apinstein

#!/bin/bash
# RESULTS, feel free to repro yourself
#
# noLock                    FAIL
# accessShare               FAIL
# rowShare                  FAIL
# rowExclusive              FAIL
# shareUpdateExclusive      SUCCESS
# share                     FAIL+DEADLOCKS
# shareRowExclusive         SUCCESS

apinstein

# debug performance stuff
SetEnvIf Request_URI ^/panel panel-request
LogFormat "%h %l %u %t %T \"%r\" %>s %b \"%{Referer}i\" \"%{User-agent}i\"" debug_performance
CustomLog /Users/alanpinstein/dev/sandbox/virtualtour/log/debug_performance_log debug_performance env=panel-request

apinstein

#!/bin/sh
# This should result in very different completion times for the 2 runs.
# This script is useful to be sure that ionice works on your kernel since it's possible to have the
# ionice command yet using it has no effect
# you can check your scheduler with something like 
# cat /sys/block/sda1/queue/scheduler

# NOTE: I wrote this to help me figure out if ionice was working on my system. It doesn't seem to be, since the
# output of this script indicates that the priorities, while adjusted, don't get reflected in the real world timings.
# I don't know if my script is wrong or if my server is mis-configured.

apinstein

How S3 Permission Work
- The AWS account that creates a bucket owns it.
- The owner of a bucket can never be changed.
  - All billing for object usage goes to bucket owner account by default. That's one reason ownership cannot be changed.
- Note that objects in the bucket can have permissions that would prevent even the bucket owner from editing/deleting it.
- There are three styles of permissions:
  1. Bucket Policies  
    - Allows access control to be specified for AWS Accounts or IAM Users
    - Specified in Access Policy Language
      - Can DENY or ALLOW

apinstein

#!/bin/sh
# Try this... kinda a hack but better than any other alternative I know of:

for i in `netstat -an | grep EST | awk '{ print $5; }' | sort | uniq | grep -v 127.0.0.1 | grep -v ::1 | sed -e 's/\([0-9]*\.[0-9]*\.[0-9]*\.[0-9]*\).*/\1/'`; echo -n "Checking IP: $i => " && dig +short -x $i && echo -n " PROCESS: " && (lsof -i | grep $i) && echo ""

apinstein

zend_extension=/path/to/xdebug.so
; enable starting debug with XDEBUG_SESSION_START=1
xdebug.remote_enable=1
; enable starting profiler with XDEBUG_PROFILE=1
xdebug.profiler_enable_trigger=1
; good idea to do this explicitly b/c it's hard to tell where it went otherwise
xdebug.profiler_output_dir=/tmp/
; To generate TRACES of script execution; see http://www.xdebug.org/docs/execution_trace
; this will happen ALWAYS (or set to 0 and use xdebug_start_trace; sadly no XDEBUG_XXX method to enable this via URL)
xdebug.auto_trace=0

apinstein

// Magic-ness to only run the save callback after no saves have been issued for a while
var executeAfter = function(f, ms)
{
    var timer;
    var wrapper = function()
    {
        var passedArguments = arguments;
        if (timer)
        {
            window.clearTimeout(timer);

apinstein

// navigator.geolocation.getAccuratePosition() is an improved version of navigator.geolocation.getCurrentPosition()
// 
// getAccuratePosition() is designed and tested for iOS.
//
// The problem with getCurrentPosition() is that it returns an inaccurate position even with "enableHighAccuracy" enabled.
// The problem with watchPosition() is that it calls the success handler every second, and it is resource-intensive.
// 
// getAccuratePosition() calls the callback only once, but uses watchLocation() internally to obtain a position that meets your accuracy needs.
// If the timeout is exceeded before a position meeting your accuracy is needed, the best position is returned to the
// success callback, and the error callback does not fire. If you really care about accuracy, you should check it

apinstein

After a lot of different workflows, we've finally settled on this workflow for using git internally. It offers several benefits and avoid a bunch of pitfalls we were running into. It isn't perfect for everyone, of course, but it's worked very well for us.

Benefits:
- Makes it easy to share a topic branch among multiple developers without lots of conflicts and rebase hell.
- Preserves logical history of "started working on topic at this point, merged it into master at this point).
- Still allows for frequent rebasing against master to ensure you stay up-to-date with the mainline.

---> master
 +---> topic-integration (shared across users via github)
   +----> topic-devA (local branch in devA's repo)

apinstein

#!/bin/sh

# shell script quoting problem demonstration

# I need to be able to set a shell variable with a command with some options, like so
PHP_COMMAND="php -d 'include_path=/path/with spaces/dir'"
# then use PHP_COMMAND to run something in another script, like this:
$PHP_COMMAND -r 'echo get_include_path();'
# the above fails when executed. However, if you copy/paste the put from this line and run it in the CLI, it works!
echo "$PHP_COMMAND -r 'echo get_include_path();'"