Mailing-List: contact dev-help@cordova.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cordova.apache.org
Received-SPF: pass (nike.apache.org: domain of mike@ustadmobile.com designates
 88.198.162.115 as permitted sender)
To: <dev@cordova.apache.org>
Subject: CB-8471: Impossible to resume FileTransfer
MIME-Version: 1.0
Content-Type: text/plain;
 charset=UTF-8;
 format=flowed
Content-Transfer-Encoding: 7bit
Date: Fri, 13 Feb 2015 01:20:19 +0100
From: Mike Dawson <mike@ustadmobile.com>
Message-ID: <ee68fb58358f6eda24ca21155df7744e@ustadmobile.com>
User-Agent: Roundcube Webmail/0.7.1

Hi,

This has been discussed briefly before:

http://mail-archives.apache.org/mod_mbox/cordova-dev/201311.mbox/%3CCAK_TSX+dKpLr7vCCacN9kMv1d0=WN2z1DfapheyEam3R_DVG1Q@mail.gmail.com%3E

Problem is when using the FileTransfer plugin to download a file, if 
the file is interrupted it will get automatically deleted.  This is what 
happens on Android (FileTransfer.java line 948):

// Remove incomplete download.
if (!cached && result.getStatus() != PluginResult.Status.OK.ordinal() 
&& file != null) {
     file.delete();
}

This will make it impossible to resume from where the file left off, 
and on many networks/situations for a lot of people in countries / areas 
where the networks are not so good the probability of ever succeeding at 
downloading a sizeable (e.g. 30MB+) file will be very low and it would 
waste a lot of bandwidth.

It is true that resume is a complex business and perhaps out of the 
scope of such a plugin (what happens when file changes, how many 
retries, etc).  If we had just an option like keepIncompleteDownload 
which would default to false this (ie. keep the current behaviour) would 
allow developers to implement their own logic in Javascript to handle 
resume if they wish to - e.g. try the download again, sending a range 
header, to another file and at the end concatenate all the file pieces 
together.

I have to make a working version of this behaviour for an app we 
currently develop just for Android; so I can write that behaviour there. 
I don't see any tests that describe the desired behaviour in this 
scenario.

Would be interested to see if anyone else has any alternative 
suggestions.

Regards,

-Mike

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@cordova.apache.org
For additional commands, e-mail: dev-help@cordova.apache.org