[Koha-bugs] [Bug 11703] Convert checkouts table to ajax datatable

bugzilla-daemon at bugs.koha-community.org bugzilla-daemon at bugs.koha-community.org
Thu Feb 13 22:09:50 CET 2014


http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=11703

--- Comment #11 from Kyle M Hall <kyle at bywatersolutions.com> ---
(In reply to Owen Leonard from comment #9)
> Some issues I see:
> 
> - checkouts.js contains English strings. They're escaped for translation,
> but the translation script doesn't touch js file at all. Translatable
> strings must be in a .tt or .inc file.

Noted

> - Does this eliminate use of the UseTablesortForCirc system preference
> altogether? If so it should be removed.

Yep, I had forgot to take care of that.

> - I disagree that submit buttons for circulation actions need to be restyled
> with Bootstrap styles and icons. I'm in favor of pretty conservative use of
> button icons. I don't think they lend added meaning in this case, just
> visual clutter.

Fair enough. I feel like they have better "weight" considering every other
button on the page has an icon, but that's just my personal opinion. I can
remove the icon.

> - This patch eliminates the separation of today's checkouts from previous
> days' checkouts. I don't think that's a good thing. I think maintaining that
> separation is important to the way the circulation staff does their job.

I was waiting to see if this would come up. I'm not surprised. I should be able
to add that back in.

> - It looks like tables don't load with any particular sort by default? I
> would like to see it stick with the previous behavior of default sorting.

The default sort is issuedate descending, or at least it should be!

> - I think it's problematic that the "processing" overlay obscures the page,
> especially on the patron detail page where there is more to obscure. Can we
> get some kind of loading indicator within the table itself instead? That
> would give a clearer indication to the user what they're waiting for.

I completely agree. We can easily turn it off, but it would be much better to
have it float over top the table itself.

> Overall this looks like it works well, but I'm concerned that the
> performance hit we get for full page render, including tables. I tested with
> a patron who had only 10 checkouts and the AJAX version was a few extra
> seconds to fully load the table of checkouts.
> 
> Testing a patron with 250 checkouts I found the time to page render with
> non-AJAX was about 20 seconds, time with AJAX was 45 seconds.
> 
> On the other hand full page render is only part of what the staff is looking
> for with this page. We also want to see reduced time between one item
> barcode scan and the next. In my tests with the 250-checkout patron I found
> the AJAX response time to be mixed, with some faster than standard and some
> not. I think this requires more testing.

I'm a bit surprised by your results. Mine has showed a steady improvement. A
patron with 200+ checkouts went from 40 seconds to 1.5 seconds for rendering.
The fact that you had results switching sides is interesting. There must be
something else going on. However, the primary purpose of this patch is to
improve the responsiveness of the circulation system, rather than the overall
page load time. The total time to load including the checkouts table will
likely be slightly higher, but the I think the increased speed at which
librarians will be able to check out books will be well worth it. In addition,
we can later tune the ajax script to be as fast as possible. I've found that
about half the time is spent on GetIssuingCharges, CanBookBeRenewed, and
GetRenewCount.

> One worrisome thing: A couple of times in my test submitting a barcode in
> the AJAX version resulted in me getting kicked back to a blank checkout
> screen as if I had submitted a blank barcode. Other testers please watch for
> that.

That is odd, I haven't encountered that particular issue.

> I'm marking Failed QA so that those bullet points can be considered, but I
> hope others will continue to test this.

Agreed! Keep on testing!

-- 
You are receiving this mail because:
You are watching all bug changes.


More information about the Koha-bugs mailing list