Hello,
With KMP it is possible to search all occurrences of a m length string, into a n length string, using O(m+n) machine operations. Next page: http://www.inf.fh-flensburg.de/lang/algorithmen/pattern/kmpen.htm describes the algorithm.
The KMP works well with the utf-8 encoding of unicode. One property of utf8 is that the encoding one unicode symbol is not a substring of another utf8 substring. This property allows to take the utf-8 encoding of the string you wish to search, and to find this utf8 encoding string, in the utf8 encoding of the text string. Geany uses scintilla, and scintilla uses utf8 to encode the document it displays, and scintilla has a command that gives the raw utf8 byte array for a [start, end) range. So, KMP gives great speed for searching all occurrences, and may be used with the underlying text representation of scintilla used by geany. The utf-8 encoding of a unicode string of length n, is less than 6n, each utf8 encoding is at most 6 bytes.
I also think that including this functionality/feature into Geany core would be a good choice. It would be a small tradeoff between keeping the core small, and adding this new functionality, but this is your choice.
If you wish to extend automark, then this is good choice too. If you wish, and if it helps, please reuse any part of the implementation provided here: http://sourceforge.net/p/geanyhighlightselectedword/code/HEAD/tree/trunk/Gea... If needed, I would help.
What should I do next? Should I not do the pull request for GeanyHighlightSelectedWord? It is okay with me. GeanyHighlightSelectedWord would then be still available at sourceforge until Geany provides this functionality from its core, or from automark.
Have a great day, Marius Ioan Buzea
-------------------------------------------- On Fri, 5/29/15, Thomas Martitz kugel@rockbox.org wrote:
Subject: Re: [Geany-Devel] pull request on GitHub, to add GeanyHighlightSelectedWords, into Geany Plugins To: devel@lists.geany.org Date: Friday, May 29, 2015, 10:03 AM
Am 29.05.2015 um 00:10 schrieb marius buzea:
Hello,
I had a look at
https://github.com/sblatnick/geany-plugins/blob/master/quick-search/src/quic....
This plugin does
something similar to GeanyHighlightSelectedWord.
The quick-search.c
calls Geany's search_find_text several times in one
processing, and each time a regex would be
recompiled in search_find_text.
This
is, I guess, a small cost when the regex is just a string.
GeanyHighlightSelectedWord implements search using KMP.
Designs differ, but
functionality is alike. I would keep both this plugins,
and not try to merge them.
Can you describe this KMP algorithm, and why it should be superior?
Anyway, the existing plugin (automark) should look into adopting it (if
it is indeed an improvment) instead of having multiple plugins with the same functionality.
PS: I also agree with providing it by the core.
Best regards _______________________________________________ Devel mailing list Devel@lists.geany.org https://lists.geany.org/cgi-bin/mailman/listinfo/devel