svn commit: r396248 - head/games/scummvm

[Adam Weinberger]

> On 7 Sep, 2015, at 20:03, Kubilay Kocak <koobs at FreeBSD.org> wrote:
> 
> On 8/09/2015 12:00 PM, Kubilay Kocak wrote:
>> On 7/09/2015 5:08 PM, Alexey Dokuchaev wrote:
>>> On Mon, Sep 07, 2015 at 08:46:51AM +0200, Mathieu Arnold wrote:
>>>> +--On 7 septembre 2015 06:25:11 +0000 Alexey Dokuchaev <danfe at FreeBSD.org>
>>>> wrote:
>>>> | New Revision: 396248
>>>> | URL: https://svnweb.freebsd.org/changeset/ports/396248
>>>> | 
>>>> | Log:
>>>> |   - Ensure that build logs in bulk (package) mode are useful by passing
>>>> |     the --enable-verbose-build argument to configure script
>>>> 
>>>> Why not always make the logs useful ? If I go and run make, I'll have some
>>>> output that's useless for me to debug, right ?
>>> 
>>> Typically yes, I agree; I also prefer to always have full logs.  However,
>>> a lot of people prefer to see them when needed, not always; this perception
>>> could be popularized by Linux kernel build being quite by default, OpeWRT,
>>> etc.  Our own Uses/cmake.mk defines CMAKE_VERBOSE for patch/package builds
>>> only.  On the other hand, with pkg(8) and binary packages becoming widely
>>> adopted, perhaps we can assume that building from the ports is expert mode
>>> and thus --enable-verbose-build/CMAKE_VERBOSE=yes/etc. should be rightful
>>> default.
>>> 
>>> I was trying to preserve existing defaults, yet not pessimize cluster logs
>>> usefullness.  I think if we could have consensus among ports folks that we
>>> always prefer verbose logs, I'm all for removing the checks and just always
>>> pass it.  In fact, this will make quite a few ports less cluttered.
>>> 
>>> ./danfe
>>> 
>> 
>> That (preserving/respecting current defaults) is the right thing to do.
>> 
>> I advocate verbose by default across the board in the ports I maintain
>> already. Our defaults should be conducive to debug-ability, particularly
>> for the case of users providing feedback for maintainers.
>> 
>> How about we exp-run for GNU_CONFIGURE=yes ports:
>> 
>> CONFIGURE_ARGS+=	--disable-silent-rules
>> MAKE_ENV+=		V=1 (maybe even MAKE_ARGS)
>> 
>> OR:
>> 
>> Templates/config.site = enable_silent_rules=no, and
>> MAKE_ENV+= V=1 (maybe even MAKE_ARGS)
>> 
>> AND maybe even:
>> 
>> Quoting from:
>> https://www.gnu.org/software/automake/manual/html_node/Automake-Silent-Rules.html
>> 
>> "Still, notwithstanding the rationales above, a developer who really
>> wants to make silent rules enabled by default in his own package can do
>> so by calling AM_SILENT_RULES([yes]) in configure.ac. "
>> 
>> Could we also REINPLACE_CMD that?
>> 
>> Quoting relevant snippets from above link:
>> 
>> Passing --enable-silent-rules to configure will cause build rules to be
>> less verbose; the option --disable-silent-rules will cause normal
>> verbose output."
>> 
>> At make run time, the default chosen at configure time may be
>> overridden: make V=1 will produce verbose output, make V=0 less verbose
>> output.
>> 
>> Also, in projects using libtool, the use of silent rules can
>> automatically enable the libtool’s --silent option
>> 
> 
> There's also no reason we couldn't wrap what we do in a user
> option/knob, so that builds could be made verbose/silent in particular
> cases, or at the users command.

Very much +1'd here. Putting PORTS_VERBOSE_BUILD=yes into /etc/make.conf makes it real easy for people to enable it if they want. Or the opposite if we want it defaulting to on.

It also becomes something that poudriere can set intelligently. bulk would likely want it defaulting to quiet (normal) mode, and testport would likely want it defaulting to verbose mode.

I just don't think that "this would be helpful to me" should have to equal "therefore everybody is going to want this change."

# Adam


-- 
Adam Weinberger
adamw at adamw.org
http://www.adamw.org