[dtaht/deBloat] 5e9fe7: Start at latency under load test and related suppo...

[noreply at github.com]

  Branch: refs/heads/master
  Home:   https://github.com/dtaht/deBloat

  Commit: 5e9fe71c15df860860683f161e962aa9a530f383
      https://github.com/dtaht/deBloat/commit/5e9fe71c15df860860683f161e962aa9a530f383
  Author: Dave Taht <d at taht.net>
  Date:   2011-12-31 (Sat, 31 Dec 2011)

  Changed paths:
    M .gitignore
  A test/ab.gnu
  M test/db.lua
  M test/db.sql
  A test/editor.lua
  A test/gnuplot.org
  A test/latency_load.lua
  A test/multiplot.gnu
  M test/netperf.lua
  A test/ping2.gnu
  A test/pingplot.gnu
  M test/pingplotter.lua
  A test/random.gnu
  A test/testtikz.gnu
  A test/tree.lua

  Log Message:
  -----------
  Start at latency under load test and related support

Some do top down design, some bottom up,
some do both at the same time and try to meet in the middle.

My own rule is if I need a piece of code more than 3 times,
make it general purpose. I work in four phases when working
without a specification, but understanding the task at hand.

1) top down,bottom up,middle out

2) get something that works

3) tag it as a working version

then refactor, refactor, refactor.

Git has changed how I work significantly in that I do like
to make commits as often as I can. Unlike some, I like
to see the mistakes I made, reflected in the repository
so I don't make them again.

I am suffering mostly from not grokking how oo works in lua,
as well as argv handling. I have thoughts towards making the
testbed work at the command line, in emacs, and on the web, and
that complicates matters somewhat.

I'm not happy with gnuplot. Nor do I want to be dependent
on google charts - as cool as they are - because most of the
time I'm offline while doing a test.

* Added a bunch of gnuplot examples to refer to
* Revised sql handling somewhat
* revised db somewhat
* wrote stubs for netperf, editing, iperf, ab
* wrote stub for latency under load test