Files are being corrupted on upload.
I load a zipped file regularly and have done for years. Same software, same platform but suddenly the files are corrupt. I have proved it is Yahoo doing this - started on 10th December - possibly 9th
Issue fixed in December 2013
Donald Hellen commented
This could be a disaster for many groups that access files often.
I can't really test this because I am unable to download files. I have not been able to download files since Yahoo changed it from being able to view them on the group site.
Our group seems to affected as well. Recent uploaded files are unreadable.
Alex Windancer commented
Moderators and owners are having a difficult time with this problem. Our members use the files, and none of us can read a file now. We depend on the files to share information. Members are leaving groups due to file problems. Without files much of our information is not available to our groups members
On my group, files uploaded on December 12 were corrupted. However, files uploaded since December 13 are OK.
My group is having the same issue. I'll describe it here...
There has been a clear line of a breakdown in viewing uploaded files. Those that were uploaded (and downloaded) prior to today are impossible to open. They are clearly jpeg images, as they have been for the past 11 years, yet there is an error message that reports they are not jpegs. Anything uploaded prior to last night will still open, however, It isn't related to a browser. I use Chrome. One member reporting the problem uses IE. Another reporting the problem uses Firefox.
The error message reports a decode error. It states < Not a JPEG file: starts with 0xc3 0xbf >
One member's analysis of the problem is: < For me it looks as if the first bytes of the image are wrong.
A correct JPEG image usually starts with (in hexadecimal)
FFD8 FFE0 0010 4A46 4946 ...
the images with the problem start with
C3BF C398 C3BF C3A0 0010 4A46 4946 ....
You can see there are four extra bytes at the beginning and they have wrong content. So, the image cannot be interpreted by the image handling programs. I have no idea how this may happen, but probably during the upload process. >
Additional technical input adds:
C3BF = FF That is, it is the UTF-8 encoding for the code value FF.
C398 = D8
C3BF = FF
C3A0 = E0
So, apparently this is yet more of the character set silliness, carried to the extreme of applying it to the content of a binary file. In other words they treated the file as a text file in ISO-8859-1 and converted it to UTF-8.
That comment continues with:
The good news is that one could, in principle, write a simple program to translate the file back. Any byte with the value C0 to C3 is the lead-in of a two-byte sequence that converts to a single byte with a value between 80 and FF. Any byte with a value from 00 to 7F stands as itself. If I'm correct there won't by bytes of any other value range.
The bad news, of course, is that if it is an upload bug then likely any file uploaded while the bug is in place is likely lost (barring a decoding app).
Is there any way to escalate this question from the forum to someone working within Yahoo's engineering who can solve this problem? Two major bugs have now rendered our group useless after 11 years of relatively smooth operation. Not good.