Fork me on GitHub
#ldnclj
<
2015-07-15
>
jamiei09:07:55

I know it's grey/raining, but it can't be that bad, surely @mccraigmccraig? simple_smile

mccraigmccraig09:07:53

my daughters nursery showed me her 'learny journey' book today, which they compile from their notes and photos... they looked at me funny when i asked them which of the 3 obvious spelling variants they used 'lourney journey', 'learny jearny', 'learny journey'

korny09:07:08

re: SQL - well, my words were prophetic, we found a performance issue yesterday because a simple Hibernate lazy join wasn’t actually lazy; it would have been a trivial custom SQL query but Hibernate still fscks it up.

xlevus10:07:36

oh neat. I was trying to build a thing that looks like it was probably just going to end up being 90% of what yada is. I can now probably just write a shim around yada.

mccraigmccraig10:07:48

korny: hibernate is a different beast to korma & honeysql though... shoehorning a relational model into OO was always an idea of dubious merit

mccraigmccraig10:07:38

xlevus: yeah, i'm playing with yada now, looks like i'm gonna throw away my half-assed solutions

xlevus10:07:59

throw away those half-asses. And implement yourself some fresh half-asses!

mccraigmccraig10:07:02

all solutions are asymptotic to a whole ass

korny10:07:47

mccraigmccraig: true, but I’m still rather skeptical. I know that if I’m using SQL under the covers, the time will probably come where I need to tweak some of the SQL because of something particular to the particular SQL flavour I’m using; and when that happens, I don’t want to find that my SQL is being generated, unless it’s a real simple generation like coffeescript is to JavaScript.

korny10:07:57

… but maybe it’s just PTSD and someone will produce something that proves me wrong simple_smile

mccraigmccraig11:07:42

korny: ha, i think we've all got ptsd from a variety of causes... yesql meets your constraints, in that it's basically just a sql template. korma and honeysql less so, but then i tend to only use these things now when i want to be able to compose query fragments, and i've got at least as much ptsd from writing query-composition stuff as i have from fighting restrictive sql generation simple_smile

xlevus11:07:18

I don't quite get yesql. It doesn't feel like it fits in with the clojure way.

malcolmsparks16:07:04

@mccraigmccraig: yada 0.5.x isn't that far off, I've had quite a bit of holiday/travel time to hack on it recently and I expect it will be the 1.0

mccraigmccraig16:07:52

@malcolmsparks: i'm playing with it now... no problems so far, though i haven't done a lot yet

malcolmsparks16:07:05

i've totally redesigned some parts of the interface however, so even though it's close to 0.4.2 there's some significant differences

malcolmsparks16:07:17

especially with content negotiation, which has been redesigned from the ground up

mccraigmccraig16:07:02

@malcolmsparks: should i use the 0.5.x branch then ?

malcolmsparks16:07:56

not yet, i'm still fixing up some tests,

malcolmsparks16:07:23

but i'd appreciate any comments, initial impressions

malcolmsparks16:07:46

I also need to update the docs, which will be quite an overhaul - probably next week

malcolmsparks16:07:09

use the 0.4.2 master for now, then upgrade to 0.5.x in about 2 weeks

malcolmsparks16:07:22

it won't be a major amount of work, depending on how many endpoints you're creating

malcolmsparks16:07:41

if there's anything you find to be an issue with 0.4.2, do let me know

mccraigmccraig16:07:43

great, i'll just carry on then simple_smile