y.st./source/pages/en/weblog/2016/11-November/22.xhtml

<?php
/**
 * <https://y.st./>
 * Copyright © 2016 Alex Yst <mailto:copyright@y.st>
 * 
 * This program is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 * 
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
 * GNU General Public License for more details.
 * 
 * You should have received a copy of the GNU General Public License
 * along with this program. If not, see <https://www.gnu.org./licenses/>.
**/

$xhtml = array(
	'title' => 'Student debt',
	'body' => <<<END
<img src="/img/CC_BY-SA_4.0/y.st./weblog/2016/11/22.jpg" alt="A smashed-up minivan" class="weblog-header-image" width="811" height="480" />
<h2 id="general">General news</h2>
<p>
	I awoke to find an awful letter.
	My student loan payments are coming due now.
	The timing couldn&apos;t be worse.
	I&apos;m trying to find a place to live, and now, about \$81 $a[USD] is going to be siphoned away each month.
	This sucks.
	It&apos;s already going to be a struggle to pay the rent.
	I logged into the loan repayment website to make a payment, and it seems that the website no longer likes my email address and wouldn&apos;t allow me to send a payment until I changed it.
	That&apos;s idiotic.
</p>
<p>
	As I was headed to out to go to work, my mother got toed home.
	They had hit a deer, mangling the front of their car! Later in the day, someone sent by the insurance agency came by to haul the vehicle away for examination.
	Making sure that my mother was okay and discussing the event with them took some time, and I thought that I was going to be late to work.
	However, I still had fifteen minutes to spare! The shift leaders were nearly late though, only showing up a few minutes before everyone needed to be ready to clock in.
</p>
<p>
	I think that I better understand the decision in $a[PHP] not to have <code>\\var_export()</code> simply serialize any objects that it comes across and wrap them in calls to <code>class::__set_state()</code>.
	If you wanted the object serialized, you&apos;d simply use the <code>\\serialize()</code> function instead! As a bonus, that function can even serialize multidimensional arrays nested objects, and combinations of the two.
	However, that doesn&apos;t mean that I agree with the decision to implement <code>\\var_export()</code> this way.
	Simply put, <code>\\var_export()</code> attempts to output a string representing the declaration of a variable&apos;s value.
	However, objects aren&apos;t like other variable values and, at least as the syntax of $a[PHP] currently exists, objects cannot simply be &quot;declared&quot;.
	The concept of declaring an object doesn&apos;t make any sense, given the way that objects and classes work.
	Any attempts to &quot;declare&quot; an object is going to come out messy, as it does with <code>class::__set_state()</code>.
	The most important issue with <code>\\var_export()</code> is probably that while it serializes objects in a way, it doesn&apos;t have a <strong>*pre-serialization*</strong> method acting as a counterpart to <code>class::__set_state()</code>; a <code>class::__get_state()</code> method, as it were.
	This method should be part of a pair, like <code>class::__sleep()</code> and <code>class::__wakeup()</code> are.
	In other words, the way that <code>\\var_export()</code> works is a bit incomplete.
	Adding to this incompleteness is the way in which <code>\\var_export()</code> names its array keys.
	Inherited properties overwrite child properties, potentially causing information loss.
	If <code>\\var_export()</code> used the same array key structure that <code>\\serialize()</code> uses, property names from multiple classes would have unique names when need be.
	Also though, <code>\\var_export()</code> assumes that every class has a <code>class::__get_state()</code> method.
	This isn&apos;t always the case, and shouldn&apos;t be assumed.
</p>
<p>
	My <a href="/a/canary.txt">canary</a> still sings the tune of freedom and transparency.
</p>
<h2 id="include.d"><a href="https://git.volatile.ch./y.st./include.d/releases">include.d</a></h2>
<p>
	I took a look at my <code>\\st\y\uri</code> class, and I think that most of it, if not all of it, will need to be entirely written from scratch.
	What a pain.
	This will be the second major overhaul of that class.
	If it weren&apos;t for the fact that the built-in <code>\\Closure</code> class bypasses the protections offered by the <code>protected</code> and especially the <code>private</code> key word, this rewrite wouldn&apos;t be necessary.
	However, as it is, I don&apos;t think that I can get out of it.
	Furthermore, this rewrite will make a big difference for the <code>\\var_export()</code> function, causing it to produce a better array from objects of the $a[URI] classes.
	In all honesty though, I&apos;d ignore the <code>\\var_export()</code> optimization if I wasn&apos;t dealing with the <code>\\Closure</code> class.
</p>
<p>
	I don&apos;t have the time, the energy, or the patience to deal with rewriting my most complex class.
	You know what? I give up.
	<code>\\Closure::bind()</code> and <code>\\Closure::bindTo()</code> break $a[API]s.
	There is nothing that I can code that will fix that, as anything that I build can be broken again with these two methods.
	As for <code>\\var_export()</code>, attempting to optimize my class for that one use case pretty much <strong>*de*</strong>-optimizes it for every other use case that I can come up with.
	If rewrite my class, I&apos;ll get cleaner output from that one function, but that function is entirely messy.
	I should instead keep my code working as best I can with mostly the same $a[API] that I&apos;ve already built; there&apos;s not much to fix about it.
	I am going to need to make some key design decisions in for the $a[URI] classes before I move on with include.d development in the $a[URI] branch though.
</p>
<p>
	While trying to get everything working in the branch relating to the <code>__set_state()</code> method, I tried to create an abstract static method.
	Apparently, that&apos;s not allowed, which is pretty stupid.
	Stupider still, there actually <strong>*is*</strong> a way to get the desired result, it just involves splitting the abstract class into both an abstract class and an interface.
	The interface can declare as many methods as you want it to, be they instance methods or static methods.
	If a class implements the interface, any methods defined by the interface are considered to be abstract.
	Voilà! The abstract class can define any non-abstract methods that it needs to, relying on the interface to define the abstract static methods.
	Either the interface or the abstract class can define any abstract instance methods that are needed.
	Why isn&apos;t there a structure that can define full methods <strong>*and*</strong> abstract static methods? This seems like a big oversight.
	According to someone on Stack Overflow not only explained the reason for the problem, but also says that it&apos;s been <a href="https://stackoverflow.com./questions/999066/why-does-php-5-2-disallow-abstract-static-class-methods#answer-31235907">fixed in $a[PHP]7</a>.
	All that I have to do is wait, then this will work itself out.
</p>
<h2 id="university">University life</h2>
<p>
	I think that I&apos;ve about finished my <a href="/en/coursework/BUS1101/Behavior_management_at_the_SAS_Institute.xhtml">essay</a> for the week, but I should check it over again tomorrow if there&apos;s time before I head into work.
	By the time that I get home from work, the essay will have already been due.
</p>
<h2 id="apartment">Apartment hunting</h2>
<p>
	I spoke with the property manager today, and it seems that the local property has washer/dryer hookups in each unit, but I&apos;d need to find a way to get a washer and dryer home to install.
	This seems like an almost insurmountable feat.
	They also seem to think that the properties in Eugine have similar setups, though they don&apos;t seem to know for sure.
	They gave me a list of the property names, but they didn&apos;t have addresses for them.
	I&apos;m not sure if I&apos;ll be able to look them up online by name or not.
</p>
END
);