#13612: BTRFS: Adding write support
----------------------------------+----------------------------
Reporter: hyche | Owner: korli
Type: enhancement | Status: new
Priority: normal | Milestone: Unscheduled
Component: File Systems/Btrfs | Version: R1/Development
Resolution: | Keywords: gsoc2017
Blocked By: | Blocking:
Has a Patch: 1 | Platform: All
----------------------------------+----------------------------
Comment (by hyche):
Replying to [comment:10 pulkomandy]:
In the first change (style fixes), in btrfs_read_attr_dir:B_OK instead. Now it will fail. Is that intentional?
{{{
+ if (status != B_OK)
+ return status;
+
if (status == B_ENTRY_NOT_FOUND) {
*_num = 0;
return B_OK;
- } else if (status != B_OK)
- return status;
+ }
}}}
The previous code would ignore the B_ENTRY_NOT_FOUND error and return
I remember I was the one suggesting BTreeNode and BTreePath, but readingthis code again I > think using BTree::Node and BTree::Path may make more
Now we have a Logical() here that is not the same thing as the Logical()in
8ee5e1382f7cc0bf66bb317e84a7295e8c07b84d. What was wrong withLogicalRoot()?
I assume the goal of the "read" flag is to create the entry if it doesnot exist, for write >support. If that is the case, then createIfNotFound
Add more line breaks in commit message.Done.
Removing the whole test of {{{ if (to < from || from < 0 || to >=ItemCount() || ItemCount() == > 0) }}}
sounds a bit dangerous, why not remove just the test for ItemCount() ==0 from it?