[Tcl-bugs] [ tcl-Bugs-2003130 ] failing fCmd.test tests on new fileserver

[SourceForge.net <no...@so...>]

Bugs item #2003130, was opened at 2008-06-26 14:35
Message generated for change (Comment added) made by ferrieux
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110894&aid=2003130&group_id=10894

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: 37. File System
Group: development: 8.5.3b1
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Don Porter (dgp)
Assigned to: Vince Darley (vincentdarley)
Summary: failing fCmd.test tests on new fileserver

Initial Comment:

My Tcl sandboxes got moved to a new
fileserver, and all branches now show
these test failures:

==== fCmd-6.16 CopyRenameOneFile: TclpCopyRenameOneFile fails FAILED
==== Contents of test case:

    file mkdir [file join td1 td2]
    createfile [file join td1 td2 tf1]
    file mkdir td2
    file rename -force td2 td1

---- Result was:
error renaming "td2" to "td1/td2": illegal operation on a directory
---- Result should have been (glob matching):
error renaming "td2" to "td1/td2": file *
==== fCmd-6.16 FAILED



==== fCmd-6.18 CopyRenameOneFile: errno != EXDEV FAILED
==== Contents of test case:

    file mkdir [file join td1 td2]
    createfile [file join td1 td2 tf1]
    file mkdir td2
    file rename -force td2 td1

---- Result was:
error renaming "td2" to "td1/td2": illegal operation on a directory
---- Result should have been (glob matching):
error renaming "td2" to "td1/td2": file *
==== fCmd-6.18 FAILED



==== fCmd-9.9 file rename: comprehensive: dir to non-empty dir FAILED
==== Contents of test case:

    file mkdir tds1
    file mkdir tds2
    file mkdir [file join tdd1 tds1 xxx]
    file mkdir [file join tdd2 tds2 xxx]
    if {![testConstraint unix]} {
        testchmod 555 tds2
    }
    set a1 [list [catch {file rename -force tds1 tdd1} msg] $msg]
    set a2 [list [catch {file rename -force tds2 tdd2} msg] $msg]
    if {[testConstraint unix]} {
        set w2 0
    } else {
        set w2 [file writable tds2]
    }
    list [lsort [glob td*]] $a1 $a2 [file writable tds1] $w2

---- Result was:
{tdd1 tdd2 tds1 tds2} {1 {error renaming "tds1" to "tdd1/tds1": illegal operation on a directory}} {1 {error renaming "tds2" to "tdd2/tds2": illegal operation on a directory}} 1 0
---- Result should have been (glob matching):
{tdd1 tdd2 tds1 tds2} {1 {error renaming "tds1" to "tdd1/tds1": file *}} {1 {error renaming "tds2" to "tdd2/tds2": file *}} 1 0
==== fCmd-9.9 FAILED



==== fCmd-9.13 file rename: comprehensive: can't overwrite target FAILED
==== Contents of test case:

    file mkdir [file join td1 td2] [file join td2 td1 td4]
    file rename -force td1 td2

---- Result was:
error renaming "td1" to "td2/td1": illegal operation on a directory
---- Result should have been (glob matching):
error renaming "td1" to "td2/td1": file *
==== fCmd-9.13 FAILED


These are just differences in error message,
perhaps reflecting different values of errno.
I suspect the way to fix is to make the tests
care less about the exact error message wording.


----------------------------------------------------------------------

Comment By: Alexandre Ferrieux (ferrieux)
Date: 2008-06-26 15:33

Message:
Logged In: YES 
user_id=496139
Originator: NO

Indeed when I compare behavior between a local disk and an NFS share, old
versions of Tcl would consistently say "file already exists", while HEAD
says "file busy" on NFS. After a bit of strace, I see that the underlying
errnos are respectively ENOTEMPTY and EBUSY.

Notice that the test case already account for this slight variability (see
comment on fCmd-9.9: Test can hit EEXIST or EBUSY, depending on underlying
filesystem), by using "file *" as error pattern.

However your case is a bit nastier. Can you strace/truss and find the
syscall returning EISDIR (responsible for "illegal operation on directory")
?


----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=110894&aid=2003130&group_id=10894