Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 24 Aug 2012 16:12:26 -0700 (PDT)
From:      Jeffrey Bouquet <jeffreybouquet@yahoo.com>
To:        freebsd-ports@freebsd.org
Subject:   Fw: RE: pkgng default release schedule (s/ABI/API/g ...typo)
Message-ID:  <1345849946.58806.YahooMailClassic@web111314.mail.gq1.yahoo.com>

next in thread | raw e-mail | index | archive | help
s/ABI/API/g   Sorry! =0A=0A=0A=0A--- On Fri, 8/24/12, Jeffrey Bouquet <jeff=
reybouquet@yahoo.com> wrote:=0A=0A> From: Jeffrey Bouquet <jeffreybouquet@y=
ahoo.com>=0A> Subject: RE: pkgng default release schedule (contd...)=0A> To=
: "Chris Rees" <utisoft@gmail.com>=0A> Date: Friday, August 24, 2012, 1:28 =
PM=0A> Comments below. I've no idea how to=0A> fix the quoting system so I'=
ll try to do it manually, Maybe=0A> in a few weeks... it is webmail and its=
-all-text seems no to=0A> function at the moment.=0A> =0A> --- On Fri, 8/24=
/12, Chris Rees <utisoft@gmail.com>=0A> wrote:=0A> =0A> From: Chris Rees <u=
tisoft@gmail.com>=0A> Subject: RE: pkgng default release schedule (contd...=
)=0A> To: "Jeffrey Bouquet" <jeffreybouquet@yahoo.com>=0A> Date: Friday, Au=
gust 24, 2012, 10:57 AM=0A> =0A> =0A> =0A> On 24 Aug 2012 15:37, "Jeffrey B=
ouquet" <jeffreybouquet@yahoo.com>=0A> wrote:=0A> =0A> >=0A> =0A> > Comment=
s inline below...=0A> =0A> >=0A> =0A> > --- On Fri, 8/24/12, Chris Rees <ut=
isoft@gmail.com>=0A> wrote:=0A> =0A> >>=0A> =0A> >>=0A> =0A> >> From: Chris=
 Rees <utisoft@gmail.com>=0A> =0A> >> Subject: RE: pkgng default release sc=
hedule=0A> (contd...)=0A> =0A> >> To: "Jeffrey Bouquet" <jeffreybouquet@yah=
oo.com>=0A> =0A> >> Cc: "FreeBSD Mailing List" <freebsd-ports@freebsd.org>,=
=0A> "freebsd-current" <freebsd-current@freebsd.org>=0A> =0A> =0A> >> Date:=
 Friday, August 24, 2012, 4:37 AM=0A> =0A> >>=0A> =0A> >>=0A> =0A> >> On 24=
 Aug 2012 11:08, "Jeffrey Bouquet" <jeffreybouquet@yahoo.com>=0A> >>wrote:=
=0A> =0A> >> >=0A> =0A> >> > A few more reasons (unless I have not seen=0A>=
 some relevant >>documentation to the contrary) to not=0A> mandate pkgng as=
 the default...=0A> =0A> >>=0A> =0A> >>=0A> =0A> >> Why don't you phrase th=
is as "How can one ..." so=0A> you sound less >>negative?=0A> =0A> >>=0A> =
=0A> >Please fix your mailer's quoting.=0A> It is webmail. Sorry. I am fixi=
ng part of it as I go along=0A> but I've no idea if it is making it worse..=
.=0A> =0A> =0A> =0A> >> I am unequivocably opposed (reasoning below=0A> inc=
luded, after I wrote >>this paragraph; ) to pkg=0A> mandated as the default=
, until convinced=0A> =0A> >>=0A> =0A> >> otherwise, and trying to sound ni=
cer, by making the=0A> sentences >>shorter.=A0 If that is=0A> =0A> >>=0A> =
=0A> >> stubborness-from-ignorance, I=A0 sincerely=0A> apologize, in advanc=
e.=A0 But >>I see no one else=0A> =0A> >>=0A> =0A> >> vocalizing opposition=
 at the moment... vs those in=0A> favor, and am >>simply in my=0A> =0A> >>=
=0A> =0A> >> view voicing arguments to delay the not-inevitable=0A> depreca=
tion of >>/var/db/pkg...=0A> =0A> >Possibly because your attitude as I ment=
ion above makes=0A> people think =0A> >you're trolling.=0A> =0A> I am disma=
yed by the thought that the package management=0A> system is=0A> broken.=A0=
 We have the Makefile directories (starting=0A> point) /var/db/pkg=0A> dire=
ctories (registration feature) portmaster/portmanager/=0A> custom .sh/=0A> =
portupgrade (automation feature) and the installed binaries=0A> (end point)=
.=0A> Pkg introduces an ABI making=0A> portmaster/portmanager/portupgrade/t=
o-be-portupgrades/.sh=0A> costlier to implement in time, ABI experience, =
=0A> and maybe buggier.=A0 Meanwhile, the registration feature=0A> (present=
 in=0A> BSD, absent AFAIK in most Linux, abstracted in Windows),=0A> would =
be=0A> abstracted in BSD also.=A0 =0A> =0A> Suppose I rightly viewed that t=
he ports tree was not working=0A> well=0A> enough for the majority of users=
.=0A> =0A> There is a new tool, portsnap-textonly (psto)=0A> Its use is eas=
y!=0A> psto view makefile editors nano=0A> psto configure editors nano=0A> =
psto view pkgplist editors nano=0A> psto view editors=0A> =0A> I'm fully ju=
stified in putting it somewhere (a port.)=0A> It is only when I remove the =
