This is an early test. It may have huge bugs.
This search engine uses sh scripts for crawling and searching, wget, grep & sed, elinks for parsing html and plaintext as database
I am aware that large searches are often unsucessful: grep takes too long to search through the files
when they are not in cache. When they are, it's still slow because of the cpu. Only solution: a real database
No combination of search terms possible at the moment, multiple search terms only searches for an exact
match of the phrase
Results can be badly truncated. I try to "de-html-ify" the pages with broad paragraphs but it doesn't always work
Since I don't match b32 and normal i2p urls there will often be duplicate results.
Layout with multiple results sucks for now
Presenting of search results now done by little C program - much faster