[PATCH 2 of 3 STABLE] tests: use -f when clearing largefile usercache

[Matt Harbison]

Mads Kiilerich wrote:
> Matt Harbison wrote, On 10/23/2012 06:04 AM:
>> # HG changeset patch
>> # User Matt Harbison <matt_harbison at yahoo.com>
>> # Date 1350962187 14400
>> # Branch stable
>> # Node ID a233e59d55e75f810488a0c94ff2e1e17696ba23
>> # Parent 6269e84db360cec75cb70982079f84fba4e9bf21
>> tests: use -f when clearing largefile usercache
>>
>> A subsequent test will cause an rm without -f to report an error, so
>> all uses
>> were converted to avoid future issues like this.
>
> Which subsequent test will cause which error?
>
> Is it just because one of the rm's will become redundant because you add
> another rm?
>
> I think it is slightly better that we know that the rm's are there for a
> purpose and actually do remove something. If one of them no longer is
> needed then remove it ... or add a -f and a comment there.

Ack.  Sorry, I managed to make this pretty context free, especially with 
the series not threading properly.  (Does patchbomb thread by the commit 
message topic?  I usually use thg, but used 'hg email -r 'mq()' --flag 
STABLE' here, and it printed out the sending lines in order.)

Yes, one of the rm commands became redundant (specifically the existing 
one after the one added in patch #3 [1]).  Changing another unrelated 
test seemed like it could be considered multiple changes so I split it 
out.  That seemed too trivial, so I figured I'd change them all.  But I 
see your point about knowing they are actually doing something, and agree.

I'll resubmit #1 and #3 with the single -f change.

--Matt

[1] 
http://www.selenic.com/pipermail/mercurial-devel/2012-October/045621.html