New Code!
There is a small stack of defects and output quirks in discount that I’ve been churning my way through, and I’ve decided to release the code in two steps so that if I’ve horribly broken anything in the simple changes they will be reported before I mangle the code for the larger changes (better handling of *, **, _, and __, as well as a CONFIGURE.BAT file for Windows) that remain to be done.
So, discount has been pushed up to version 1.2.0 with the removal of one memory leak defect (code blocks leaked the blank lines at the end of the block) and two subtle output changes. The defect was nothing exciting – I was merely setting t->next = 0 without bothering to delete the unwanted lines that were attached to t, but the output changes have the potential of biting the unwary:
- The contents of a
<html tag>are now parsed as code, not as text. This keeps<ahref="linky">from being written as<ahref=“linky”>, and may thus confuse source that was written with\"’s to get around the smartypants extensions. - When printing
<email@address>automatic links, ensure that the human-readable part of the link doesn’t include amailto:, just the mail address.
This code has been run through the standard test suite and failed to fail, plus the memory leak does not cause exciting core dumps, so it’s tanned, rested, and New Code! for your coding pleasure.