Assume the Stat size of links is zero (and read them instead)
On some virtual filesystems (such ash LucidLink), reading a link size via a
Stat call always returns 0.
However, on unix it reads as the length of the text in the link. This may
cause errors like this when syncing:
Failed to copy: corrupted on transfer: sizes differ 0 vs 13
Setting this flag causes rclone to read the link and use that as the size of
the link instead of 0 which in most cases fixes the problem.
Fixes#4950
Signed-off-by: Riccardo Iaconelli <riccardo@kde.org>
TestParseDuration relied on an elapsed time calculation which
would vary based on the system local time. Fix the test by not relying
on the system time location. Also make the test more deterministic
by injecting time in tests rather than using system time.
Fixes#4529.
This is an attempt at rewriting the rclone filter documentation page.
I have drawn largely from what appears to be the strong original
structure of the page; existing text, and forum comments.
The term flag is used throughout rather than differentiating `--`
options with more complex arguments. That diverges from some standard
practice but is consistent with messages in the rclone binary and `go`
documentation.
The term directory not folder is used throughout.
I tried referring to objects more broadly rather than files and it
just did not seem to work. Apart from a note at the top the
explanations refer entirely to paths, directories and files. My
justification is that bucket store users understand the concept of
files. Not all users of directory aware storage are so familiar with
objects, keys and metadata.
Many of the changes I have made involve moving issues into what seemed
to me to be more relevant parts of the original page structure. I
still find the content repetitious and overly long but that may be
inevitable when users can only be expected to read the section of the
page they think most relevant.
I have eliminated the rsync section from the original structure. It is
hard enough explaining how rclone filters work without also setting
out how they do not. Comment on sync is instead relegated to a
paragraph in the directory filter section.
The structure of the page is intended to work with a hugo toc card
from html Header2 to Header3.
My original intention was to establish a separate examples section. I
have instead retained examples in each section, added to them and
tried to make clear what is documentation and what example.
The changes draw on Github and Forum issues too numerous to mention.
for instance:
https://forum.rclone.org/t/certain-exclusion-flags-seem-to-be-ignored/20049/2
I am **especially** grateful for
https://forum.rclone.org/t/object-key-remote-directory-filter-clarification/20386/2
for making sense of directory filters for me.
@ncw has a fun (and useful) online filter app at
https://filterdemo.rclone.org/ I have not referred to it at this stage
though I particularly like the fact that it is tied to the same
codebase as an rclone version.
I have added cautions about mixing the `--filter...` flags with
`--exclude...` or `--include...`. The same issues seem to arise as
already recognised between the latter two.
The formal summary of glob syntax introduced at the top of the page is
shamelessly stolen from https://godoc.org/github.com/gobwas/glob
I have tried not to alter too many header descriptions and thereby
break existing links to them.
The reference to 'lass' in the example has been retained to confuse
all those not of Scottish or Yorkshire heritage.
Some of my activity was to remove ambiguity and I anticipate
suggestions to roll that back where it has become overly complex.
I tried particularly to bring together and make clear material about
directory filters. It was previously scattered throughout the page and
I couldn't understand it. I am particularly grateful for the
explanations I received about directory filters though any remaining
errors are entirely my own.
Removed erroneous references to non existent `--filter...` flags.
In some ways the best person to write this page would be one with no
knowledge whatsoever of how rclone filters work. The further I got
into it the better qualified I found myself to be.
E&OE
Before this change using --vfs-cache-mode full and --buffer-size 0
together caused the vfs downloader to open more and more downloaders.
This is fixed by introducing a minimum size of 1M for the window to
look for an existing downloader.
Fixes#4892
This patch changes to using the default page limit for listing
unfinished multpart uploads rather than 1000. 1000 is the maximum
specified in the docs, but setting anything larger than 200 gives an
error.
- fix test case FsNewObjectCaseInsensitive (PR #4830)
- continue PR #4917, add comments in metadata detection code
- add warning about metadata detection in user documentation
- change metadata size limits, make room for future development
- hide critical chunker parameters from command line
Before this patch chunker required that there is at least one
data chunk to start checking for a composite object.
Now if chunker finds at least one potential temporary or control
chunk, it marks found files as a suspected composite object.
When later rclone tries a concrete operation on the object,
it performs postponed metadata read and decides: is this a native
composite object, incompatible metadata or just garbage.
