When downloading a file with the size of 84 megabytes is downloaded, then more than its size, then less. This shows an incorrect percentage of downloads.
http://funkyimg.com/i/2xXQW.jpg
In version 183 this bug is not.
http://funkyimg.com/i/2xXR1.jpg
Hard for me to test because my connection is sufficiently fast to download a 50 MB file in less than ten seconds, but I see no issues on (Pre) RC1 with this very large file:
https://dl.google.com/android/repository/android-ndk-r15c-linux-x86_64.zip
@Mike, perhaps server sends invalid value for size of the file, hard to tell without some public test case.
Perhaps it's some change in QtNetwork, weekly 183 most likely uses older version.
I downloaded this file to another file sharing service and from there it is downloaded correctly. But initially, before downloading to Yandex, I downloaded the file from another site, and there were also problems with statistics and size. From this I conclude that the error appears if the processor is heavily loaded during the download (interface brakes). Checked version 187, there is also this bug. http://funkyimg.com/i/2xZss.jpg But in version 183 the file is downloaded correctly, even if you load the processor, for this I used the opera presto with the open site ulmart.ru. Here is the file:
https://yadi.sk/d/C5wH1-Gr3NTGAA
@Mike, thanks for test link, unfortunately I was unable to reproduce it under both Windows and Linux.
Maybe this bug will appear when you download a file larger than this.
@Mike, dunno, I need some reproducible test case to find out what exactly went wrong.
File 500 MB.
https://yadi.sk/d/1wRBEG7h3NoSxv
@Mike, still no luck, but it might be somehow XP specific too, I'll try to test it there too (testing that build under Windows 10 doesn't show any issues with values).
No difference when testing it under XP, both files, they end with 100% and bytes number matches.
It is written here (http://forum.ru-board.com/topic.cgi?forum=5&topic=46573&start=500#8) that if the download is stopped and then resumed, the change in interest is not shown until the end of the download. This bug and incorrect upload of files can be connected.
@Mike, but you didn't have to resume download to reproduce this issue, right?
Yes
@Mike, no idea about that then, cannot reproduce, with resuming it might be possible.
When downloading large files from a folder on my computer, a bug is also there.
@Mike, is there anything unusual about your instance of Windows XP? Which SP do you have?
Windows XP Home SP2. I did not use any programs to change the parameters of the system.
@Mike, I see.
Perhaps we could try using some special build that collects debug information while downloading files.
Where can I download such an special build?
@Mike Nowhere yet.
@Emdek will have to create it first. ;)
Btw, is there a particular reason to use SP2 instead of SP3?
SP2 runs a little faster and does not send data to the network about the processor.
It's faster for you? Interesting, thanks. In my experience pre-SP1 XP was surprisingly bad (compared to NT5.0/2000), SP2 was fantastic and SP3 was slightly better still with more efficient memory use, better security and perhaps even being slightly faster. But that last one is probably very much ymmv. I was on a Core 2 Duo on all my regular-use systems by 2008. Are you using an old AMD CPU or some such? :)
I had similar problem downloading drivers from nvidia, like size mismatch, it would just report an error at the end of the downloading,
leaving an unusable file.
@dim, under 64 bit Windows 7?
Could you please find that link and retry?