Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 12 Sep 2011 23:55:56 +0200
From:      Michal Varga <varga.michal@gmail.com>
To:        gabor@FreeBSD.org
Cc:        ports@freebsd.org
Subject:   Thank you (for making the ports less boring).
Message-ID:  <1315864556.1747.103.camel@xenon>

next in thread | raw e-mail | index | archive | help
Dear maintainer of databases/gdbm,

Thank you for committing the latest update of gdbm with a (nowhere to be
found mentioned in UPDATING) shared library bump, which again makes my
very dull FreeBSD installation a little bit more fun to maintain, and
especially - *use*

After all, it's been now running for close to a week without almost any
minor ports breakage, so I presume that alone warrants a quick fix to
that whole very boring and uneventful situation, of course:

$ mplayer
/libexec/ld-elf.so.1: Shared object "libgdbm.so.3" not found, required
by "libpulse.so.0"

[And I'm not going bother posting the remaining epic crashing cascade of
all the other running PA applications, because then I might like, even
start to sound somewhat negative-ish for a short moment, and who would
possibly want to do that in the wholly beautiful day like this?]

Anyway, what I was going to say in the first place...

Ah, right - because, there is like, no possibility that a thing like
gdbm might be actually used as a dependency somewhere, like for example,
this tiny little unimportant piece of whateverish whatever called
pulseaudio, I presume?

I mean, what the heck, it's not like it's something as important as,
say, a major sound system heavily employed in at least one of the two
largest desktop environments of the current Unix world, so who could
even care?

Or, why even bother with a quick grep or some recursive checking? It's
not that if the update breaks anything, nobody will step up to complain,
and not that those crazy people out there can be seriously using FreeBSD
as their main desktop OS, heck, FreeBSD even haz sound nao?

And, anyway, it's Monday, real men would *hate* to start their weeks
without spending an hour or few fixing a cascade of new port breakages,
that goes without a question, no arguments from me.

"So yeah, ok," one would probably mutter to himself, "I see it's simply
that time of week again, so I'll just, like, rebuild everything
recursively to catch the latest bump as we do now every odd week anyway
and just get over it, not that there is anything better to do as the
system is already totally unusable (again).."

So yeah, that's easy fix, here we go then:

--->  Reinstalling 'pulseaudio-0.9.22' (audio/pulseaudio)
--->  Build of audio/pulseaudio started at: Mon, 12 Sep 2011 21:40:13 +0200
--->  Building '/usr/ports/audio/pulseaudio'
===>  Cleaning for pulseaudio-0.9.22
===>  License check disabled, port has not defined LICENSE
===>  Found saved configuration for pulseaudio-0.9.22
===>  Extracting for pulseaudio-0.9.22
=> SHA256 Checksum OK for pulseaudio-0.9.22.tar.gz.
===>  Patching for pulseaudio-0.9.22
===>   pulseaudio-0.9.22 depends on package: libtool>=2.4 - found
===>  Applying FreeBSD patches for pulseaudio-0.9.22
===>   pulseaudio-0.9.22 depends on executable: gmake - found
===>   pulseaudio-0.9.22 depends on file: /usr/local/libdata/pkgconfig/x11.pc - found
===>   pulseaudio-0.9.22 depends on file: /usr/local/libdata/pkgconfig/sm.pc - found
===>   pulseaudio-0.9.22 depends on file: /usr/local/libdata/pkgconfig/xtst.pc - found
===>   pulseaudio-0.9.22 depends on file: /usr/local/libdata/pkgconfig/ice.pc - found
===>   pulseaudio-0.9.22 depends on package: libtool>=2.4 - found
===>   pulseaudio-0.9.22 depends on file: /usr/local/bin/intltool-extract - found
===>   pulseaudio-0.9.22 depends on executable: pkg-config - found
===>   pulseaudio-0.9.22 depends on shared library: samplerate.1 - found
===>   pulseaudio-0.9.22 depends on shared library: speexdsp.1 - found
===>   pulseaudio-0.9.22 depends on shared library: dbus-1.3 - found
===>   pulseaudio-0.9.22 depends on shared library: gdbm.3 - not found
===>    Verifying install for gdbm.3 in /usr/ports/databases/gdbm
===>  Extracting for gdbm-1.9.1
===>  Patching for gdbm-1.9.1
[...]
===>  Installing for gdbm-1.9.1
===>   Generating temporary packing list
===>  Checking if databases/gdbm already installed
===>   gdbm-1.9.1 is already installed
*** Error code 1

Stop in /usr/ports/databases/gdbm.
*** Error code 1


Yeeey.

So I guess not.


Now is it what we think it is?

$ grep gdbm /usr/ports/audio/pulseaudio/Makefile
		gdbm.3:${PORTSDIR}/databases/gdbm \

Color me surprised, at this point.


But I know. I know... It's not like the shared library bump was actually
mentioned in the initial PR in the first place or anything...

http://www.FreeBSD.org/cgi/query-pr.cgi?pr=ports/160250

Oh. Right. It was.

So, like... It's not that there is, actually, somehow, anyhow, possible
to, dunno, maybe - NOT COMMITTING THE DAMN THING ALONE when there's
absolutely positively 100% guarantee that stuff depending on it WILL
INSTANTLY AND COMPLETELY BREAK.

Because, not like that a one single quick look would reveal that.

And I'm not even mentioning some actual real testing AND properly
solving the case for the whole dependency tree first, BEFORE committing
just the one port alone, because... If that was happening, what would
then be all those end-users even doing all those empty days, other than
rolling back ports and catching up/fixing breakages? For myself, I
surely can't even imagine.

So again, thank you for taking your part in ensuring that my days with
FreeBSD (the remaining few, so to say) won't become too boring. It's
much appreciated, really.

m.

-- 
Michal Varga,
Stonehenge (Gmail account)





Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1315864556.1747.103.camel>