CraigBuchek has joined #rom-rb
snusnu has quit [Quit: Leaving.]
dkubb has joined #rom-rb
CraigBuchek has quit [Read error: Connection reset by peer]
CraigBuchek has joined #rom-rb
skade has joined #rom-rb
skade has quit [Quit: Computer has gone to sleep.]
skade has joined #rom-rb
robertjpayne has joined #rom-rb
CraigBuchek has quit [Quit: CraigBuchek]
robertjpayne has quit [Remote host closed the connection]
jfredett-w has quit [Ping timeout: 240 seconds]
kleech has joined #rom-rb
kleech has quit [Remote host closed the connection]
kleech has joined #rom-rb
kleech has quit [Ping timeout: 245 seconds]
kleech has joined #rom-rb
kleech has quit [Remote host closed the connection]
kleech has joined #rom-rb
kleech has quit [Ping timeout: 240 seconds]
onewheelskyward has quit [Ping timeout: 252 seconds]
dkubb has joined #rom-rb
snusnu has joined #rom-rb
onewheelskyward has joined #rom-rb
kenphused has quit [Ping timeout: 240 seconds]
CraigBuchek has joined #rom-rb
jfredett-w has joined #rom-rb
mbj has joined #rom-rb
<
mbj>
dkubb, snusnu: Experiences from Kenya cellphone network: 12k ms round trip time, stdvar 6k ms. Totally overrun ;)
<
dkubb>
mbj: whoa, you mean 6 second ping times?
<
mbj>
dkubb: No 6 seconds standard variance ;) 12 seconds ping in mean ;)
<
dkubb>
I hope you don't plan on playing any first person shooters
<
mbj>
dkubb: They have HUGHE bufferbloat in their network.
<
mbj>
dkubb: So TCP algorithms on both sides dont see package drops and increase window size swaming the lines even more
<
mbj>
dkubb: And because they have HUGHE queue in the cellphone towers TCP ramps up window size.
<
mbj>
dkubb: I'd love to hit their network engenieers with a stick ;)
<
mbj>
dkubb: And no, me is not a gamer.
<
mbj>
dkubb: Over the day I had 3% package drops + this huge RTT I think bufferbloat is the root of the problem.
<
dkubb>
mbj: I'm at github issue zero on my main axiom-* repos
<
dkubb>
oh wait, I spoke too soon.. axiom-types has 1 issue
<
dkubb>
it's an easy one though
<
mbj>
Hah, I have far to much issues. I'd love to go 0 issues after my stay here.
<
mbj>
But I also have more active repos than you ;)
<
mbj>
dkubb: I also had 3 blackouts this day, lasting at max 3 hours ;)
<
mbj>
dkubb: But cellphone towers and my notebook are powered from battery / generator.
<
mbj>
And my child is so excited from "blackout" ;)
<
mbj>
I should do a blog: "Coding from Kenya - Experiences with crappy infrastructure"
mbj has quit [Ping timeout: 240 seconds]
mbj has joined #rom-rb
mbj has quit [Ping timeout: 240 seconds]
mbj has joined #rom-rb
<
mbj>
dkubb: Back online on other carrier
<
mbj>
dkubb: With much better delivery
<
mbj>
dkubb: BTW 1GB is 2€ in prepaid data plans.
<
mbj>
dkubb: In germany you pay much more.
<
mbj>
dkubb: Also im now on 3g and not EDGE/HSDPA
<
dkubb>
mbj: so I'm going to finish up wrap/unwrap and then try to do some work on sql
<
dkubb>
mbj: unless you think think of something more pressing
<
mbj>
dkubb: Still fighting infrastructure, tomorrow is a "full free day". Today we had a wedding.
<
dkubb>
oh, and I'm also going to refactor Nest and Wrap to use a Nesting class internally to handle the nesting.. the main ops are the same
<
mbj>
dkubb: But speaking about planning: Yeah we can start with SQL.
<
mbj>
dkubb: not my wedding, some friends wedding.
<
dkubb>
I may also rename nest/unnest to group/ungroup to be closer to the RA
<
mbj>
dkubb: A kenyan wedding is actually a very enjoable ceremony.
<
mbj>
dkubb: I'd love to add group/ungroup support to axiom arango adpater.
<
mbj>
dkubb: But first lets add a generic axiom-ast in between. Porting my axiom-sexp thing to produce an axiom-ast should be MOST easy.
<
dkubb>
I thought about the confusion, and I think it will only be momentary, vs continuous confusion with RA documents and other libraries
<
dkubb>
sure, I can work on that if you want to take on the sql stuff first
<
dkubb>
or we can work on both together
<
dkubb>
I just want to minimize contention
<
mbj>
dkubb: I think transforming axiom-ast to sql-ast is many times more easy than transforming axiom-relation to sql-ast.
<
dkubb>
axiom is an AST so it should be easy to adapt a #to_ast method for each kind of ops
<
mbj>
dkubb: So I'll wor here.
<
dkubb>
mbj: should we do any round-trip testing?
<
dkubb>
axiom -> ast and ast -> axiom
<
mbj>
dkubb: Yeah, sure.
<
dkubb>
it'd be nice to update axiom-fuzzer to work with the new ast too
<
mbj>
dkubb: Dunno if I really like #to_ast
<
dkubb>
but that's not as critical. I would want to use it after getting the basics
<
dkubb>
mbj: it's part of the ast interface
<
mbj>
Especially when we plan to use whitequark/ast internally sometime.
<
dkubb>
mbj: every node responds to #to_ast. it makes sense to have Axiom::Relation#to_ast
<
mbj>
dkubb: Sure, but somehow I dislike all the #to_something stuff.
<
mbj>
dkubb: As an entrypoit +1
<
mbj>
dkubb: But the implementaiton should be "external"
<
dkubb>
but if it's part of the ast it should be alright
<
mbj>
Axiom::Ast::Generator.call(self)
<
dkubb>
yeah, I would mix it in from axiom-ast
<
dkubb>
axiom-sexp will probabl go away
<
dkubb>
or it'll just add to_sexp and do to_ast.to_sexp
<
mbj>
axiom-sexp was a "debug thing2
<
dkubb>
yeah, that makes sense
<
mbj>
Many times more easy to follow than Axiom::Relation#inspect
<
mbj>
whitequark/ast also has nice pretty printing build in
<
dkubb>
I want to add a table outputter
<
dkubb>
for seeing the actual data
<
dkubb>
ascii table I mean
<
mbj>
dkubb: Yeah, makes sense for materialized relations
<
mbj>
axiom-sexps prints the data also, but not as "pretty"
<
dkubb>
I like how alf uses the table output as inputs too, that's kinda neat
<
dkubb>
it allows him to make small cli tools that allow piping of the data
<
dkubb>
ok, bbiab, gotta take my daughter to a birthday party
mbj has quit [Ping timeout: 248 seconds]
skade has quit [Quit: Computer has gone to sleep.]
zekefast has joined #rom-rb
skade has joined #rom-rb
skade has quit [Quit: Computer has gone to sleep.]
zekefast has quit [Quit: Leaving.]