yoush.homelinux.org

There are lots of wikies in the web running on MediaWiki software. Most of those are far not as large as WikiPedia, but contain only several 10s or 100s of pages.

Today I faced a situation when it would be good to download entire content of a MediaWiki installation, for futher offline browsing.

Doing so using variants of wget --mirror turned to be next to impossible, because wget ended downloading all edit- and special-links, and was unable to complete within an hour, then I interrupted it. Using --reject flag did not help because it does not stop wget from following and downloading link, it just removes the file after download.

Here is a solution that I found:

wget -k -p -r -l 1 --html-extension --reject '*=*' http://linux-raid.osdl.org/index.php/Special:Allpages

This command completed in several seconds, and produced exactly what I wanted – cross-linked copies of all wiki pages, and almost no garbage.

If there is a better method, please let me know

Why in the world your webmasters think that you should decide if your reader will read wide or narrow text pages?

Ever heard that web browser windows are resizable?

If I have my window resized narrow – and I may have that for a reason! – why in the world I’m seeng horizontal scrollbar and text is behind right window border, while half of window contains nothing than background?

If I have my window resized wide – and I may have that for a reason! – why in the world I have to look at huge empty pads at both sides while text (e.g. code examples) is not fitting in the central field, for the best having ugly horizontal scrollbar attached, but more often just garbling other text on right?

P.S.
Still searching a nice-looking non-fixed-width wordpress theme with internationalization support …