actual ports tree to a=0A> xml-based or=0A> sql-based repository, that I be=
gin to de-minimalistic=0A> de-potentially-better =0A> Freebsd, and fork it =
to something that only may serve better=0A> a subset=0A> (however large) of=
 its users. Even if I obtain the=0A> majority opinion that it is expedient,=
 and=0A> even if it *is* actually expedient for the=0A> majority, I'd maybe=
 want to quell the=0A> concerns of those very reluctant to pursue the parad=
igm=0A> shift that=0A> it is *not* in their best interests... or do more co=
ding=0A> beforehand=0A> to compensate for the impact it may have upon their=
 use of=0A> the operating system. =0A> =0A> =0A> =0A> =0A> =0A> =0A> >> Cer=
tainly not trying to "just complain" for the=0A> sake of impeding >progress=
. =0A> =0A> >>=0A> =0A> >> > Nowadays, one can save time by installing two=
=0A> ports which >officially or unofficially conflict, and=0A> have /var/db=
/pkg entries for >both, and even=0A> =0A> >> > local workarounds (for insta=
nce, moving the=0A> duplicate binary >elsewhere before the second install)=
=0A> (Perchance removing the line in >the Makefile).=0A> =0A> >>=0A> =0A> >=
> Currently you can still do this, at least until=0A> STAGEDIR.=0A> =0A> =
=0A> =0A> >>=0A> =0A> >>=0A> =0A> >> STAGEDIR? =0A> >Not of concern at the =
moment, but it's in the FreeBSD=0A> wiki if you're >interested.=0A> Thank y=
ou for the information.=0A> >>=0A> =0A> >> > A failed "make install (regist=
er)", one can=0A> check the /var/db/pkg/ >>directory(ies) to double=0A> che=
ck visually to what extent it did NOT >>fail.=0A> =0A> >> > Similarly for a=
 failed "make deinstall=0A> (unregister)"...=0A> =0A> >>=0A> =0A> >>=0A> =
=0A> >> The error messages are perfectly clear.=0A> =0A> >>=0A> =0A> >> "Th=
e error messages are perfectly clear", yes, once=0A> double-checked >>again=
st=0A> =0A> >>=0A> =0A> >> the directories in /var/db/pkg to make them=0A> =
less-terse.=A0 If you mean >>pkg's =0A> =0A> >>=0A> =0A> >> error messages,=
 maybe *this* trepidation is from=0A> inexperience with >>pkg.=0A> See belo=
w.=0A> >>=0A> =0A> >> > pkgdb -F to fixup stale dependencies and=0A> resolv=
e dependency >>information.=0A> =0A> >>=0A> =0A> >>=0A> =0A> >> Unnecessary=
 with pkgng.=0A> =0A> I'll take your word for it.=A0 Thanks for the=0A> inf=
ormation. Howsoever, I've=0A> encountered stale dependencies with pkgdb -F =
after=0A> portmaster had=0A> fixed them up, so I am a little skeptical.=0A>=
 =0A> >>=0A> =0A> >> Okay, I am unaware.=0A> =0A> >Let me say this once mor=
e.=0A> >You have no right to complain about something until=0A> you've prop=
erly >tested it.=A0 =0A> =0A> I am not trying to complain about pkg(ng).=A0=
 I am trying=0A> to dissuade=0A> the deprecation of /var/db/pkg/ as a plain=
-text hook into=0A> the =0A> registration of ports/packages.=A0 =0A> =0A> T=
here is no way I can properly test pkg and a deprecated=0A> /var/db/pkg=0A>=
  without more time and effort.=A0 If you mean to say I=0A> should first in=
stall pkg fully on a machine and compare the=0A> =0A> result to what I am d=
oing now, I decided a few hours ago to=0A> do just=0A> that.=A0 However, I'=
ve a lot of other stuff to do so it=0A> is not a=0A> priority.=A0 If you me=
an to say I should delay stuff like=0A> this posting=0A> until I've tested =
further, I'll apologize here and=0A> now.=A0 But it is=0A> easier and poten=
tially more informative to others not=0A> decided, in=0A> my humble opinion=
, to read these views, if I state them=0A> politely =0A> enough, than for m=
e to write them down as a draft to post=0A> later.=0A> Maybe I will benefit=
 by someone refuting them to my=0A> satisfaction,=0A> and saving you the ef=
fort of another reply.=0A> =0A> =0A> =A0 =0A> =0A> =0A> >You will find that=
 most of the scripts you currently use=0A> are to >compensate for brokennes=
s in pkg_install and are=0A> no longer necessary.=A0 =0A> =0A> Most of the =
scripts I use are to upgrade, I've way too many=0A> ports=0A> installed to =
ever use the pkg_install command without a=0A> front end.=0A> The only ones=
 I usually use are pkgdb -F and pkgdb -u and =0A> portmaster -d -B -P -i -g=
 ...=A0 and the shell does most=0A> of the=0A> heavyweight work, as I save =
