Zachary R. Dev Log

Trying out wordpress

2011-04-10T20:52:00.000-04:00

This is basically just a post with a link so I don't forget about making the blog:
http://yabwritt.wordpress.com/

Emacs Review

2011-03-15T16:09:00.001-04:00

Emacs -- "The extensible self-documenting editor"
This is all I knew about emacs after deciding to install emacs23.

I started with the emacs tutorial, after finishing that I had a basic
knowledge of how to navigate and an knowledge of the concepts of emacs,
one very interesting concept emacs has is the modes, essentially turning
a dumb text editor into a word processor, modes are also very helpful for
having programming language specific settings.

On the note of modes one very-very useful mode is org-mode which
is the best way to organize notes and thoughts I have used.

I also very much enjoy emacs features like a plugin that detects
tab style/settings automatically(which makes it simple to edit other's code).

Here are the links to the plugins I mentioned:
http://www.emacswiki.org/emacs/GuessStyle
http://orgmode.org/

So basically emacs is great, oh and this was written with emacs.

My ZSH Configuration

2011-03-02T12:03:00.003-05:00

This is my ZSH configuration.

Thoughts on C++

2010-09-21T12:02:00.001-04:00

I have been learning/using C++ for just a few days so far, I have been learning from "Thinking in C++ Vol. 1", which seems to explain the language nicely.

I really like C++ so far it seems less tedious than C (obviously because C++ is intended to be a bit higher level than C) and overall most of the code I have written has compiled the first time around nicely (whether it works as intended is another question).

Release 0.1 of AsciiEditor

2010-09-03T21:12:00.000-04:00

Here is release 0.1 of AsciiEditor, a tiny software project I have been working on over the last few days.

Download: http://github.com/Mach1723/ascii_editor/downloads

The readme:

An ncurses based image stlye editor for ASCII art.
Requirements:
Python 2.5 or newer(may not work on Python 3)
Curses libraries.
Running:
At a terminal invoke: python asciieditor.py

To run this on windows I reccomend:
cygwin with python, ruby and the terminator terminal

Controls:
m for movement mode
p for paint mode
e for erase mode
c followed by any character to change the paintchar
s to save a file called 'out' which contains the current drawing area
r to read the 'out' file.

Arrow keys for movement

Image Version Control, How Could it be Done?

2010-07-31T23:39:00.003-04:00

How could you create a version control system ( like git, svn or cvs ) for images?
A special text representation of an image would need to be made, a file describing every pixel, and a diff format describing an associated change ( i.e resize, color change etc. ).

The version control system would need integration into an image editor(i.e gimp) and would work something like this:

User 1 changes image, saves image as an special text format, this format would list every pixel, its associated color, operations done on it(resize, color change whatever)
User 2 changes same image(after User 1 commits) when user 2 commits the file would be changed as normal
if User 1 and User 2 commit to the same file, whoever is reponsible for the repo would need to create a merge(if neccesary) which would invoke an supported image editor which would be used to make whatever changes are neccesary to represent both User 1 and User 2's changes(if possible)

Considering images aren't code, user commited changes would be no where near as easy to accept and would be more of a guideline than an exact way of handling things.

This post is very early ideas.

The Design of the Unix Operating System Notes 1

2010-07-04T17:50:00.005-04:00

These are my notes from "The Design of the UNIX Operating System" book these notes are incomplete, i will add more as i write them, though writing them is going slowly.

Chapter 1( General Overview of the System) :

The system is mostly written in a high level language(like C).

It is simple but powerful.

It provides simple libraries( and programs) to code more complex software with.

It is consistent with data, for example, even if a file may be stored in different locations on disk, the kernel makes it look like a stream of bytes to applications.

It is easy to port to other platforms, because of its very few machine dependent parts.

Programs in *nix have no notion of the internal format that the kernel stores file data in, they treat the data as an unformatted stream of bytes.

References for Chapter 1:

http://en.wikipedia.org/wiki/Unix_philosophy

Chapter 2( Introduction to the Kernel) :

The two major parts of a *nix kernel( and any kernel) are the: process control subsystem and the file subsystem.

The system call library(sys/*.h) provides functions that look like ordinary C function calls, the system libraries map these functions to system calls.

Functions from the system libraries elevate the process to kernel level for the time they are running.

Every library will use system calls, although most library's will use the standard library not system calls directly.

The file subsystem manages files, allocates file space, administers free space, controls file access, and gets file data for users.

