Presumably, we create the ~/.buildr directory so buildfiles/plugins can
write into it.

I changed the code such that the directory is only created if ENV['HOME'] is
set, the directory exists and is writable.

I do think buildfiles/plugins should handle this condition on a case-by-case
basis: either fail, warn the user or continue silently depending on what
they're doing.  The existence of ~/.buildr should be optional.

alex


On Thu, Mar 4, 2010 at 8:43 AM, Antoine Toulme <[email protected]>wrote:

> No rush. Here is maybe one way to look at it (I didn't look at how you
> modified the code yet though):
>
> You can expect that folder to be created and existing if ENV['home'] is
> set.
>
> WDYT ?
>
>
> On Thu, Mar 4, 2010 at 08:36, Alex Boisvert <[email protected]>wrote:
>
>> Yes, known issue...  I was waiting to make up my mind on it.  I'll clean
>> it up.
>>
>> alex
>>
>>
>>
>> On Thu, Mar 4, 2010 at 8:21 AM, Antoine Toulme 
>> <[email protected]>wrote:
>>
>>> This spec fails ./spec/core/application_spec.rb:28 with this
>>> message 'Buildr::Application home_dir should point to existing directory'
>>> FAILED expected true, got false
>>>
>>> I think this is related to the change you introduced recently Alex, on
>>> one
>>> of Ittay's bugs, where you removed the line that creates the .buildr
>>> folder
>>> in the home folder because in some setup there is no home folder.
>>>
>>> Should we just remove that spec ?
>>>
>>> Thanks,
>>>
>>> Antoine
>>>
>>
>>
>

Reply via email to