BALL

This page has three main sections. (Note that the listings could be empty.) more… Tip: Remember the back-button! And don't click the number on the left.

  1. The first item displays a term, URL, text or picture.
  2. Followed by a (still unordered) listing of statements about this subject. (for details follow ">"-link)
  3. Separated by a horizontal rule a "reverse" listing of statements referring to this item in object position.
9
1167 comment >

Copying a 35MB file back after upload ran into a "negative destination offset" (from mmap handling under Chicken).

The process of splitting the input into blocks seems buggy.

When the first (the "directory")-block is full (that is in the current default configuration 200 blobs 32KByte each), an indirect directory block is created. The latter holds another set (200 here) of directory blocks. When checked from the failed upload, this first indirect directory blob was found empty.

A stripped down test case using a) 2 blobs of 2 Byte each and b) 8 blobs of 2 byte each per directory appeared to confirmed the splitting process to work correct. Turned out that was not the case. Another bug inhibited the the logic which would once upon a time flush the data as read so far to disk (for such a small blob size).

Once the issue was fixed, it turned out, that the garbage collector for persistent objects could also step in and delete blobs not yet held by any place…

1160 is a > issue
1163 title > Bug when serving large files.
1165 state > solved

ball.askemos.org/Bugs comment
issue container member