Feature request (resume + chunks)

[vlad]

Hi,

I would really like to see the "resume" parameter working with the chunked uploads.
The server could return the id/number of the last successfully uploaded chunk and the applet would then continue the upload from that chunk.

Regards,

Vlad

[support]

The current workaround is to use:
<PARAM NAME="resume" VALUE="true">
<PARAM NAME="overwrite" VALUE="false">
<PARAM NAME="policy" VALUE="ignore">

Each chunk already uploaded will be skipped. The current chunk will be resumed and next chunks will be uploaded.

Could it help ?

[vlad]

Customizing the server side script to skip the chunks that were already uploaded is easy. However, when struggling to upload a 2GB file over a bad connection, starting the transfer from the beginning (uploading chunks that are already on the server) will waste a few more hours so is not the best option.

[support]

So, you suggest that the server-side script returns the id of the last chunk (for the HEAD request sent for resume) and then the applet will restart from this last id?

[vlad]

So, you suggest that the server-side script returns the id of the last chunk (for the HEAD request sent for resume) and then the applet will restart from this last id?

Yes. That would make the applet a more reliable upload tool and it doesn't seem very difficult to implement.

[support]

Ok, we're making some improvements and I think we could provide a testing release with this feature next week. Drop an email to support(at)jfileupload(dot)com.

[vlad]

That would be great.

P.S. Is there a way of sending also the expected total file size of the file with each chunk? Or at least with the last chunk, so that the server side script can make sure the file doesn't miss any bytes.

[support]

Will see if it can be added too.

[support]

Added in 2.5. They is a new multipart parameter in each upload request. It looks like:
chunkbasesize=124587812

[support]

We're working on a solution to resume chunks. It will be based on "content-range". Each chunk willl be appended to the final file on server. There won't be any intermediate chunk stored.