On Tue, Apr 07, 2015 at 09:38:37AM -0700, garduino wrote: > Is SmallCMS1 published somewhere?
I'll share some SmallCMS1's inner workings here: - Content is in plain text file, written in Markdown. Metadata is in JSON at the start of the content file. - I use Seaside's HTML rendition engine. Entire site is rendered in Smalltalk code, like how a Seaside application does. Currently calls out to the Discount C library's executable through PipeableOSProcess to get HTML from the Markdown content. I seem to recall that Petit Markdown exists, but haven't tried it. - I found Seaside's HTML-in-code style not so convenient when trying to implement one particular fine looking open source Bootstrap skin. Maybe using a templating mechanism like Mustache will make it easier. - Because my site is small, SmallCMS1 reads the entire content directory tree into a tree-like (duh) structure. URLs are matched by walking the in-memory tree using simple string matching. I was going to say this prevents the dot dot trick, but I thought I'd verify. Putting a halt in my code, I now notice that Firefox processes ".." before it sends the request, so SmallCMS1 doesn't see the ".." at all, at least when using Firefox. - Comment system is outsourced to Disqus. - I VNC into the image and use inspectors, workspaces, and browsers as SmallCMS1's admin interface. I run my own VPS so I have Unix command line access to the content directory. SmallCMS1 is thus quite tied to the way I work. I have not published SmallCMS1. Maybe SmallCMS2, if it ever happens... :-) Pierce