Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 28 Jun 2014 01:03:00 +0200
From:      John Marino <freebsd.contact@marino.st>
To:        Gerald Pfeifer <gerald@pfeifer.com>, John Marino <marino@freebsd.org>
Cc:        svn-ports-head@freebsd.org, svn-ports-all@freebsd.org, Gerald Pfeifer <gerald@FreeBSD.org>, ports-committers@freebsd.org
Subject:   Re: svn commit: r359090 - head/lang/gcc47
Message-ID:  <53ADF824.4070401@marino.st>
In-Reply-To: <alpine.LSU.2.11.1406272352330.30120@tuna.site>
References:  <201406241448.s5OEmqQ9057556@svn.freebsd.org> <53A991D8.1040403@marino.st> <alpine.LSU.2.11.1406272352330.30120@tuna.site>

next in thread | previous in thread | raw e-mail | index | archive | help
On 6/28/2014 00:10, Gerald Pfeifer wrote:
> On Tue, 24 Jun 2014, John Marino wrote:
>>> [snip]
>>>   Also, since this is now final and stable, no longer bootstrap this port.
>> I'm not following the "don't bootstrap anymore" logic here.
>>
>> This implies that one bootstraps due to active development, which is 
>> not what I would claim.
>>
>> I would say unless GCC 4.7.x or later is building this gcc 4.7 port,
>> then bootstrapping by default is a good idea.  And it seems that gcc
>> 4.2.1 or clang 3.x is what builds this port by default.
> 
> Bootstrapping (or not) always is a tradeoff.
> 
> To provide faster builds for users of the default lang/gcc and
> older branches, which should be rather stable, I am defaulting
> those to not bootstrap.  This is based on feedback I received
> over the years in favor of faster build times.  
> 
> Users can always opt to enable that option if they desire so, 
> of course.

Well, again, the (in)stability of a branch is not why would choose to
bootstrap.  In the case of gcc 4.2.1 building current gcc, I would
*always* bootstrap.  In other words, it has mainly to do with the
compiler used to build gcc.  For that matter, I would also bootstrap if
clang was building gcc.

This is why changing the bootstrap option default simply because no more
releases on a branch are coming makes zero sense to me.  If branch
stability is the criterion for the option default, you might as well
switch it on the first release as the subsequent ones are strictly bug
fix releases.

John



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?53ADF824.4070401>