test-mq on FreeBSD

Adrian Buehlmann adrian at cadifra.com
Wed May 28 10:08:31 CDT 2008

On 28.05.2008 16:56, Maxim Dounin wrote:
> Hello!
> On Wed, May 28, 2008 at 04:05:18PM +0200, Adrian Buehlmann wrote:
>> On 28.05.2008 15:22, Maxim Dounin wrote:
>>> Hello!
>>> On Wed, May 28, 2008 at 02:59:02PM +0200, Adrian Buehlmann wrote:
>>>> Just a minor point. When running the testsuite on FreeBSD 6.2, I always get:
>>>> ERROR: test-mq output changed
>>>> --- Expected output
>>>> +++ Test output
>>>> @@ -499,7 +499,7 @@
>>>> abort: local changes found, refresh first
>>>> % apply force, should not discard changes with empty patch
>>>> applying empty
>>>> -/usr/bin/patch: **** Only garbage was found in the patch input.
>>>> +patch: **** Only garbage was found in the patch input.
>>>> patch failed, unable to continue (try -v)
>>>> patch empty is empty
>>>> Now at: empty
>>>> Could we agree on using that output on all unixes by doing
>>>> the patch below?
>>> Hmm, looks like you patch isn't FreeBSD native /usr/bin/patch.  
>>> :-)
>> I don't understand you here. I simply do
>> %python run-tests.py test-mq
>> using crew tip and get that annoying changed output I posted.
> Yes.  But the output you get are from gnu patch, not from
> /usr/bin/patch (it's normal - Mercurial uses gpatch instead of 
> patch if it's here).  Without gnu patch on you system you will get 
> another output (I posted it in previous message).

Ah. Thanks for that info. If it helps, I don't mind requiring FreeBSD users
wanting to run the testsuite having to install gnu patch.

For example, I also have to use gmake (not make) to build on

> [...]
>> I don't even understand what the implementation of that hg command
>> is calling here. All I can say is that obviously on my system, that
>> command seems *not* to prepend its install path in the error output.
>> It's the only test that fails on FreeBSD 6.2. All I would appreciate to
>> have changed is not to have to repeatedly stumble upon that output
>> each time I run the testsuite.
>> Since I do not have another unused computer, installing a linux/ubuntu/
>> debian/whatever is not an option for me and wouldn't be worth wasting
>> time for me anyway.
>> But as I said, it's only a minor problem. If my proposal is causing
>> other problems, then ignore it.
> It isn't, but it does solve the problem only for your specific 
> installation.  The patch I suggest just a bit more generic and 
> should fix all such problems.

Ok. You might want to ask on irc about the status of your patch.
Doing so has proven to be helpful in the past.

More information about the Mercurial-devel mailing list