Syntax Highlighter Upgraded

2011-12-11

That Syntax Highlighter library I added yesterday has been upgraded from the version found on Google Code. The latest version by the author has a few more syntaxes, a different approach to calling for highlighting, an autoloader (which I won't use as I'm doing all the work myself already), and a slight change to the syntax selection.

I've upgraded the library files, am skipping the autoloader, skipping the legacy proxy, and am curious about the change to the toolbar space as it appears functionality may be lost.

	
		
	
	
		
Hello again, world!

	

I've also stripped out the functionality of adding the Syntax Highlighter code to the RSS feed as it has no use there, code blocks will still be rendered in the default manner but will not be very pretty.

Syntax Highlighter Added

2011-12-10

A few years ago I played around with a javascript library that would perform syntax highlighting on identified code snippets. It looks like the project has progressed and is available now on Google Code, so I added it to the site.

	
		
	
	
		
Hello, world!

	

In blog posts I look for any identifiable code snippets, collect the languages being represented, and add the Syntax Highlighter libraries needed to work on them. Above the fold is a code snippet for HTML, and below is also a JavaScript example. In the snippet on a blog page only the HTML library will be called (unless other snippets on the page have code examples), and the full blog posts will have JavaScript as well.

RSS posts will have post-level Syntax Highlighter library references, as well.

Keeping separate lists of snippet/full-post code language references was necessary to make this work in different ways where blog post content would show up, and to prevent duplicate library references. I also had to go inside each "pre" tag and alter < characters due to a limitation with the Syntax Highlighter code.

Fixed RSS blog post linkage, moved .post files to new path

2011-12-08

Today I fixed a bug with the post links in the RSS feed, and moved my posts to subdirectories for better organization.

The RSS bug was that blog posts had a link back to the full post page, but the path was incorrect. It included part of the build path where the final .html content gets prepared before being deployed to the web service path.

I also decided that putting all of my blog posts into a single directory was a bad idea, and have adopted a YYYY/MM model for keeping the clutter at bay.

After some code restructuring I found some other bugs with sequence of path operations, blog post URL errors, and more RSS file content issues. Over the course of a few hours I restructured the code some more and set limitations on how deep blog post formatting would go, and expanded the count and scope of RSS files.

I now have rss.xml files populated at all tree levels where a blog post resides down a farther branch. This means each of my major directories can be have blogs in them, which may come in handy as a mechanism for creating and maintaining multiple separate blogs. The blog posts will only link with other posts that same major branch down from the root path, which provides local scoping to that major branch. The root level's rss.xml can capture all blog posts, so it could operate as a master site feed.

I will probably have to do more work within the RSS files if ever I do start to distinguish other blog paths and use them, the content in the header information of the RSS feed is not dynamic to the source of the content.

Now with RSS

2011-12-05

Today I added an RSS feed to the blog. It includes the full post content, not just the snippet.

One of the earlier design decisions I made was to have content in my blog post files be reusable as content to be displayed on the main blog page, various flavors of blog pages, blog posts (permalink, full post version), and items in an RSS feed. For the last of these I wanted the content to be plain, devoid of frills like Prev/Next linkage, and complete instead of just being a snippet.

RSS can be a powerful and simple way to share content, though it is by intention very stripped down. I appreciate that philosophy, and though it doesn't work with every need of content providers it is enough for the simple kind of content I intend to provide.

One of the problems I ran into while implementing the RSS file construction was formatting the dates of my posts into RFC822 standard, which looks like this: "Tue, 06 Dec 2011 05:30:04 GMT". I have been keeping dates in human-readable YYYY-MM-DD or similar formats, and to go from that format to RFC822 required some work, most of which I received help on from an old Perlmonks article.

Now that I have an RSS feed I'll have to be more careful when I restructure my blog post files into date-based subdirectories. I should have instituted a better method than dumping the files all in the /blog path where they will accrue endlessly. Another task for another day.