Processes interact with the file subsystem via a set of calls, on *nix systems these calls are usually, open( returns a file descriptor for R/W ), close, read, write, stat ( gets file attributes ), chown, and chmod.

The file subsystem accesses file data using a buffering mechanism, which regulates data flow between the kernel and the storage device.

The buffering mechanism interacts with block device drivers to start data transfer to and from the kernel.

Device drivers are kernel modules that control peripheral devices.

Block devices are random access storage devices, the file subsystem also interacts directly with raw I/O device drivers, without using the buffer mechanism.

Raw devices, are also referred to as character devices.
Character devices are all devices that are not block devices.

The process control subsystem is responsible for process synchronization, IPC, memory management, and process scheduling.

The file subsystem and process control subsystem work together when loading an executable file into memory for execution.

Some system calls for controlling processes are fork, exec, exit, wait( waits for state changes in child process) , brk and signal.

The memory management module controls memory allocation, if the system does not have enough memory for all processes the kernel moves them between physical memory, and secondary storage ( usually a hard drive's swap partition ), so that all processes get a chance to execute, this is referred to as swapping.

The internal representation of a file is given by an inode, which contains a description of the disk layout of the data and data such as file data , access permissions and access times.

When a process attempts to access a file by name, the kernel has to check that it has permission to access the directories the file is in and retrieves the inode if it does.

Inodes are stored in the filesystem, but when the kernel is manipulating them the kernel loads them into a an kernel space buffer, known as the inode table.

The kernel has two other file structures known as the "file table" and the "user file descriptor" table.

The file table is an global kernel structure, the user file descriptor table is allocated per process.

When a process opens a file the kernel allocates a entry from the file table, user file descriptor table and the inode table.

The file table keeps track of byte offset in the file, to determine where the next write/read will start, and access permissions allowed to the process.

The user file descriptor table keeps track of what files are opened by the process.

A file system ( your partitions ) has the following layout, the boot block which contains bootloader code for booting( or is empty if this is not the boot partition ), the super block which describes: how large the file system is, how many files it can hold, and where free space can be found, the inode list which has a number of inodes that is specified at filesystem creation time, and the top of the inode list is the root inode[1], which branches out to inodes below it, the data blocks start at the end of the inode list, a data block can only be allocated to one file in the system.

When source code is compiled into an executable it is formed of several parts:

headers that describe the file attributes
the program text
a machine language representation of data that has inital values when the program starts, and an indication of how much space the kernel should allocate for uninitialized data called bss[2]( which is initalized to 0).
other misc sections.

Processes have two stacks, the userspace stack and the kernel space stack.

A stack frame ( in userspace ). Contains the current function, the function parameters, the functions local variables and the data needed to go back to the previous stack frame, including the value of the program counter and stack pointer at the time of the function call.

A stack frame ( in kernelspace ). When a system call is run, the process switches to kernel space, executes the kernel code and uses the kernel stack, the kernel stack's stack frames are for functions executing in kernel mode, the data in the kernel stack is for the system functions and data in the kernel, it is nulleduserspace stack

Every process has an entry in the kernel process table, and each process is allocated an u area[3] that contains private data only changed by the kernel, the process table points per process region table, whose entries point to entries in a region table.

A region is a continuous area of a processes address space, such as text, data and stack.

Region table entries describe the attributes of a region, such as whether its shared or private, and where the data of the region is located in memory.

The extra level of indirection ( from the per-process region table to the region table ) allows processes to share regions.

The u area contains information that is only needed when executing, such as:

a pointer to the process table slot of the current process
parameters of the current system call, return value and error codes.
file descriptors for all of the open files.
internal I/O parameters
current directory/current root
process and file size limits.

The kernel can access the u area of the currently executing process, but not the u area of other processes.

The lifetime of a process can be divided into several states, the important ones to remember are:

The process is executing in kernel mode
The process is executing in user mode
The process is not executing , but is ready to run as soon as the scheduler chooses it.
The process is sleeping, a process puts itself to sleep when waiting on an event, like waiting for I/O to complete.

When a process is sleeping the kernel doesn't check the process constantly to see if it is able to use whatever resource it is sleeping for, sleeping processes are said to be waiting for an event, so if resource A becomes available and process B needs it, the kernel checks to see if any processes need to use resource A and if process B is a higher priority than other processes it wakes up process B.

