Always check that downloads are complete

[triffid]

curl can leave partial downloads after network errors and the script will not retry (despite -C -) due to if ! -e $1 on the previous line

This pullreq makes curl always retry, since we don't have upstream filesize information available to compare against stat -c %s $1

Since curl returns error code 33 when the file is complete (claiming that server doesn't seem to support byte ranges), we trap this error, consider it a success, and continue

[FreddieChopin]

Hello again!

Thanks for your second contribution, which fixes the problem with partial downloads. I've pushed your change with very minor modifications:

• I've simplified the logic expressions a bit,
• there's no point in printing the error code, as curl does that automatically (at least on my system - let me know whether it behaves differently on your system),
• I've added some limits to download (max 30 seconds for connection, max 30 seconds with speed below 1 kB/s), which allow for simple test of the new feature - just pull the Ethernet cable (or shut down your router) in the middle of a download (;

I hope that my additions/changes are OK - if I messed something up, just let me know!

[triffid]

Looks lovely 👍 thanks!