Kevux Git Server - koopa/log

Cleanup: make sure stylesheet/external/ and related directories are ignored by git repository

THe stylesheet/external/ directory will be for storing 3rd-party (aka: external) css.

I am currently investigating different css libraries to include, such as:
- full_calendar
- jquery_mobile
- html5_shiv

The same goes for the images/external/ directory.

Cleanup: make sure javascript/external/ directory is ignored by git repository

THe javascript/external/ directory will be for storing 3rd-party (aka: external) javascript.

I am currently investigating different javascript libraries to include, such as:
- fullcalendar
- jquery_mobile

Cleanup: re-structure the database sql for standard and reservation

To avoid having duplicate queries, I moved all the permission related information for standard into a separate sql file.
- This allows for referencing the standard sql files as part of the install, without having to modify them for name changes.

The only files that should need to be replaced from standard are standard-first.sql, standard-permissions.sql, and standard-last.sql.

I changed all references of u_standard and r_standard to u_reservation and r_reservation to generate the reservation specific versions of standard-first.sql, standard-permissions.sql, and standard-last.
- These changes were saved in their respective reservation equivalent files: reservation-first.sql, reservation-permissions.sql, and reservation-last.

While making these changes, I noticed that there is a way to disable all triggers for a table so that I can alter content without a trigger being 'triggered'.
- This made it easy for me to move some special cases to the standard-permissions.sql file.

Progress: standard path design, database accounts, database logging, and other fixes

Move all of the execute parameters into the c_standard_path class so that it does not have to be passed to every function.
- Down the road, I may just have the execution function without parameters and use a separate function for assigning the parameters to the class.

Make sure the database accounts exist in the user table.
- Ldap synchronization on login is now functioning.
- Non-LDAP does not synchronize or change account settings on login.
- There is currently no 3rd-party database/structure/design to use as a basis for auto-assigning roles, therefore roles must be manually assigned to the each user.
- This manual assignment will still auto-update the postgresql roles, so only the is_* columns need to be altered and postgresql roles will automatically follow via triggers.

The first part of the database logging is setup and working.
- There is much more work to do.

Other fixes and changes.

Cleanup: do not explicitly specify public when referencing v_users_self_locked_not

Cleanup: remove some debug code accidentally committed

Progress: log tables and related code out of sync and other fixes

The log settings became noticably out of date between the code and the database.
Significant changes are made, dealing with the database and its usage.
Switched to timezone that supports timezone by default and then set standard behavior to use UTC timezone.

Bugfix: incorrect session name in use

Progress: move code from reservation to standard and other fixes

I've now determined what would be a good basis for using a 'standard' structure.
Standard would be a minimal website used by most of my projects here, but it is not part of the base so that other developers do not have to use it at all.
This allows for more development flexibility while still providing a common default.

Most of the reservation example program code was moved into standard, including sql code.

Cleanup: use 'minute' and 'second' (without s) for time- css class

Progress: more development with paths, additional html processing,

More work with paths.
Add additional data to HTML, restructuring code as necessary.
- Now includes class names by date and path (assigned to <body>).
- Add sanitizer, I used drupal 7's css sanitizer as a reference (just their preg_replace() part).

Functions like get_is_root() is a bit awkward, change the behavior and name to be more suitable.
- get_is_root() and set_is_root() are both now is_root() such that is_root() functions as get_is_root() when no parameters are specified and functions as set_is_root() when parameters are specified.
- this is repeated for all other similar functions.

I decided to separate /a/ paths into two paths:
- /a/ - administrative paths (for advanced and technical administration).
- /m/ - management paths (which is where most of the non-user related website logged in access will be involved in).

Other fixes and changes.

Update: put logged in status inside of c_base_session and cleanup function ordering

Progress: work on http request processing, path processing, and other fixes

Process the HTTP request methods, such as GET, POST, HEAD, OPTIONS, etc...
Add validation against these request methods on a per path basis.

Improve path processing logic (there is still more work to do).

Other fixes and improvements.
I specifically added an nginx example configuration file (it still requires separate PHP-fcgi configuration).

Bugfix: http request uri and http response location fixes

The request uri should include the scheme and authority where possible.
Fix a syntax error containing '$$' instead of '$'.

The response location needs to fully support an array.