history across reboots.=A0=0A> The only=0A> breakage I am aware of are the =
usual ones, ports waiting to=0A> be=0A> fixed, conflicts, clang vs gcc, dif=
ficultes in fetching=0A> should=0A> it be an https://, etc.=0A> =0A> =0A> =
=0A> > Scripting your way around bad design decisions is=A0 not=0A> a way t=
o run >an OS.=0A> Where is the bad design decision specifically?=A0 I view=
=0A> an ABI into the=0A> plain-text port registration as a cure worse than =
the=0A> disease for=0A> the majority of cases.=A0 I see Good Design specifi=
cs...=0A> portmaster=0A> for example.=A0=A0=A0I can run several in several=
=0A> xterms at once and they=0A> hardly ever collide...=0A> =0A> =0A> >> > =
A proven method in the portmaster manpage to=0A> reinstall all ports.=0A> =
=0A> >>=0A> =0A> >> You want to talk to the portmaster author about=0A> tha=
t.=0A> =0A> >>=0A> =0A> >> I never use the procedure, I've too many ports.=
=A0=0A> But others may be >>interested.=0A> >OK, please sort out your own v=
iews, then start worrying=0A> about other >people.=0A> Again, I've tried to=
 state them nicely above, and put forth=0A> a few=0A> reasons why I am stat=
ing them now rather than later. =0A> =0A> >> *Anything* that circumvents=A0=
 portmaster's=0A> fine-grained updating of >>the =0A> =0A> >>=0A> =0A> >> e=
ach-port-its-own-database-directory-in-var-db-pkg,=0A> which is=A0 a >>grea=
t help=0A> =0A> >>=0A> =0A> >> to myself daily, is in my view, unwarranted,=
=0A> untimely, bloat, >>contrary to the=0A> =0A> >>=0A> =0A> >> keep-it-sim=
ple port-upgrade tools "fixing it when=0A> it breaks" >>methodologies that=
=0A> =0A> >>=0A> =0A> >> exist at the present time, etc. =0A> =0A> >You mea=
n, let's cobble together yet more hacks and=0A> continue to have a >package=
 management system that is=0A> fifteen years out of date?=0A> I view portma=
ster, used above, as=0A> cutting-edge.=A0=A0=A0If the /var/db/pkg=0A> flat-=
file is out-of-date vs an ABI into an equivalent,=A0=0A> I see that=0A> as =
a fork away from minimalism, and as a disservice to=0A> ground-up=0A> based=
 port-management tools.=A0 =0A> If you are trying to say that the pkg based=
=0A> port-registration will=0A> work less-as-a-hack, than portmaster and sh=
ell scripting, it=0A> could=0A> be that I may switch to that view after tes=
ting, in which=0A> case I=0A> am fully too dismissive of pkg's potential.=
=A0 But in my=0A> real-world=0A> experience so far, I view it as only a rem=
ote possibility,=0A> since=0A> it is not the day-to-day stuff I view as par=
amount, but the=0A> =0A> recourses one has upon unexpected failures when re=
building,=0A> say,=0A> perl from one version to another.=A0 I've no reports=
 of=0A> someone =0A> attempting that with pkg so far...=0A> =0A> =0A> >I'm =
sorry that you enjoy messing with these files and=0A> are not going to >be =
able to do it in future.=A0 The hard=0A> fact is that the people who do >th=
e work in bringing you=0A> ports are sick of the inadequacies of the >curre=
nt=0A> system, and are improving it.=0A> =0A> Statement taken as an opinion=
 unless specifically elaborated=0A> upon... I kind of have *fun* doing a =
=0A> cat ffil_0812 | grep p5 | awk '{print $1}' | xargs -J %=0A> portmaster=
 -d -B -P -i -g --update-if-newer % =0A> ...typing only a few characters in=
 it (from history), rinse,=0A> repeating=0A> for /devel/, /ftp/ ... etc... =
whereas before portmaster=0A> existed,=0A> portupgrade would upon the lower=
-power machines monopolize=0A> the CPU, etc=0A> since I can *automate* it a=
t a machine I am not physically=0A> at during most of the several hours...=
=A0 I see that as a=0A> not-inadequacy... it=0A> is, again, the command-lin=
e (non-abi) hook into the=0A> registration details=0A> that speeds (here) f=
ix of breakages.=A0 So maybe define=0A> inadequacy =0A> more explicitly?=0A=
> =0A> >Do some research and try it out.=A0 You'll be amazed.=0A> =0A> As I=
 mentioned, I may be able to install it on one=0A> machine.=A0 But =0A> I a=
m very pessimistic that its abstraction layer will be=0A> compensated for b=
y any more-than-empirical improvement...=0A> Thanks=0A> for the optimism th=
ough.=0A> =0A> =0A> =0A> >Chris=0A> =0A> 



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1345849946.58806.YahooMailClassic>