|
From: <rg...@sd...> - 2003-07-06 04:21:21
|
>>>>> "Ben" == Ben van Klinken <be...@vi...> writes: Ben> Ben> Hi Rob, Ben> Sounds like a good idea, Rob. Linux Gazette sounds good! No Ben> specific files have been used before - just whatever i come Ben> across first. Documentation of something generally. Ben> Ben> You'll currently find some differences between the outputs of the Ben> java and c++ version. It's not a major problem but should be Ben> fixed anyway. The numbering of documents (or segments i think) Ben> some how differs. I looked into this but could not find why at Ben> the time. This doesn't affect the search results though. Ben> Ben> Also, don't forget when running searches that the Ben> unicode/non-unicode versions will have slightly different output Ben> (depending on the input of course). The java version uses unicode Ben> - and in c++ is optional. There is a gcc script version of Ben> clucene in the pipeline so it would be great if someone could Ben> test unicode on that :) Ben> Ben> Thanks Rob for the offer. Look forward to hearing the results. Ben> Ben> cheers, Ben> Ben Ok, I will start with the Linux Gazette. Eventually I would like the test body to be large enough that we can time indexing and searching to see if changes affected runtime. But for now, just a few sanity checks that can be verified with grep would be nice. I'll let you know shortly. --Rob |