[sword-devel] Fastest way of counting verses
Manfred Bergmann
bergmannmd at web.de
Thu Mar 12 08:46:20 MST 2009
Hi Troy.
The speed of the library is good but we're producing some overhead for
display like adding markers to the text in order to synchronize
between views and such.
The overall speed is still ok but if the user pulls more books like
1-5Mose then it can take 7 - 10 secs.
However that's for your snippets, I could speed up some things here.
Here is something interesting:
(a not existing key is used in the examples).
---------------------------
sword::VerseKey vk;
sword::ListKey lk = vk.ParseVerseList("Par.1.1", vk, true);
lk.Persist(true);
swModule->setKey(lk);
for ((*swModule) = sword::TOP; !swModule->Error(); (*swModule)++) {
// do something
...
}
--------------------------
The above crashes at (*swModule) = sword::TOP.
---------------------------
sword::VerseKey vk;
sword::ListKey lk = vk.ParseVerseList("Par.1.1", vk, true);
//lk.Persist(true);
swModule->setKey(lk);
for ((*swModule) = sword::TOP; !swModule->Error(); (*swModule)++) {
// do something
...
}
--------------------------
This doesn't crash but iteration starts at Gen 1.
Note, Persist() is commented.
--------------------------
sword::VerseKey vk;
sword::ListKey lk = vk.ParseVerseList("Par.1.1", vk, true);
for (lk = sword::TOP; !lk.Error(); lk++) {
swModule->setKey(lk);
// do something
}
--------------------------
This produces the correct output - nothing, loop is not entered.
Regards,
Manfred
Am 12.03.2009 um 13:19 schrieb Troy A. Griffitts:
> Manfred,
>
> Are you specifically asking about how to obtain the ordinal count of
> a traversable SWKey?
>
> If so, the official answer is, "you can't without counting
> yourself". SWKey's which return true for isTraversible() don't
> contractually offer their set count. This is fairly common with
> optimized object containers who's set count is not easily derived
> (think of linked lists).
>
> Having said this, SWKey offers a bookmarking facility with long
> SWKey::Index(), SWKey::Index(long).
>
> The longs can _usually_ be used to practically bound a 'range'.
>
> e.g.
>
> VerseKey *parser = (VerseKey *)kjv->getKey();
> ListKey keySet = parser->ParseVerseList("matt-jn", parser, true);
>
> keySet = TOP;
> long start = keySet.Index();
> keySet = BOTTOM;
> long end = keySet.Index();
>
> progressBar.setStart(start);
> progressBar.setEnd(end);
>
>
> for (keySet = TOP; !keySet.Error(); keySet++) {
> progressBar.setValue(keySet.Index());
> // process text here...
> }
>
>
> // Having said this, a few points:
>
> obvious fail case: "gen; rev22:21"
>
> true ordinal count:
> long i = 0; for (keySet = TOP; !keySet.Error(); keySet++) i++;
>
> finally, processing a set of display data should really not take
> humanly noticeable times on even old desktop computers (or even
> reasonably recent mobile devices). We've actually, recently been
> trying to optimize things on mobile devices for SwordReader and as
> we continue to make improvements for these devices you should see
> improvements on all platforms. Please let us know with a snippet of
> code which demonstrates an iteration which takes any noticeable time
> to complete.
>
> -Troy.
>
>
>
>
>
> Manfred Bergmann wrote:
>> Hi.
>> Another APi question:
>> In order to show a progress indicator to not bore the user when the
>> cursor ball spins and a text pull takes a little longer I want to
>> know how many verses need to be pulled before actually doing it.
>> This can be for any reference the user enters, "gen" or "gen 1,6,9"
>> or anything like that.
>> What is the fastest way to do that?
>> Regards,
>> Manfred
>> _______________________________________________
>> sword-devel mailing list: sword-devel at crosswire.org
>> http://www.crosswire.org/mailman/listinfo/sword-devel
>> Instructions to unsubscribe/change your settings at above page
>
>
> _______________________________________________
> sword-devel mailing list: sword-devel at crosswire.org
> http://www.crosswire.org/mailman/listinfo/sword-devel
> Instructions to unsubscribe/change your settings at above page
More information about the sword-devel
mailing list