Dan:
What a surprise! The top 1/3 is my code. The remaining is completely unknown
to me.
I do not know if it works or how it works. My hats off to the unknown author.
Unfortunately,
I have not the time to read and test the code right now. I am knee deep in
Judy1/L.
However, I am delighted to see someone take the brave step into Judy code.
Do you know how to download the whole package? I would like to see if any other
modifications have been made to Judy code and add the additions to my test
programs.
I told someone about a year ago I would have a new version of JudyHS or
equivalent.
I haven't gotten very far. So I have given up on predictions of release dates.
Api? I had no say in this one. I haven't had the time to even look at it.
Api changes
in the future are up in the air until I get close to release, all inputs are
welcome.
The judy-1.0.3 release, I think, is the lastest code changes. 1.0.4 and 1.0.5
were
just build changes to handle more platforms.
Doug Baskins <[email protected]>
________________________________
From: daniel zvinca <[email protected]>
To: Doug Baskins <[email protected]>
Sent: Mon, March 21, 2011 12:45:51 PM
Subject: Re: judy arrays hs parsers
Hello Doug,
Check this
out:
http://svn.openfoundry.org/pugs/third-party/judy/Judy-1.0.3/src/JudyHS/JudyHS.c
Is that OK? (for sure is not your code, but how reliable is it...)
I am very glad to hear about the new faster version. Do you have any idea about
the new release date?
Are they any benchmarks available? I am sure a lot of people would be very
happy
to hear about it.
Are they expected any changes in the parameters for api functions?
(too many questions, i'm afraid)
Best regards,
Dan
On Fri, Mar 18, 2011 at 8:38 PM, Doug Baskins <[email protected]> wrote:
Dan:
>
>
>>Hello Doug,
>>
>> Your judy array structures are still one of the fastest dictionary like api
>>ever
>> written. I plan to use your library as base for some fast hierarchical
>>structures
>> in a commercial application. One of the issues regarding the JudyHSapi is
>> the missing parsing possibilities (first, next, prev, last) which were
>>available in
>> 1.03 version.
>
>I will probably do something about this problem someday. There was no version
>of JudyHS
>that had a way of doing a search. It does not make much sense on random order
>of Indexes.
>I am on making Judy1/L faster now.
>
>
>>
>> For me would be a huge help to parse the whole array in a
>> "sorted" order passing a pointer to a simple function like func_compare(void
>> *item1, uintlen1, void *item2, uintlen2) on retrieval stage - preferable for
>> flexibility - or on the storage stage. (see unicode strings, arbitrary
>>structures,
>> etc). Do you have any suggestion or trick to reach this?
>
>I havent got a clue.
>
>
>>
>> I have to say that one of the reasons of using your library is the memory
>> friendly behavior, so having double lists, is not much of an option :(
>>
>> Second question is related to a "fast" mode of filling a judy array from a
>disk
>> like storage.
>
>If the data on the disk is in sorted order, then the normal JudyHSIns() will be
>probably be faster than your disk read.
>
>Sorry, I haven't been much help.
>
>Thanks for your interest,
>
>Doug
>
>>
>> Kind regards,
>> Dan
>
>Doug Baskins <[email protected]>
>
>
>
------------------------------------------------------------------------------
Enable your software for Intel(R) Active Management Technology to meet the
growing manageability and security demands of your customers. Businesses
are taking advantage of Intel(R) vPro (TM) technology - will your software
be a part of the solution? Download the Intel(R) Manageability Checker
today! http://p.sf.net/sfu/intel-dev2devmar
_______________________________________________
Judy-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/judy-devel
