Max recursion level (500) exceeded.

[Adam Weinberger]

Adam Weinberger wrote:
> Lowell Gilbert wrote:
> 
>> Lowell Gilbert <freebsd-ports-local at be-well.ilk.org> writes:
>>
>>
>>> Kris Kennaway <kris at FreeBSD.org> writes:
>>>
>>>
>>>> On Tue, Mar 15, 2005 at 05:58:57PM -0500, Lowell Gilbert wrote:
>>>>
>>>>
>>>>>>> I get it on "make clean" for editors/openoffice-1.1, but I haven't
>>>>>>> managed to track the problem down.  With an empty environment and an
>>>>>>> empty make.conf.  But that takes so long to build anyway that I 
>>>>>>> don't
>>>>>>> really care very much...
>>>>>>
>>>>>>
>>>>>> Let us know if/when you study it in more detail knowing the above
>>>>>> probable cause :-)
>>>>>
>>>>>
>>>>> I knew the "usual cause."  That's why I tried it with an empty
>>>>> make.conf and a minimal environment, under various shells.  And a
>>>>> simple "make conf" still goes into spasms of recursion.  
>>>>
>>>>
>>>> It doesn't for me, so you'll still need to debug this some more on
>>>> your own or at least post some details.
>>>
>>>
>>> So far, I've tracked it down to calculating the all-depends-list in
>>> lang/gcc32.  And if I deinstall gcc32, the problem goes away.
>>
>>
>>
>> The problem comes from the USE_GCC variable in the gcc32 Makefile.
>> Is USE_GCC broken for "ORLATER" if a later gcc is already installed?
> 
> 
> ENOPARSE on that last line, sir.
> 
> That USE_GCC line shouldn't be creating any sort of BROKEN assignment. 
> And it shouldn't choke if gcc > 3.2 (or gcc > 2.7) is installed.
> 
> The only thing I'm confused about is why openoffice-1.1 has an explicit 
> BUILD_DEPENDS on gcc32, rather than just using USE_GCC=3.2.
> 
> However, you are most certainly correct in that when I deinstalled 
> gcc32, make clean went smoothly and quickly.

The problem lies in the gcc32 port. It redefines USE_GCC. It, and the 
other gcc ports, need not to do that.

# Adam


-- 
Adam Weinberger
adamw at magnesium.net || adamw at FreeBSD.org
adamw at vectors.cx    ||   adamw at gnome.org
http://www.vectors.cx