Re: [Wget-dev] wget2 | Progress reports >100% file size during recursive downloads with HTTP/2 (#339)

[Josef Möllers]

Sorry for the somewhat long delay ... I had slightly more urgent tasks to 
attend to.

My analysis so far is that the bar's reference is set to one of the files' size 
(the first?), but the update (progress) value used is from whatever is 
currently updated. If these happen to be identical files or the file sizes 
happen to be identical, then the total is 100%. If, however, the reference is 
set from a very small file (eg to 8 from some .Po file), and later the progress 
of a file of different size is displayed, then the percentage is off.
Does this make sense?

My proposal is to amend "struct _body_callback_context" with an additional 
field "uint64_t file_size" and set this in _get_header().
Then, in _get_body(), pass this value to bar_set_downloaded() and further down 
the chain such that when _bar_update_slot() calculates the percentage, the 
reference is correct.
I'm currently implementing this.

As an alternative, the size of a file to download should be added to 
slotp->file_size and the number of bytes received in a DATA packet should be 
added to slotp->bytes_downloaded, such that, in the end, the sum of all file 
sizes is the reference and the sum of all DATA packets is the measurement 
value. This would mean that the percentage would drop whenever a new file's 
HEAD is received and is added to slotp->file_size. In the end, the two sums 
should be equal and the percentage would be 100.

-- 
Reply to this email directly or view it on GitLab: 
https://gitlab.com/gnuwget/wget2/issues/339#note_94868348
You're receiving this email because of your account on gitlab.com.