There's been several problems and attempts to fix PHP syntax highlighting in Geany. Perhaps it would be beneficial to write a small parser instead of relying on regex pattern matching. I don't think such a parser would have to be 100% complete... just sufficient enough to *properly* identify keywords, function/class constructs, and maybe a few other tidbits like namespaces. I know the desire is to keep the core code small, but a hand-written parser should be fast and rather lightweight.
Taking it a step further would be start thinking in terms of extensions. I don't know if it would be possible to override the internals of highlighting and populating the function list currently, but if the code were refactored to provide an API we could have a screen in the configuration options to specify colors of various tokens (keywords, variables, etc). Then geany would call out to the extension and return back a list of positions, lengths, and token types. It can then go through and set characters appearing at position x for length l with the color indicated by token type t. The work would be done in the external module and geany would be able to apply the coloring relatively easily in one pass. Then it calls out to get a list of classes and functions and such which the extension identified while it was parsing to populate the side list.
I'm just brainstorming here and not sure how it would work with multi-language documents like html/javascript/php. But might be worth discussing since this does seem to be a problem that comes up every now and then. Thoughts?
-Tim
Timothy Boronczyk, ZCE tboronczyk@gmail.com
Dnia czwartek, 16 grudnia 2010 o 06:44:04 Timothy Boronczyk napisał(a):
There's been several problems and attempts to fix PHP syntax highlighting in Geany. Perhaps it would be beneficial to write a small parser instead of relying on regex pattern matching. I don't think such a parser would have to be 100% complete... just sufficient enough to *properly* identify keywords, function/class constructs, and maybe a few other tidbits like namespaces. I know the desire is to keep the core code small, but a hand-written parser should be fast and rather lightweight.
There is no need to write a PHP parser. Geany can use the tokeniser delivered with PHP. (However, it is worth remebering that PHP is not only PHP code. It is also phpdoc, and it is also the master (domain) language, mostly but not necessarily (X)HTML, and that it is possible to mix multiple domain languages in the same file, and that HTML may contain CSS and Javascript, and…)
HTH, Chris