Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 14 Jul 2013 11:50:50 -0700
From:      Adrian Chadd <adrian@freebsd.org>
To:        Devin Teske <dteske@freebsd.org>
Cc:        Chris Rees <crees@bayofrum.net>, "freebsd-current@freebsd.org" <freebsd-current@freebsd.org>, Garrett Wollman <wollman@hergotha.csail.mit.edu>
Subject:   Re: [HEADSUP] No more pkg_install on HEAD by default
Message-ID:  <CAJ-VmomBya3arWpbDPm4zxXaRvykkPUudrN_DxisMBCn75B58w@mail.gmail.com>
In-Reply-To: <13CA24D6AB415D428143D44749F57D7201FC59E8@ltcfiswmsgmb21>
References:  <20130712231637.GS85556@ithaqua.etoilebsd.net> <13CA24D6AB415D428143D44749F57D7201FC2DBD@ltcfiswmsgmb21> <20130713080732.GV85556@ithaqua.etoilebsd.net> <13CA24D6AB415D428143D44749F57D7201FC3AA2@ltcfiswmsgmb21> <13CA24D6AB415D428143D44749F57D7201FC3C92@ltcfiswmsgmb21> <CA%2B7WWSe7C2UXAsEX3OZH7nPJJbxotpQ_iYBxgake4R68sbUtLA@mail.gmail.com> <13CA24D6AB415D428143D44749F57D7201FC3FAA@ltcfiswmsgmb21> <CAGE5yCoH2auer_kKpUT_caFUZPpVM5TdAFH5tJcGgF4Ji12f0g@mail.gmail.com> <201307140613.r6E6Dsov002016@hergotha.csail.mit.edu> <201307140706.r6E76Kg0002959@hergotha.csail.mit.edu> <13CA24D6AB415D428143D44749F57D7201FC51FE@ltcfiswmsgmb21> <7325EE70-8821-4350-9D8A-E5CAAC548FE9@bayofrum.net> <13CA24D6AB415D428143D44749F57D7201FC59E8@ltcfiswmsgmb21>

next in thread | previous in thread | raw e-mail | index | archive | help
... I bet you could do that. I bet you could build the rpm inside a
linux jail and have the relevant uname bits overridden in the right
way.