[1] this inode is the root directory of the system aka "/"

[2] http://en.wikipedia.org/wiki/.bss

[3] user area

References for Chapter 2:
http://sourceware.org/gdb/current/onlinedocs/gdbint/All-About-Stack-Frames.html
http://en.wikipedia.org/wiki/Stack_frame
http://en.wikipedia.org/wiki/Stack_(data_structure)
http://en.wikipedia.org/wiki/.bss

Chapter 3 ( The Buffer Cache ):

Chapter 3.0:

The kernel could write and read, directly to and from the disk, but due to slow disk response rates this would be impractical.

Instead the kernel minimizes disk access rates by keeping a pool of internal data buffers, known as the buffer cache, which contains the data of recently asked data blocks. [1]

Chapter 3.1:
During system initialization the kernel allocates space for a number of buffers according to memory size and system performance constraints.

A buffer consists of two parts, a memory array that contains the data from the disk and a buffer header that identifies the buffer.

The data in a buffer corresponds to the data in a disk block on the file system, and the kernel identifies the buffer contents by examining identifier fields in the buffer header.

The buffer is the in-memory copy of data from a disk block, a disk block can never be mapped into more than one buffer at a time, because the kernel has no way of knowing which data is newer.

The buffer header contains a device number field and block number field, to allow it to write back to the block of data on the disk. [2]

The buffer header also contains a pointer a buffers data array who's size must be at least as large as the disk block, and a status field that indicates the buffer's current status.

Chapter 3.2:

1 - And in more modern OS's i think, the buffer cache might have a bit more intelligence to what files it keeps in cache.

2 - device number field refers to a logical disk(i.e /dev/sda0 ), not a physical disk number.

Links related to the linux kernel

2010-06-27T22:17:00.000-04:00

Okay this is a collection of links i need to read on the linux kernel.

http://www.redhat.com/magazine/001nov04/features/vm/
http://www.ibm.com/developerworks/linux/library/l-linux-slab-allocator/
http://www.ibm.com/developerworks/linux/library/l-linuxboot/index.html

Short Slackware 13.1 Review

2010-06-23T12:33:00.001-04:00

Introduction:
Slackware is a Linux distro i have read of from time to time, and I figured I would try installing it( on a secondary computer ), heres a quick review of it.

Installation:
The installation process went smoothly and I learned how to use fdisk, i really love the way slackware groups its core packages, i do some basic coding, and its nice to have basically everything i need/want.

Slackware groups its packages by letters, D contains all of the development packages, and the default installer just installs every package in each chosen group, I chose to customize it( via the "menu" option ).

First Boot:

On restarting the computer you get a nice looking boot screen with a Slackware logo and the menu option "Linux", once I got to a login prompt, I logged in as root and created my user account, of course most distro's create a user account for you but thats another thing I love about Slackware is it doesn't assume much, so the configuration is easier to customize.

Networking Setup:
I had quite a bit of trouble setting up wireless networking, i figured out the kernel module for my wireless card was loaded, and installed "wicd"( from my install disk in /extras ), wicd makes configuration much easier, but the problem was my modem, I ended up changing WEP encryption to WPA( in the modem ) and it worked.

KDE Setup:
Since Slackware is fairly generic and installs a lot of applications, I ended up having a lot of unecessary menu entries to cleanup.

In my experience with KDE (OpenSUSE and Kubuntu), I always regarded it as way too slow and shiny, but it actually runs better than GNOME on this computer, its great and has some very nice options.

Foreword:
Overall Slackware is great, its not the most simple or easy distro, but its highly customizable and a great learning experience, it is wonderful for servers, desktops and any other computer and contains around 5gb of useful software to use.

identi.ca

2010-06-10T14:02:00.001-04:00

This will be the last mini-post on this site( big/relevant posts only
now) since i have decided to try out identi.ca for all my random tiny stuff.

Links related to operating system development

2010-06-10T11:35:00.000-04:00

Some good threads(stack overflow) on os development, which is related to the fact that i might be developing a OS as a learning project soon.

http://stackoverflow.com/questions/43180/how-to-get-started-in-operating-system-development/

Blender and Build Note

2010-06-10T08:41:00.005-04:00

Blender is a great software.. and interesting for anyone who is
interested in graphics development
http://www.blender.org/development/
http://wiki.blender.org/index.php/Dev:Doc/New_Dev_Info
http://wiki.blender.org/index.php/Dev:Doc/Building_Blender

Unfortunately it doesn't work on ubuntu 10.04 in my case.

