[cfe-commits] [PATCH] Adds a RecursiveASTVisitor test

[Manuel Klimek]

On Wed, Apr 18, 2012 at 5:20 PM, Benjamin Kramer
<benny.kra at googlemail.com> wrote:
>
> On 18.04.2012, at 17:08, Manuel Klimek wrote:
>
>> On Wed, Apr 18, 2012 at 5:06 PM, Benjamin Kramer
>> <benny.kra at googlemail.com> wrote:
>>> On 18.04.2012, at 16:40, Manuel Klimek wrote:
>>>
>>>> Hi,
>>>>
>>>> the attached patch adds a test for RAV based on the tooling
>>>> infrastructure. It also includes some FIXME tests which require fixes
>>>> in RAV itself / the AST and which are basically impossible to test
>>>> with an integration level test.
>>>>
>>>> Some of the classes at the beginning in the file are probably going to
>>>> be pulled out when other tests want to use them (for example, in the
>>>> tooling branch I use the TestVisitor for tests of the refactoring
>>>> library). I have no idea where to pull them though, and wanted to get
>>>> feedback on the general idea first. My general feeling is that it's
>>>> still too much overhead to pull out a test like this, but I'd rather
>>>> have smaller steps here than overarchitect a solution.
>>>>
>>>> Thoughts?
>>>
>>> Can this be merged into the "Tooling" unittest binary? We already have enough binaries around that link virtually everything of clang, slowing down build times for no reason.
>>
>> I can link it any way you want :) In fact, we can link together a
>> whole bunch of unit tests into a single binary, if you're concerned
>> about linking times...
>>
>> Do you propose to also put the code into Tooling/ or leave it under AST/?
>
> I think our build system requires to have them all in Tooling/ so it's not like we have an option. ;)

Oh, the makefile build - I really need to not forget about it all the time ...
Attached an updated patch.

Cheers,
/Manuel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ravtest.patch
Type: application/octet-stream
Size: 7761 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20120418/136077a5/attachment.obj>