Stefan Hajnoczi writes:
> On Wed, Apr 18, 2012 at 2:37 PM, Lluís Vilanova <vilan...@ac.upc.edu> wrote:
>> Stefan Hajnoczi writes:
>>
>>> On Sat, Apr 14, 2012 at 12:19:03AM +0200, Lluís Vilanova wrote:
>>>> - rm -f $(GENERATED_HEADERS)
>>>> - rm -f $(GENERATED_SOURCES)
>>>> + rm -f $(foreach f,$(GENERATED_HEADERS),$(f) $(f)-timestamp */$(f)
>>>> */$(f)-timestamp)
>>>> + rm -f $(foreach f,$(GENERATED_SOURCES),$(f) $(f)-timestamp */$(f)
>>>> */$(f)-timestamp)
>>
>>> Why */$(f) */$(f)-timestamp?
>>
>> Some of the files are generated in immediate subdirectories (e.g.,
>> libuser/trace.c-timestamp).
>>
>> I could use the results of a call shell to "find" instead, so that it will
>> always find the right victims no matter where they are.
> Ah, I get it. This actually adds something that was missing previously.
> I'm not a make expert, I think if we try we can find issues with any
> approach to removing the files. What you've done seems fine to me :).
What I meant is the following, so that it will always be correct no matter how
directories are organized in the future:
rm -f $(foreach f,$(GENERATED_HEADERS),$(shell find $(BUILD_DIR) -name $(f)
-o -name $(f)-timestamp))
rm -f $(foreach f,$(GENERATED_SOURCES),$(shell find $(BUILD_DIR) -name $(f)
-o -name $(f)-timestamp))
I'll resend this patch with the aforementioned changes.
Lluis
--
"And it's much the same thing with knowledge, for whenever you learn
something new, the whole world becomes that much richer."
-- The Princess of Pure Reason, as told by Norton Juster in The Phantom
Tollbooth
