Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 27 Aug 2015 13:42:58 +0200
From:      John Marino <freebsd.contact@marino.st>
To:        Dmitry Marakasov <amdmi3@amdmi3.ru>, marino@freebsd.org
Cc:        Alexey Dokuchaev <danfe@FreeBSD.org>, ports-committers@freebsd.org, svn-ports-all@freebsd.org, svn-ports-head@freebsd.org
Subject:   Re: svn commit: r395409 - head/astro/geographiclib
Message-ID:  <55DEF7C2.4080901@marino.st>
In-Reply-To: <20150827113301.GH41937@hades.panopticon>
References:  <201508270944.t7R9ia0R067164@repo.freebsd.org> <20150827100953.GA78107@FreeBSD.org> <20150827105912.GF41937@hades.panopticon> <55DEEFC2.40309@marino.st> <20150827113301.GH41937@hades.panopticon>

next in thread | previous in thread | raw e-mail | index | archive | help
On 8/27/2015 1:33 PM, Dmitry Marakasov wrote:
> * John Marino (freebsd.contact@marino.st) wrote:
> 
>>>>> New Revision: 395409
>>>>> URL: https://svnweb.freebsd.org/changeset/ports/395409
>>>>>
>>>>> [...]
>>>>>  GRAVITY_NAMES:=		egm84 egm96 egm2008 wgs84 grs80
>>>>> -MAGNETIC_NAMES:=	wmm2010 wmm2015 igrf11 igrf12 emm2010
>>>>> +MAGNETIC_NAMES:=	wmm2010 wmm2015 igrf11 igrf12 emm2010 emm2015
>>>>
>>>> Why := (assignment with immediate expansion) is used here, esp. for constant
>>>> values?
>>>
>>> Ask maintainer.
>>
>> I thought it had been clearly established (by mat@ IIRC) that this kind
>> of response is not acceptable as a reason for ports mistakes.  The
>> committer is supposed to review the port he or she is committing and
>> catch errors.  The maintainers simply don't have the same level of
>> experience as committers do.
> 
> I'll repeat the same thing I've said back then: I don't see it as
> a mistake, that is exactly why I've committed it without modification.
> 
> "There's no difference whether variable is assigned with or without
> expansion as long as the assigned value does not contain other variable
> references" - that's what I'd answer if you asked _why_ I committed it
> as-is, but you've asked another question, and that question may only
> be directed to maintainer.
> 

I consider it a mistake that got past me (and several maintainers that
committed to those lines after me).  Now that I'm thinking about it, it
may have been part of a complex loop to define it where ":=" was
necessary and I simplified it but didn't change the assignment operator.

Regardless, there's been a pattern of "well, in my opinion this is fine"
defenses even when it's against the PHB and I don't get the point of
even having documentation if people are going to use their judgement
even if contradicts 90% of everyone else.

I think judgement comes into play when there's no documentation.  I
would be interested in knowing if documentation states to prefer lazy
evaluation over immediate assignment.  Obviously that would affect this
ongoing discussion that seems to show up in multiple forms.

John



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?55DEF7C2.4080901>