Bugfix: uri processing code is not handling authority correctly

The authority is now being processed after scheme is found in a url.
The authority now properly returns success when a forward slash is found.

Update: implement custom language support in regular paths in the same way internal paths implement custom languages

Bugfix: redirect arguments in c_reservation_path_user_login() are incorrect

Update: add s_return() as a non-valued alternative to s_value() for error handling

Bugfix: internal pages were not processing language after redesign and use fix paths output handling

When I wrote the language handling into the initial setting load function and then changed the output handling to be more dynamic, I forgot to update p_include_path().
Get rid of the loop when deciding the language include file suffix and only use the last entry in the aliases array (generally the shortest).
There were some cases where I was still using $this->html instead of $this->content.
Be more consistent by using $this->output instead of $this->content.

Progress: more path management changes, imrpove languages support, and other fixes

Continue developing path handling functionality.
I've decided to implement language handling as sub-classes, using japanese as the first language to test with.
Other bugfixes and tweaks.

Progress: designing path handling, designing form processing, and other fixes.

Path management is done via classes.
The form processing code is just a rough draft of ideas and is subject to massive changes.

Update: add basic documentation on my url path naming practices

Progress: improve language support when using global data, add headers in some tests programs, and fix language related issues

When I moved languages into the global defaults class, I missed a few chunks of code that needed to be updated.
I noticed that some of the tests were not printing any headers.
Allow for setting content-language with multiple languages, as is the case with one of the tests using japanese and english.

Update: implement and use more global values, simplify http response return values, implement content-range, other changes/fixes

The time functions have been moved into the global default class.
The http language class has been moved into the global default class.

The return values for HTTP response functions have been set to their specific type, instead of always FALSE.
This makes it easier to handle the return value without caring about errors.
Any unsupported functionality continues to return FALSE until such time I read the appropriate part of the standards and implement the functionality.

Content-Range response is now implemented.

I noticed that I have been inconsistently using $ps_ and $s_ for private static class variables.
I removed $ps_ that in favor of $s_ only because it was the simpler change.
I may end up coming back later and changing all private static to $ps_ (or not..).

While making the time values global, I noticed that PHP does not always guarantee microseconds to be populated.
Special handling is necessary and so I created more global static functions to handle this complicated situation.

I've observed and fixed some simple mistakes, bugs, typos, etc...

Progress: work on path handling, other minor changes

I changed my approach half-way through this and will need to come back and clean out a lot of unecessary code.
Some decisions still need to be made on certain aspects.

Progress: more work on path libary, fix problems with some return values

Did a small amount of grunt work in setting up the path class.

I noticed that I had, in a good number of cases, forgot to include the 'new' reserved word when returning c_base_return_false().

Cleanup: set the default examples and program settings for connecting to the database to non-ssl

This simplifies the work necessary for testing and initial setups.
Setting up a cert is extra work, so do no use it by default for database calls.

Update: improve return values used in c_base_utf8 class functions (and related)

I believe I was returning TRUE/FALSE types out of habit and for consistency.
For string processing functions, it makes far more sense to always return a c_base_return_string.
This simplifies the code the caller has to write.

The caller can still identify errors using the c_base_return::s_has_error() function.
The caller need only to call ->get_value_exact() if they do not care whether an error occured.
This will always return a string (and on error, it returns an empty string).

Progress: major database work, begin adding path handling, and other changes

More focus on database changes.
Begin adding path handling.
Custom utf8 function changes.
Other changes.

Cleanup: improve wording of the project description in the github readme file

Update: rework access control to be automatic based on is_* roles in the t_users table

Instead of trying to use select queries on every statement based on a users role, have the database automatical grant/revoke roles based on column settings in the s_tables.t_users table.
u_public and r_public have been renamed to be u_reservation_public and r_reservation_public to keep all roles self-contained.
name_machine fields should now only allow word characters.

Progress: improve access control database structure and other changes

Numerous changes, but there has been a huge focus on access control changes.
It occurred to me that if an account was locked or deleted, then they should be denied access to all of the views (with exception of the current user views).
This required a notable structural change.

The tables are now stored in the s_table schema.
In case accounts get locked, even administer accounts, access should be disabled.
I did not get the time to convert all of the administer and manager roles to be restricted to views but will eventually do so.

