There was a small API change between libjansson1.3 and 2 that needed to be compensated for. These patches address that and should also remain backward compatable with jannson1.3 http://www.digip.org/jansson/doc/2.0/upgrading.html
This usage of the HTML entity was there for a reason. It was a quick eyeball test to see if utf-8 AND html decoding were still working. Since the output should always be the same, I'm reverting this back to the previous. This reverts commit 4b092af.
(note: spc doesn't care for ORG markup, so what's there might not even be valid ORG markup.)
You can now specify the input/output character set if you aren't using UTF-8 natively.
This is done to make sure that when this goes multithreaded, things don't break because setlocale() isn't thread safe.
Remind me to turn off kill-a-lawyer.el while writing new code.
Trailing whitespace is now dead. Very dead. Let's hope it does not return.
One more pass left to go.
This is just part of a brain dump, I need to find the original paper that linked together the sha hash concepts.
-D is now the short arg for a debug option -d adds support for the as yet uncompleted --description field which is intended to make "titles" be more independent (shorter) for presentation.
It was possible for an old, obsolete engine to crash gnugol. Null check added. Also cleaned up whitespace issues
Nobody but me cares about searching my own sites, removed, put elsewhere.
Command line options will take precedence/add to the options on GNUGOL_OPTS. Had to break the command line parsing into two stages, and parse GNUGOL_OPTS to create an argc/argv style "command line".
Given the functions changed: gnugol_init_QueryOptions() gnugol_reset_QueryOptions() gnugol_free_QueryOptions() when will you *ever* pass a NULL QueryOptions_t to it? You won't, unless it's a programming mistake, but in checking for NULL and doing stuff if it isn't hides bugs. One tenent of programming I do: THOU SHALT GIVE VALID PARAMETERS TO A FUNCTION (and in return, they'll return valid data) So, with that in mind, get rid of the unnecessary checks for NULL (remember: you won't be passing in NULL, now will you?) Also, free() accepts a NULL (it won't do anything) so there's no need to check for that condition. And besides, those pointers won't be NULL anyway, right? Right. If this gives you the heebie-jeebies, you haven't programmed enough with this style. It really does make for rock solid code. But you do need to learn to live with assert().