162177 - error in the downloading windows

Status: VERIFIED FIXED

(SeaMonkey :: Download & File Handling, defect)

Description

[regisr]

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.0.0) Gecko/20020802
BuildID:    Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.0.0) Gecko/20020802

During downloading the speed was incorrectly displayed:
(2.8 KB/sec 2.9 KB/sec 2.10KB/sec 3.1KB/sec ... )
the 2.10 sould be 3.0 ;-)


Reproducible: Didn't try
Steps to Reproduce:
1.download a big file
2.
3.

Expected Results:  Display 3.0 instead of 2.10

Comment 1

[Kai Lahmann (is there, where MNG is)]

could also be a temporary lag...

Comment 2

[alge]

hm. math. fun.

this is how the download manager does it: (nsDownloadProgressListener.js)
        // rate is bytes/sec
        var kRate = rate / 1024; // K bytes/sec;
        kRate = parseInt( kRate * 10 + .5 ); // xxx (3 digits)
[..]
         var fraction = kRate % 10;
         kRate = parseInt( ( kRate - fraction ) / 10 );
[..]
         rateMsg = replaceInsert( rateMsg, 1, kRate + "." + fraction );


and the old progressdialog: (nsProgressDialog.js)
    rateToKRate: function( rate ) {
         var kRate = rate / 1024; // KBytes/sec
         var fraction = parseInt( ( kRate - ( kRate = parseInt( kRate ) ) ) * 10
+ 0.5 );
         return kRate + "." + fraction;
    }

the old one seems rather complicated compared to the new (downloadmanager) one. 

blake, got a comment?
oh, and the indentation in nsDownloadProgressListener.js is off by one space.

Comment 3

[alge]

reporter: btw, did you see this in the single-file progress dialog, or the
download manager window?

Comment 4

[Kai Lahmann (is there, where MNG is)]

ok.. we have it:
the nsProgressDialog.js function returns 2.10, if it get's rate=3020 for
example. Problem: I'm not sure, if the new code is also 100% secure against
errors, where a x.99 shown as x.9 and not as x+1.0, but I can't get it to fail.

Comment 5

[Kai Lahmann (is there, where MNG is)]

ok, here comes my version of that:

var kRate = Math.round(rate/1024*10);
var fraction = kRate % 10;
kRate = parseInt(kRate / 10 );
return kRate + "." + fraction;

..what did I miss?

Comment 6

[alge]

Attachment: uses Kais New and Improved[tm] algorithm.

Comment 7

[regisr]

[e-mail]  asked me where I see this error: it was in the single-file
progress window.

Comment 8

[Henrik Lynggaard Hansen]

Adding review keyword so that someone hopefully takes a look at the patch

Comment 9

[Henrik Lynggaard Hansen]

nominating for 1.4b: a small bug that allready has patch (patch needs reviewing)

Comment 10

[Asa Dotzler [:asa]]

This looks like a good one to get. Blake said that the patch looked fine but
we'd need to get reviews on it. We're not going to hold the 1.4b release for it
though.

Comment 11

[Henrik Lynggaard Hansen]

There should be amble time for a review during the final cycle.I think we should
really try and get those bugs with patches checked in.

Comment 12

[alge]

Attachment: same patch, now with diff -u

Comment 13

[alge]

Attachment: new cleaner patch, thanks to dmose

Comment 14

[OOO until Jun 29 - Dan Mosedale (:dmosedale, :dmose)]

Comment on attachment 121879 [details] [diff] [review]
new cleaner patch, thanks to dmose

sr=dmose

Comment 15

[Asa Dotzler [:asa]]

Use Math.floor, not parseInt, and we'll reconsider for 1.4final.

/be (typing as Asa)

Comment 16

[alge]

brendan: The patch doesn't use parseInt(), it uses toFixed() 
(parseInt was part of the orignial problem)

Comment 17

[Asa Dotzler [:asa]]

not gonna block 1.4 for this but if we can get a review (from jag?) then seek
approval to land the patch by setting the approval1.4? flag and drivers will
evaluate for landing.

Comment 18

[jag (Peter Annema)]

Comment on attachment 121879 [details] [diff] [review]
new cleaner patch, thanks to dmose

r=jag

Much much better. How about using this in other places that try to do this?

Comment 19

[Asa Dotzler [:asa]]

Comment on attachment 121879 [details] [diff] [review]
new cleaner patch, thanks to dmose

a=asa (on behalf of drivers) for checkin to 1.4

Comment 20

[Christian :Biesinger (don't email me, ping me on IRC)]

Checking in nsProgressDialog.js;
/cvsroot/mozilla/embedding/components/ui/progressDlg/nsProgressDialog.js,v  <--
 nsProgressDialog.js
new revision: 1.28; previous revision: 1.27
done

Comment 21

[alge]

fixed. yay.

Comment 22

[Chris Petersen]

Verified on 2003-06-05-08 win32 and macho trunk builds