Progress: major database work and other changes

Progress: continuing development, http headers work

More progress with writing HTTP header processing code.

Other minor bugfixes and changes.

Progress: continuing development, http headers work

Fixed some syntax and logic issues.
Implemented more functionality (there still a lot more to finish).

Began added a little more documentation.

I've noticed some incomplete code in places and marked them with @todo.

Other minor fixes/changes.

Bugfix: operating on non-existent array index in p_s_split() of c_base_utf8

This causes PHP to crash as it does not like operating against non-existant indexes.

Progress: continuing development, http headers work

There is still a lot more to go, but this resolves some of the @todo's in the http headers class.

I have noticed that in some cases I called a prepare_token function and the return value has parsed in the name.
- The parsed name is used for parsing functions and generally returns an array.
- Based on the name, I could easily mistake the purpose of the return value for the prepare_token (or similar) functions.
- These prepare functions now use prepared in place of parsed to make reading the code more consistent.
- There are some functions that make this incorrect interpretation and they must be rewritten accordingly.

There are also other minor changes.

Progress: continuing development, updated test example, added language testing, miscellaneous fixes

Did some testing and made sure the test example was working.

Added some languages for testing.
- I specifically added japanese as a language to test for error messages because it is so different from english.
- It is not practical to implement the error messages class for many languages right now because I expect major refactoring as I continue coding.
- Another simple language test is added, testing the languages: russian, spanish, japanese, and english.

More work done on in developing the database.

Other fixes and changes.

Progress: continuing development, database work, use c_base_return on more classes, and fix inconsistencies

Further work on building the database structure.

Added some missing roles:
- Requester
- Editor

A number of inconsistencies have been identified and resolved.

After reviewing my use of c_base_cookie as extending the class c_base_return, I have decided that to do that everywhere possible.
This simplifies the return structure that I complicated when I decided to use the c_base_return return type development design.

I noticed that there is still some cookie functionality that I could implement.
This adds support for 'same site' and 'host only'.
These are functionally equivalent except that I think 'same site' is a little more granular.

This also adds one of my test scripts under examples.
It is very likely not completely up to date with the database changes.

Many other changes.

Progress: continuing development, more http header support, implemented error reporting

I noticed that there are a lot of incomplete functions in the http class.
This, in particular, implements cookie support.
There is still a lot more incomplete code in the http class.

My initial error reporting work is done.
Expect major refactoring as I continue developing.

After some testing, I have foind the php-xz module to be lacking and defunct.
I also noticed that there are multiple relevant repositories to choose from (added both to documentation).

Numerous other changes.

Progress: continuing development, more documentation and database work

Begin working on the database for the reservation program.

Write some documentation.

I am just starting to get work done on adding error messages (multi-lingual friendly at that).

Many other changes.

Bugfix: silence socket_connect() warnings/errors

Cleanup: reservation_session.php is not following programming style

The referenced code is not calling unset() as necessary according to this projects coding practices.

Cleanup: remove extra/unecessary postgresql comments for create_ldap_users

Bugfix: database user 'public users' should instead be 'public_user'

Update: add custom/experimental sessionize accounts and postgresql account auto-creation programs/scripts

This adds some experimental helpers for either creating postgresql accounts via ldap or providing session management independent of PHP.

Some of the C code is not using threads but likely should.
I will eventually come back and make a threaded version.

Progress: continuing development, adding examples and example program: reservation

For the sake of functionality and time, I implemented a straight-forward HTML class.
I will come back later to work on CML that will be an attempt to separate the context from the content, even in the code.

There is a (currently) experimental program called reservation.
This is a reservation/scheduling system that is currently under design and will be used as a functional example for using this site.
It currently requires LDAP and a few other custom/experimental software that is to follow.
The sql code is postgresql-specific.

Other examples are included in the example direction.

There are some theme files that I am probably not going to use and may remove in the future.

Update: initial import, including incomplete work

I wanted to complete all of the common/base work first, but I ended up getting side-tracked.
With the holidays, I realize that I would not finish this in time and I needed to at the very least ensure all of my hard work is not lost.

Much of the base work is complete except for the following:
- HTML and HTML tag related processing.
- Form processing.

I have not decided how I want to design forms and I expect this to be the weakest part of my code design.
How this will end up being developed will directly affect how the HTML processing code is built.

Initial commit