Conflicts:
plugins/category_generator.rb

I now support both forms of CASE expression. I changed the name of the
Case class to MatchCase a...

CASE is a fairly standard expression found on all database platforms.
There are some variants av...

Use the Github API repo sorting.

T-SQL variants use the modulus operator rather than MOD.

The Github API can sort by pushed time, no need to do it manually.
http://developer.github.com/v...

There was no reason to use regex when formatting output. I already know
the type of each token (...

There were really only two jobs being done for parsing tokens. The first
was tokenizing a comman...

Some functions support an optional window clause: COUNT(1) OVER
(PARTITION BY Blah1 ORDER BY Bla...

In some places I was limiting it to select expressions (no order by
clause). However, it appears...

I did some performance tuning on the parser to avoid determining a
token's type multiple times. ...

I finished implementing the CommandBuilder class. I am concerned with
some of the performance is...

There's no reason to recompute the SqlGrammar's regex.

I have the first edition of the the CommandBuilder implementation for
creating ISelectBuilder in...

I am in the middle of building the command builders.

I realized that my precedence rules were m...

I decided that Top can accept any IProjectionItem. This eliminated the
need for the IArithmetic ...

I've always debated supporting EXISTS. I finally decided to add support,
if not just to handle S...

Decided to forego worrying about fancy output.
Made generation an adhoc process with free-form f...

I decided that the way code is currently generated is WAY too
complicated. I started converting ...

All of the commands now support table aliasing.
I now support @, : and ? in identifiers, for the...

I changed the grammar to handle aliases in DELETE statements. I fixed
some whitespacing issues.

I wanted to make sure the correct SQL was generated when placing columns
throughout SQL.
- Fixed...

I want the ability to use the same token registry to both tokenize an
input string and detect to...

I made sure the grammar properly handled logic and arithmetic
precedence.
I fixed some ambiguiti...

I broke out the Parser class into a Parser and ParseAttempt class. This
allows for late binding ...

I am in the process of generating the actual SQL as part of traversing
the AST. I am creating cl...

Namespace:
The Schema object was replaced with a more generic Namespace class that
supports as m...

I realized that I was moving towards a general-purpose parsing tool when
I starting thinking AST...

Giving a more helpful error message instead of silently failing.

In order to simplify the parser, I defined the TokenType enumeration. I
had to go back and pass ...

I was able to make some progress in parsing the schema that is returned
from the expression tree.

added async switch = TRUE to loader.
According to Google PageSpeed the Facebook JS was not load...

Making it so the build doesn't fail if a bad lexer is passed to the code highlighting.

I am working towards being able to correctly identify where I am while
parsing the expression tree.

In order to determine how to process the tokens, I am providing type
information. So, for instan...

In order to make it easier to format the output, I needed to restructure
the generated expressio...

I am trying to generalize formatting. As I went through trying to format
strings, the code becam...

More convenience functions...
More formatting options...
More realization that manually formatti...

I am still adjusting the formatting code to properly handle indentation.
I am trying to think of...

I decided that it didn't make sense how WHERE clauses were built. For
all the other clauses ther...

Whitespace

Because the .nuspec file was not in the project directory, it was not
being picked up by NuGet.

I decided it would be a lot easier to just specify the parameter prefix
as part of the parameter...

This has all of the code from the Earthworm project, plus some
modifications to make it easier t...

Escape github repo descriptions, as they may contain HTML.

This should help resolve issues for zsh users and make it easier to enter titles which freak out ...

robots.txt generation

Extract category_link filter from category_generator.rb plugin

Rakefile option for any additional rsync arguments

To support url without domain name like /path/file

Added variable `rsync_args`

Fixing English error pet-peeve.

You need to know if you would like to support this, each word in the title can not start with "/".