quit [Ping timeout: 252 seconds]
mbj has joined #rom-rb
jfredett-iv has joined #rom-rb
knowtheory has quit [Ping timeout: 252 seconds]
knowtheory has joined #rom-rb
dkubb has joined #rom-rb
<mbj>
dkubb: hola
<dkubb>
mbj: good morning
<dkubb>
solnic: I fixed that remaining bug with axiom-types
jfredett has quit [Quit: Leaving.]
<jfredett-iv>
mbj: how's mutcov coming? I've got a project w/ 100% mutation coverage and no badge wherewith to brag...
<mbj>
jfredett-iv: I'm busy with clients work. Than I'll take a 4 week break dedicated to OSS.
<mbj>
jfredett-iv: mutcov is such a simple thing, I should be able to get it up in a few days.
<mbj>
jfredett-iv: I'tll start as a simple "Badge for OSS" service.
<jfredett-iv>
mbj: cool
<jfredett-iv>
mbj: I'm a sucker for OSS Badges...
<mbj>
jfredett-iv: And I'll limit it to "n" badges. Mostly to see if there is a demand for a commercial service.
<mbj>
jfredett-iv: You maybe should follow me on twitter, I'll advertise once that thing goes online.
<jfredett-iv>
I should follow the whole gang here, honestly -- I forget about twitter a lot...
<mbj>
jfredett-iv: heh
<mbj>
jfredett-iv: I had to be "pressed" into twitter by the gang. But with a resonable choice of people to follow it works well for me.
<mbj>
jfredett-iv: Some kind of a "distributed fourier transform" filtering the noise away.
<mbj>
I use twitter exclusively for "coding conversations" and I only follow people doing the same. So I dont end up in noise.
<jfredett-iv>
yah, I probably should purge my twitter -- I was spoiled by tweetdeck + filtering, so I ended up being a bit to liberal with my follows...
<mbj>
hehe
<mbj>
jfredett-iv: The idea of a limit like "n" OSS project badges is simple.
<mbj>
jfredett-iv: I have limited resources for OSS, and I think that badge service should mirror it.
<mbj>
jfredett-iv: Doing rom, mutant, ... is lots of work and limiting the numbers of badges I give out helps me to singnal "hey that guy has limited resources" ;)
<jfredett-iv>
*nod*, makes sense.
<mbj>
I'll try to make the use of a mutcov super simple
<mbj>
no account management
<mbj>
not even a github connection.
<mbj>
Use the mutcov gem under travis/circle and you are done.
<jfredett-iv>
I like how codeclimate does that (more or less, they need the GH connection, but only inasmuch as the repo name)
<jfredett-iv>
same w/ coveralls
<mbj>
coveralls ui is imho confusing, such as gemnasium
<mbj>
Also I'm not a GUI person ;)
<jfredett-iv>
it's quite confusing, but all I really want out of it is the badge...
<jfredett-iv>
and the number, more generally
<mbj>
yeah
<mbj>
I think I can do some authentication/authorization via encoding your repo name with the GH/circle crypo api (if available for OSS)
<mbj>
So nobody can setup a sutpid curl script to submit any value ;)
lgierth has joined #rom-rb
<jfredett-iv>
This library is 110% mutation covered...
<jfredett-iv>
Chaos
<mbj>
?
<mbj>
ah, heh ;)
<mbj>
I'm doing web security reviews/work in my job, so I should not produce a very obvious problem in my public services ;)
<jfredett-iv>
probably a good idea...
<mbj>
jfredett-iv: BTW there is a #mutant channel on freenode.
<mbj>
jfredett-iv: I'm totally okay you post here and ask. (That #mutant thing is really silent, currently)
<mbj>
But we have to start filling it somehow ;)
<jfredett-iv>
joined.
<mbj>
jfredett-iv: I just dont want to scare away #rom insterested folks from this channel, with mutant talk.
<jfredett-iv>
no worries, I'm exercising my 'ii' wrapper, so joining more channels is good, I keep finding little bugs.
<mbj>
hehe
lgierth has quit [Quit: Ex-Chat]
postmodern has joined #rom-rb
snusnu has joined #rom-rb
<dkubb>
jfredett-iv: I've often thought about making a clean delegate library that is mutation covered
<dkubb>
the whole point being that you're routing a message to some other receiver
<dkubb>
I've always disliked the forwardable and delegator interfaces in stdlib
<jfredett-iv>
I've never much liked the interface either (mostly #def_delegator), but I do like the idea
<jfredett-iv>
I don't mind the SimpleDelegator(SomeClass) so much, but the non-classtemplate version (ie, using #__setobj__ or w/e it is) is toast.
<dkubb>
I like the idea of the delegator stuff. some of the interface is odd and the internals aren't so nice in delegator and forwardable
<jfredett-iv>
Never really dug into the internals (always been a little scared to) :)
<dkubb>
mutant kinds of forces you to have clean and simple internals because otherwise it punishes you since you now need to write 2-5x more tests for everything
<jfredett-iv>
I've noticed, that, it's also forced me into changing structure (generally for the better) that I otherwise wouldn't have changed
<jfredett-iv>
Though in a couple places I walked away thinking, 'Normally I'd be lazy here and be okay w/ not speccing that a #foo? method always returns a boolean, but I guess it won't hurt.'
<jfredett-iv>
Mutant testing reminds me a lot of having types, tbh. Same sort of 'sometimes I don't agree with you, mr. typechecker, but I respect you, so I'll do it anyway'
<dkubb>
I think in the end we'll find that in a few cases mutant does force you to write unnecessary tests, but it's a worthwhile trade off for all the tests that it makes you write that you should've been writing all along
<dkubb>
see, in that specific case, you'd have to mutation test all those logic branches for each delegated method
<dkubb>
heh, I've had the same feeling, but I haven't had as much experience with languages having a good type checker
<dkubb>
mostly haskell, and then pretty limited.. maybe a few aggregate weeks of experience
<dkubb>
but I've had mutant point out real, actual bugs so many times I tend to trust it now
<jfredett-iv>
It's not caught many for me yet, but I have only been truly diligent about using it on this one project so far, and it's definitely found some areas I probably wouldn't have wanted to come back to.
<dkubb>
one thing I've found is that it makes me a better spec writer. I'm able to see and spec way more cases up-front
<dkubb>
I don't yet know if it's worth it on all projects, but I definately like using it on foundational libraries
<dkubb>
otoh, I have used it on a work project with good results so far
<jfredett-iv>
I made the pitch to poor fanfare a few weeks ago, 'Our suite is already too slow, why would we add more time to it?'
<dkubb>
heh
<dkubb>
one thing that's nice is to run it in isolation with new code
<jfredett-iv>
I ran it on our business-model gem for this one client, 15% covered, mind = blown.
<jfredett-iv>
The scary thing is -- when I ran it on older checkouts it was quite a bit higher, it was just code drift, new stuff snuck in during refactoring or w/e
<jfredett-iv>
that's the real thing I'm interested in, there've been a few times where bugs snuck in becasue of drift, mutant seems like a nice way to eliminate (at least partially) that class of bugs
travis-ci has joined #rom-rb
travis-ci has left #rom-rb [#rom-rb]
<travis-ci>
[travis-ci] dkubb/axiom#239 (nested-relation - 962012e : Dan Kubb): The build was broken.