Another collab software

2010-06-09T19:57:00.001-04:00

http://gobby.0x539.de/trac/

globby seems alot more useful for smaller teams because it doesn't
require a dedicated server(though you do need to do some port forwarding)

Some of the best GNU articles

2010-06-09T18:16:00.001-04:00

A little list for my own references of some of the best articles written
at GNU

http://www.gnu.org/philosophy/not-ipr.html
http://www.gnu.org/philosophy/misinterpreting-copyright.html
http://www.gnu.org/philosophy/selling-exceptions.html
http://www.gnu.org/philosophy/danger-of-software-patents.html
http://www.fsf.org/blogs/community/war-on-sharing-riaa-lawsuits

Etherpad

2010-06-09T17:45:00.001-04:00

http://etherpad.org/

Just found this very interesting Free Software.

Stupid Idea: Userspace Kernel Sim

2010-06-09T11:06:00.000-04:00

Idea: In C code a simulation of a kernel, *nix like with all the basic
concepts, like i said stupid idea and maybe impossible but it might be a
fun learning project.

A bug from a bug (bugs #591359 and #591388)

2010-06-08T14:36:00.002-04:00

I was attempting to confirm bug #591359, and when i tryed to install
libsdl1.2debian-all, it marked ubuntu-desktop for removal, i have seen
quite a few bug reports lately that involve ubuntu-desktop being
uninstalled, this is obviously part of a larger problem, but anyway i
filed a new bug report against libsdl1.2.
Links:
http://bugs.launchpad.net/ubuntu/+source/rhythmbox/+bug/591359 - #591359
http://bugs.launchpad.net/ubuntu/+source/libsdl1.2/+bug/591388 - #591388

http://bugs.launchpad.net/~f-fzfx - and since the bug report has been closed due to a lack of info, i am sticking this here so i can remember to see if this guy re-reports.

[DONE]TODO: Check /var/log/backup_script.log after cron job finishes

2010-06-08T11:18:00.002-04:00

After the lftp cron job finishes, i need to check its logfile and see if
any thing was written there.

A great tutorial on lftp backup

2010-06-08T10:38:00.001-04:00

http://www.linux.com/archive/feed/122169

This is a great tutorial on lftp backup, and using blogger in some kind
of notepad/twitter context is pretty effective.

[DONE]TODO: Setup a home dir to remote ftp cron job

2010-06-07T22:14:00.001-04:00

Need to setup a home dir backup to ftp cron job using lftp perhaps,
going to document this project here.

LFTP's manpage, the useful command here would be "mirror", along with other things.

http://ec2-75-101-156-227.compute-1.amazonaws.com/man1/lftp.1.php

A great vim cheatsheet

2010-06-07T20:06:00.001-04:00

http://www.cmake.org/Wiki/VIM_Useful_Commands

Vim is fun but you can't remember everything.

irssi notes

2010-06-07T17:26:00.001-04:00

Threw away pidgin's irc interface in favor of irssi, i have learned how to
connect and other basic things.

Links:
http://scripts.irssi.org/
http://quadpoint.org/articles/irssi

Perl Quiz of the Week

2010-06-07T14:06:00.001-04:00

As a sort of tradition I go looking for some good challenges/quiz's when
I start learning a new language, as an extra note I try and learn new
programming languages all the time, but I have stuck with C and python
more than others, I think I am definitely sticking with Perl.
http://perl.plover.com/qotw/#archives

EDIT: Turns out that site is old.. still has challenges to do though.

Test post via email, also some interesting knowledge about gvim

2010-06-06T22:37:00.001-04:00

This post is mainly a test post via email.

It would also be interesting to note that gvim/vim can run a command
internally by executing :!commandname

Perl and Bug #588974

2010-06-06T22:28:00.000-04:00

Perl is a interesting language, haven't learned much beyond you basic variables and arrays and the like, a note below.

The "$," is a perl predefined variable it is referred to as the output field seperator and is added between array || list elements

Now for Bug #588974, Bug #588974 is a bug concerning gscan2pdf and is basically the first new bug i have tried to triage sofar i have:

Confirmed it happens in both the ubuntu repo version and the newer 0.2.30 version.

Looked at the source to see if the fix was obvious (not at all)

Reported upstream in its debian bug tracker

The bug is that the quit without saving alert quits irregardless of what option you choose, i couldn't imagine the developer never saw this so maybe its a tough fix? As a side note this bug also made me want to look into learning Perl.

Bug #588974