Blog posts linked to each other

2011-12-04 20:58:00

Earlier this morning I pushed out my latest updates and found a few problems. Today I went about fixing things, and added another feature to make blog posts have Previous and Next post links at the bottom of the page.

When I pushed my changes I noticed that my blog posts no longer showed up in the order I expected them to. This is because I was using the `stat` of the post files to order them, and it slipped by me that using version control means the mtime isn't a reliable mechanism for determining post order unless I build an elaborate scheme to record my original file mtimes and re-touch the files after checking them out from Fossil, and really scratch that!

So after making some hotfixes I spent some more of today scrapping the mtime mechanism, in favor of reading content from within the posts themselves. On single-post days I should only need YYYY-MM-DD, and on multi-post days I'll also need time information to help disambiguate and order correctly.

The method of adding Prev/Next links is something I'd like to clean up, it's a bit kludgy and was hacked in and means I now need to iterate over my list of blog post files and open them twice for edits in order to do the bi-directional linking. I might ease the process and open for edits once if I do something else, maybe store post file metadata into a SQLite DB. I'll revisit this later, I think next up is either working on the calendar content or attempting to fetch my Twitter feed and display that content in the blog as well. ]]> http://howahumanwon.org/projects/website/2011/12/BlogPostsByFile.html Sun, 04 Dec 2011 03:21:00 EST http://howahumanwon.org/projects/website/2011/12/BlogPostsByFile.html

Blog posts by file

2011-12-04 03:21:00

I have now implemented adding blog posts by creating separate .post files.

Included in this is an after-the-fold piece so that the blog front page can be trim and show snippets of posts, with links to the complete post. As demonstrated here:

There are a few other improvements to my site building script, as well as the site itself. My latest changes include:

• adding Creative Commons licensing
• moving some hard-coded configurations to a JSON config file
• beginng separation from my JustMeOverHere domain, focusing on switching to HowAHumanWon, or an even more generic option
• improving coverage of logging in the site building sub-processes
• adding logging to file
• adding replacement identifiers so that static files can contain placeholders for dynamic data, such as blog post snippets for the main blog page based on content in other .post files
• preparating for future enhancements, like Twitter feed integration

While it doesn't look like I've been very active I have been working on some of these improvements in a sandbox space while testing new functionality. There are still some migration issues going over from the sandbox to the hosting site, such as making sure all of the Perl packages I am using are available and correctly installed in both locations. My sandbox actually has issues with Data::Dumper for some reason, and I haven't found the correct flags for the compiler to generate reliable .so libraries when fetching/updating Perl packages. For this and other reasons I have pared down my reliance on external Perl packages as much as possible.

I have started using Asana as my task list organizer and am very pleased with how it works. I'm so happy with it that I've also adopted it for work at my company and have vastly improved my task management when the project count goes higher than a handful, which during the holidays it inevitably does.

I think once I get a bit further along in my site building I'll be able to post more often, and with less website navel gazing. It will be nice to write some opinion pieces and get into some of the underlying decisions I've made to guide me in the style and behavior of this site.

Later. For now I'm going to commit all of my updates to the project and go over my list to see what features I'll work on next.

Hello, world!

2011-10-22

Some years ago I purchased this domain in the hopes of using it to express myself on the Internet.

I have a feature list of things I want to have on this website that I am starting to implement. I'm going to build many of the components for constructing and maintaining this site on my own. I'm also going to document the process now that I'm at a state where it is practical to do so.

I began with building a very small framework in Perl to generate all of the website's content from source header, sidebar, menu, content, and footer files. I am learning to use the Fossil version control system, using it will help me to track my progress and the source files used for this site.

This site, and this blog is completely in its infancy. As I type this the blog consists of one static HTML page with absolutely no frills. Over time I will grow my small framework to allow me to author posts as separate files, and the main blog page will display fragments of these posts with links to the full content.

More posts will follow, as will more features to the site. I'm going to enjoy learning some new things along the way, and that's the whole point. Lastly, I'd like to point out that if you double-click anywhere you'll make Usagi Yojimbo show his angry face.