info-beamer pi 0.8.1 released

Here is another release after 3 months of development. This release will give you yet another big performance boost. You visualizations could by up to 50% faster now.

Better performance

This release changes how shaders are set up. This change will speed up rendering. You do not have to change anything in your code to benefit from that. You can read more about performance differences in the recent blog post.

If you try this new release on your existing code, I'd be happy to get your feedback.

Luajit 2.1

This version of info-beamer uses the new 2.1 branch of luajit. This change could potentially speed up your visualizations.

If you are interested in experimenting you can give the low overhead profiler provided by luajit 2.1 a try. If you set the environment variable INFOBEAMER_PROFILE to either annotate or top the profiler will run and take samples at regular intervals. When you exit info-beamer you'll get a summary of how your code behaved. These results might allow you to fix performance issues in your code. Please get in contact if you make use of this feature.

Kerning

The font rendering code now uses the kerning information provided by many true type fonts. Kerning basically changes the horizontal space between each pair of characters depending on that pair of characters. Learn more about kerning on wikipedia. You can disable kerning if for any reason this negatively affects the output of your visualization. To do that, set the environment variable INFOBEAMER_NO_KERNING to 1.

Better console output

info-beamer now has different log levels. This allows you to tune down output generated by info-beamer to a minimum if you are not interested in every detail. The default log level now produces less output than previously. Use the environment variable INFOBEAMER_LOG_LEVEL to change log levels. The default log level is 1. Set it to 3 to restore previous verbosity. You can change the log level up to 9.

UTF-8 support library

Lua doesn't natively support unicode. All strings and string functions provided by Lua treat all string as pure byte strings. While this works for latin1 or other 8bit encodings it creates problems when you try to use unicode in the form of UTF-8.

When you use info-beamer you often have external string values (for example in json files) that are encoded in UTF-8. The font rendering code of info-beamer expects UTF-8 encoded values. So usually everything works as expected. But if you want to get the length of a string in Lua, a simple string.len(value) will return unexpected results depending on the string provided.

For that reason info-beamer now bundles the luautf8 library. Usage is simple:

local utf8 = require "utf8"
print(utf8.len("ä")) -- Will correctly output 1
print(string.len("ä")) -- Will wrongly output 2