karlseguin’s gists

## gist:c264863573a6bfbf7605
loadArticle (res, req) ->
  article = yield findArticlebyId(...)
  return res.notFound() unless artilce?
  res.ok(article)

findArticleById (id) ->
    conn = yield conn.Open().
    row = yield conn.QueryAsync("....", id)
    new Article(row)

## gist:3fbe1aa749b1e58c485e
func LoadArticle(res http.ResponseWriter, req *http.Request) {
  if article, err := FindArticleById(...)
  if err != nil {
    handleError(res, err)
    return
  }
  if article == nil {
   handleNotFound(res, err)
  }
}

## gist:0ba24030fb12b10b686b
/*
bolt      5000	    277963 ns/op
redis    30000	     48081 ns/op
pg       10000	    149691 ns/op

Yes, the Bolt transactions could be batched. But so too could the PG transactions,
and the Redis work could be pipelined. And that isn't always a workable solution.
*/

import (

## init.coffee
# place this in your init.coffee
atom.workspace.observeTextEditors (e) ->
  editor = atom.views.getView(e)
  clone = editor.querySelector('.wrap-guide').cloneNode()
  clone.style.left = (editor.getDefaultCharacterWidth() * 120) + "px"
  editor.querySelector('.underlayer').appendChild(clone)

## gist:a6b441269d82690381e5
/*
Playing with building a specialized set for integers. Specifically:
- Integers are natually randomly distributed (auto-incrementing ids for millions of records would do)
- The number of items is known upfront (can be approximate)
- The set sizes changes infrequently (we're mostly interested in membership tests)

Accomodating constraints, to be sure, but if we can meet the first two requirements,
we'll end up with a little less than 1/2 the memory of a map[int]struct{}, and membership
tests taking about ~1/2 the time.

## gist:5a6a45ace2048545b6c3
   BYTE UNIX Benchmarks (Version 5.1.3)

   System: p2.kapi.io: GNU/Linux
   OS: GNU/Linux -- 3.13.0-36-generic -- #63-Ubuntu SMP Wed Sep 3 21:30:07 UTC 2014
   Machine: x86_64 (x86_64)
   Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
   CPU 0: Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz (6799.8 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
   CPU 1: Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz (6799.8 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization

## gist:a659ef87b3a4a5d590e9
   BYTE UNIX Benchmarks (Version 5.1.3)

   System: ip-172-30-0-152: GNU/Linux
   OS: GNU/Linux -- 3.13.0-36-generic -- #63-Ubuntu SMP Wed Sep 3 21:30:07 UTC 2014
   Machine: x86_64 (x86_64)
   Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
   CPU 0: Intel(R) Xeon(R) CPU E5-2666 v3 @ 2.90GHz (5800.1 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
   CPU 1: Intel(R) Xeon(R) CPU E5-2666 v3 @ 2.90GHz (5800.1 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET

## gist:0f38b9ad87ebd54375da
 BYTE UNIX Benchmarks (Version 5.1.3)

   System: ip-172-30-0-254: GNU/Linux
   OS: GNU/Linux -- 3.13.0-36-generic -- #63-Ubuntu SMP Wed Sep 3 21:30:07 UTC 2014
   Machine: x86_64 (x86_64)
   Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
   CPU 0: Intel(R) Xeon(R) CPU E5-2666 v3 @ 2.90GHz (5786.7 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
   CPU 1: Intel(R) Xeon(R) CPU E5-2666 v3 @ 2.90GHz (5786.7 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET

## bench.go
package main

// @see https://twitter.com/karlseguin/status/524452778093977600
import (
	"math/rand"
	"strconv"
	"testing"
)

const (

## gist:9686555
type PgAutoPrepare struct {
	driver.Conn
	cache map[string]driver.Stmt
}

func (pap *PgAutoPrepare) Open(name string) (driver.Conn, error) {
	conn, err := pq.Open(name)
	if err != nil {
		return nil, err
	}