XYplorer Beta Club

Jibz wrote:It's not as easy as it appears, at least if you want it to work in general.

But if you limit it to what Explorer can do, it's easy.

admin wrote:What do you mean?

Jibz wrote:

admin wrote:What do you mean?

Well, imagine one of the items you are moving is a folder with the same name as the parent folder. Then the order in which you traverse the tree suddenly matters.

http://www.hardtoc.com/archives/198

The algorithm there still has a kink; it doesn't always remove such recursive paths after moving.

Since everybody seems to be ignoring this issue, including Windows Explorer, you can probably get away with doing so too .

admin wrote: I just added the primitive version to the current beta -- could not resist. I was shamelessly encouraged by the fact that "the others" don't do it any better.

Later, after the release of 9.60, I will improve the algorithm to handle "recursive collision" or what you want to call it.

admin wrote:Well, I made it a bit cheaper: I added at least a warning (in v9.60.0002).

Jibz wrote:

admin wrote:Well, I made it a bit cheaper: I added at least a warning (in v9.60.0002).

I think that is a very good solution .

One thing though, if the destination is further up the path, it would be better to check if the entire path from destination to source is present in source instead of just the first folder. For example, if I have

\parent\foo\bar\baz\zoo\zeb\foo

and I am in \parent\foo\bar\baz\zoo\zeb and try to move foo out to \parent I get a warning even though there is no chance of an overlap.

Or in general, if Special is the relative path from Destination to Source (i.e. Source == Destination\Special), then if Source\Special exists there is a possibility of an overlap.

Jibz wrote:

admin wrote:Well, I made it a bit cheaper: I added at least a warning (in v9.60.0002).

I think that is a very good solution .

One thing though, if the destination is further up the path, it would be better to check if the entire path from destination to source is present in source instead of just the first folder. For example, if I have

\parent\foo\bar\baz\zoo\zeb\foo

and I am in \parent\foo\bar\baz\zoo\zeb and try to move foo out to \parent I get a warning even though there is no chance of an overlap.

Or in general, if Special is the relative path from Destination to Source (i.e. Source == Destination\Special), then if Source\Special exists there is a possibility of an overlap.

\---parent
    \---foo
        |   a.txt ; "first"
        |   z.txt ; "first"
        |
        \---foo
            |   a.txt ; "second"
            |   z.txt ; "second"
            |
            \---foo
                    a.txt ; "third"
                    z.txt ; "third"

admin wrote:I made a new experiment with Explorer (identical results with XY). In your first test case (repeated here below), I go into \parent\foo\foo (sic!), and move all contents upward to \parent\foo, and the result is a desaster. Of 6 TXT files, 4 are gone (2 "first" correctly overwritten, but 2 "second" simply lost). Can you confirm this?

admin wrote:I would not have expected this, and now I'm unsure how to test for this case. But maybe, hopefully, I made a mistake somewhere...

Jibz wrote:

admin wrote:I made a new experiment with Explorer (identical results with XY). In your first test case (repeated here below), I go into \parent\foo\foo (sic!), and move all contents upward to \parent\foo, and the result is a desaster. Of 6 TXT files, 4 are gone (2 "first" correctly overwritten, but 2 "second" simply lost). Can you confirm this?

I get the same here. When you move from \parent\foo\foo to \parent\foo, you will get the set of "first" files overwritten, but the set of "second" files disappear because of the overlap issue.

If I try to do the move from \parent\foo to \parent here, Explorer stops with an error in the middle of the move.

admin wrote:I would not have expected this, and now I'm unsure how to test for this case. But maybe, hopefully, I made a mistake somewhere...

I think the test is the same .. the path from destination to source (here foo) is present under source as well, so there is a possible overlap.

admin wrote:OK, I think I got it now. I got shortly lost in the thickage of bar\baz\zoo\zeb\foo...