On Wed, Feb 19, 2025 at 03:30:19PM +0100, Toon Claes wrote:
> We're about to add the use of progress through curl. Although, curl
> doesn't know the total at the start of the download, but might receive
> this information in the Content-Length header when the download starts.
>
> To allow users set the total size after calling start_progress(), add a
> function progress_set_total().
Makes sense.
> +void progress_set_total(struct progress *progress, uint64_t total)
> +{
> + if (progress)
> + progress->total = total;
> +}
I wondered if we'd need to do any other computation here (that would
have been done in start_progress() if we specified the total then). But
it looks like we don't look at progress->total until we're ready to
display something.
Would we want to call display_progress() or similar here, to update the
original view like:
Working hard: 1<CR>
to:
Working hard: 33% (1/3)<CR>
immediately, rather than waiting for more progress to be made? I guess
it probably doesn't matter that much in practice as we'd remain stale
for a brief period in most cases (particularly for transfers where we're
updating based on bytes received, which is a pretty small unit of work).
Plus I think it may be awkward, because we don't know whether to call
display_progress() or display_throughput().
-Peff
