Re: [Monotone-devel] Solaris breakage

[William Uther]

On 24/07/2007, at 4:39 PM, Richard Levitte wrote:

> Question, are you getting the exact same error in a different Solaris
> environment?  I'd like to see a "second opinion", so to say.  You see,
> I'm wondering if zfs is coming into play here:

Nah - I was looking at the buildbot output.  Sorry.

> willu.mailingLists> mtn: error: could not remove '/zfs_space/export/ 
> home/mtbldbot/slave-dir/full-sparc-solaris10-sunstudio11/build/ 
> tester_dir/clone_warning_with_multiple_heads/test_dir1': Invalid  
> argument
>
> Especially the "Invalid argument" bullshit (sorry) gives me a number
> of red flags.  Could there be a special file in there that doesn't get
> removed?

Hrm.  If by "Special file" you mean the database that should have  
been created in:

/zfs_space/export/home/mtbldbot/slave-dir/full-sparc- 
solaris10sunstudio11/build/tester_dir/ 
clone_warning_with_multiple_heads/test_dir1/_MTN/mtn.db

Removing that file is what causes problems on windows.  But there no  
error when that file is removed.  There is no error when the '_MTN'  
directory is removed.  The error only happens when 'test_dir1' is  
finally removed.

Has anyone considered altering the test to remove the '1' on the end  
of the path, like this:

check(nodb_mtn("--branch=testbranch", "clone", testURI, "test_dirA"),
         1, false, true)
xfail_if(ostype == "Windows", not exists("test_dirA"))

That shouldn't help... but?

> If you're worried about verbosity (what is it with Unixly people and
> verbosity?????), make it a debugging thingy, not a warning that get's
> output always...

It still slows down every recursive remove.  I'd prefer that it was  
added by the person debugging and then removed afterwards.   
Especially if it turns out not to be useful... which is likely.

Cheers,

Will         :-}