Jos van den Oever
jvdoever at gmail.com
Thu Jun 18 11:10:35 BST 2009
2009/6/18 Sebastian Trüg <trueg at kde.org>:
> On Wednesday 17 June 2009 16:16:21 Jos van den Oever wrote:
>> 2009/6/10 Sebastian Trüg <trueg at kde.org>:
>> > There are two solutions when using Strigi:
>> > 1. Hack Strigi to support a fastest flag -> a lot of work I think.
>> > Probably not worth it, seeing that there are still analyzers to be
>> > ported.
>> This can be done without hacking strigi but by subclassing
>> That class determines how much information should be retrieved. You
>> can there set the maximal number of bytes allowed to be read.
>> If you 'hack' the stream you feed strigi to have a maximal lenght, you
>> can receive wrong results, because strigi then thinks the file ends
> aha, nice, did not know that. And what about the analysers? Do they handle
> that or are they forced to only use the configured number of bytes by the
The analyzers *should* follow the configuration. If they do not, they are buggy.
The only analyzers that can be buggy are the endanalyzers. The other
types of analyzer do not have influence on the amount of data read.
More information about the kde-core-devel