curl throws away partial downloads on retry, even if told to resume #1084
Description
Say I want to download a large file, 1 GiB perhaps. I specify -O.
My network connection isn't very reliable, so I specify -m and --retry, so it doesn't hang up.
The problem is, on every retry, the output file is truncated, effectively throwing away all the work so far.
Warning: Transient problem: timeout Will retry in 1 seconds. 5 retries left.
Throwing away 327962289 bytes
This is wasteful, and makes --retry quite useless.
At least, if -C - is specified, it's pretty clear the user wants to always resume automatically.
The workaround is to either not to use --retry & shell-wrap curl in a for loop, or to use wget.
Relevant source, for anyone interested.
curl 7.50.3