-adrian

On 14 July 2013 09:52, Teske, Devin <Devin.Teske@fisglobal.com> wrote:
>
> On Jul 14, 2013, at 8:01 AM, Chris Rees wrote:
>
>> On 14 Jul 2013, at 08:29, Teske, Devin wrote:
>>>
>>> To give you an idea as to just how helpful this is...
>>>
>>> Imagine the following hierarchy:
>>>
>>> src/pkgbase/depend/mystuff/script1
>>> src/pkgbase/depend/mystuff/textfile1
>>> src/pkgbase/depend/mystuff/sourcefile.c
>>> src/pkgbase/depend/mystuff/Makefile
>>>
>>> You are a developer. You want to ship a package that contains "script1"=
, "textfile1", and "binary1" (which is compiled by saying "make" to turn "s=
ourcefile.c" into "binary1")
>>>
>>> You want to ship 8 types of packages:
>>>
>>> FreeBSD-4.11
>>> FreeBSD-8.1 (i386)
>>> FreeBSD-8.1 (amd64)
>>> RedHat EL 4
>>> RedHat EL 6 (i386)
>>> RedHat EL 6 (x86_64)
>>> Debian Wheezy
>>> Debian Wheezy 64-bit
>>>
>>> This is where my framework comes in-handy...
>>>
>>> cd ~/src/pkgbase/freebsd/RELENG_4/category/mystuff
>>> make
>>> # it pulled the necessary bits from "src/pkgbase/depend/mystuff" and bu=
ilt the .tgz
>>>
>>> cd ~/src/pkgbase/freebsd/RELENG_8/category/mystuff
>>> make
>>> # it pulled the necessary bits from the "depend" dir and built .tbz
>>>
>>> cd ~/src/pkgbase/redhat/rhel4/category/sub-category/mystuff
>>> make
>>> # pulled in "depend" and made .rpm
>>>
>>> cd ~/src/pkgbase/redhat/rhel6/category/sub-category/mystuff
>>> make
>>> # pulled in "depend" and made .rpm
>>>
>>> etc.
>>>
>>> Of course, *any* time the depend tree has binaries in it... you have to=
 first do a make in there on the platform you want to ship the binary for, =
and then do "make depend" in the platform-specific tree to pull in the bina=
ries. Once you've done that, you don't have to muck with the depend tree ag=
ain unless there are changes there.
>>>
>>> So, I assume that your prejudice remarks are because you haven't either=
 seen (a) such a platform or (b) such a need for said platform.
>>>
>>> Yeah, I could rewrite the freebsd-specific logic to use "pkg create", b=
ut let me tell you...
>>>
>>> When you have to touch a file that needs to get shipped out to multiple=
 platforms...
>>>
>>> It's damned nice to be able to build the FreeBSD packages under RedHat =
*BECAUSE* the redhat RPMs can't be built under anything else (building an R=
PM on FreeBSD and attempting to install it on RedHat results in an error me=
ssage similar to "this is an rpm for FreeBSD; go away").
>>>
>>> Whereas FreeBSD will never balk about a package built on another platfo=
rm.
>>>
>>> It's a huge time-saving measure... not having to jump over to each/ever=
y unique platform to package things up *IF/WHEN* you know that there are no=
 binaries in the package *or* you've already checked the pre-compiled binar=
ies into the arch-specific hierarchy.
>>>
>>>
>>>
>>>
>>>> Or you
>>>> can maintain the old cruft for your business -- just don't expect
>>>> anyone else to use it, or even want to.
>>>>
>>>
>>>
>>> I have no intention of making old-world packages... but I also have no =
intention of using "pkg create".
>>
>> You still haven't really explained at all why you can't use libpkg.  If =
it doesn't run on Debian (not tried), it's got to be easier to port it than=
 rewrite a hacked version, hasn't it???  At least then you'll also be contr=
ibuting back.
>>
>
> Simple, really.
>
> Let's take RPM for example. The RPM package format has been ported to oth=
er platforms.
>
> But, I can't take archivers/rpm4 and build on RPM on FreeBSD and install =
it on RedHat.
>
> This is because the RPM format records the platform that you "build" your=
 RPM on (not the binaries, just the RPM) *into* said RPM.
>
> This actually adds a requirement to the RPM production that the RPMs be p=
roduced on the platform that they will be installed-to.
>
> Currently, no such restriction exists for the building of FreeBSD package=
s (within our system). This would have been true if we had ported pkg_creat=
e (and may continue to be true if we ported pkg and its ilk), but let's say=
 for the sake of argument that the future of "pkg" looks bright and it gets=
 ported to all sorts of systems (ported in a fashion similar to RPM) *and* =
we find one day that the +MANIFEST starts containing a target-platform (res=
ulting in refusal to install a *.txz package because it was rolled on a dif=
ferent platform.
>
> In that case, we'd then prefer to by-pass the tools and use our own metho=
d of creating the tar-ball to lift such a restriction.
>
> ASIDE: If I knew how to force rpmbuild into creating androgynous packages=
 for other architectures, I'd be doing that to life the restriction there t=
oo, but I haven't figured out that.
>
> Basically... within our "pkgbase" tree, we like the branch within the tre=
e to dictate how a package is built... not what platform you're on. The goa=
l being that we can run a single package-build host that builds all of our =
packages from a single platform.
> --
> Devin
>
> _____________
> The information contained in this message is proprietary and/or confident=
ial. If you are not the intended recipient, please: (i) delete the message =
and all copies; (ii) do not disclose, distribute or use the message in any =
manner; and (iii) notify the sender immediately. In addition, please be awa=
re that any message addressed to our domain is subject to archiving and rev=
iew by persons other than the intended recipient. Thank you.
> _______________________________________________
> freebsd-current@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org=
"



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-VmomBya3arWpbDPm4zxXaRvykkPUudrN_DxisMBCn75B58w>