[gentoo-dev] Empty DEPEND strings in virtuals

[gentoo-dev] Empty DEPEND strings in virtuals

[Petteri Räty]

[-- Attachment #1: Type: text/plain, Size: 1309 bytes --]

betelgeuse@pena /usr/portage/virtual $ grep 'DEPEND=""' -r . | wc -l
97
betelgeuse@pena /usr/portage/virtual $ find -name "*.ebuild" | wc -l
102

betelgeuse@pena /usr/portage/virtual $ find -name "*.ebuild" | xargs
grep -L 'DEPEND=""'  | xargs grep DEPEND
./pmake/pmake-0.ebuild:RDEPEND="!userland_BSD? ( sys-devel/pmake )"
./perl-Scalar-List-Utils/perl-Scalar-List-Utils-1.19.ebuild:RDEPEND="~perl-core/Scalar-List-Utils-${PV}"
./c++-tr1-type-traits/c++-tr1-type-traits-0.ebuild:DEPEND="|| (
>=sys-devel/gcc-4.1 dev-libs/boost )"
./c++-tr1-memory/c++-tr1-memory-0.ebuild:DEPEND="|| (
>=sys-devel/gcc-4.1 dev-libs/boost )"
./c++-tr1-functional/c++-tr1-functional-0.ebuild:DEPEND="|| (
>=sys-devel/gcc-4.1 dev-libs/boost )"

so only three virtual ebuilds have a DEPEND

So it seems most ebuilds for new style virtuals don't have DEPENDs.
Anyone have any idea on how these are supposed to work or are they
buggy? PMS currently says that RDEPEND doesn't have to be installed at
the time a package is emerged so as far as I understand it this means
that DEPEND="virtual/foobar" could mean that it could happen that
nothing providing the virtual is available. So should we be changing the
DEPEND atoms to DEPEND="${RDEPEND}" or is there something I am missing here?

Regards,
Petteri


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 252 bytes --]

Re: [gentoo-dev] Empty DEPEND strings in virtuals

[Zac Medico]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Petteri Räty wrote:
> so only three virtual ebuilds have a DEPEND

According to GLEP 37 [1], they should only define RDEPEND.  The
reason that only RDEPEND is needed is that a package that has a
virtual dependency has freedom to include the virtual atom in any of
DEPEND, RDEPEND, and PDEPEND as necessary.  Note that when portage's
$ROOT feature is used, DEPEND is installed to the build platform.
The virtual ebuild itself does not make this decision, but rather
the package that has a virtual dependency.

> So it seems most ebuilds for new style virtuals don't have DEPENDs.
> Anyone have any idea on how these are supposed to work or are they
> buggy? PMS currently says that RDEPEND doesn't have to be installed at
> the time a package is emerged so as far as I understand it this means
> that DEPEND="virtual/foobar" could mean that it could happen that
> nothing providing the virtual is available.

The bit about "RDEPEND doesn't have to be installed" is only for
solving circular RDEPEND during the installation process.  When a
package has RDEPEND that is not installed, it must be considered
unusable and therefore this state should only exist for a relatively
short period of time during the installation process.

> So should we be changing the
> DEPEND atoms to DEPEND="${RDEPEND}" or is there something I am missing here?

We should not, as I've explained above.

Zac

[1] http://www.gentoo.org/proj/en/glep/glep-0037.html
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.3 (GNU/Linux)

iD8DBQFGH/fs/ejvha5XGaMRAhZoAJwPH7W04pL8EfkJ4STM+79Drl0jcQCdGjGz
n4Ee8wfFp3RgpKy25ftjoXo=
=TBuW
-----END PGP SIGNATURE-----
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Empty DEPEND strings in virtuals

[Petteri Räty]

[-- Attachment #1: Type: text/plain, Size: 226 bytes --]

Zac Medico kirjoitti:
> 
> [1] http://www.gentoo.org/proj/en/glep/glep-0037.html

We should link this info to the devmanual. Opened a bug about this:
https://bugs.gentoo.org/show_bug.cgi?id=174530

Regards,
Petteri


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 252 bytes --]

Re: [gentoo-dev] Empty DEPEND strings in virtuals

[Marius Mauch]

On Fri, 13 Apr 2007 21:50:50 +0300
Petteri Räty <betelgeuse@gentoo.org> wrote:

> betelgeuse@pena /usr/portage/virtual $ grep 'DEPEND=""' -r . | wc -l
> 97
> betelgeuse@pena /usr/portage/virtual $ find -name "*.ebuild" | wc -l
> 102
> 
> betelgeuse@pena /usr/portage/virtual $ find -name "*.ebuild" | xargs
> grep -L 'DEPEND=""'  | xargs grep DEPEND
> ./pmake/pmake-0.ebuild:RDEPEND="!userland_BSD? ( sys-devel/pmake )"
> ./perl-Scalar-List-Utils/perl-Scalar-List-Utils-1.19.ebuild:RDEPEND="~perl-core/Scalar-List-Utils-${PV}"
> ./c++-tr1-type-traits/c++-tr1-type-traits-0.ebuild:DEPEND="|| (
> >=sys-devel/gcc-4.1 dev-libs/boost )"
> ./c++-tr1-memory/c++-tr1-memory-0.ebuild:DEPEND="|| (
> >=sys-devel/gcc-4.1 dev-libs/boost )"
> ./c++-tr1-functional/c++-tr1-functional-0.ebuild:DEPEND="|| (
> >=sys-devel/gcc-4.1 dev-libs/boost )"
> 
> so only three virtual ebuilds have a DEPEND
> 
> So it seems most ebuilds for new style virtuals don't have DEPENDs.
> Anyone have any idea on how these are supposed to work or are they
> buggy? PMS currently says that RDEPEND doesn't have to be installed at
> the time a package is emerged so as far as I understand it this means
> that DEPEND="virtual/foobar" could mean that it could happen that
> nothing providing the virtual is available. So should we be changing the
> DEPEND atoms to DEPEND="${RDEPEND}" or is there something I am missing here?

As soon as a dependency chain has a DEPEND element then all lower
RDEPEND elements get converted into DEPENDs as well. Guess that might
be confusing, so example would be:

- app/bla has DEPEND="virtual/foo"
- virtual/foo has RDEPEND="lib/bar"
- in the internal depgraph app/bla gets DEPEND="lib/bar" as a result of
the transitive dependency

At least that's just how it should behave conceptually.

Marius
--
gentoo-dev@gentoo.org mailing list

[gentoo-dev] Deps (was Re: Empty DEPEND strings in virtuals)

[Steve Long]

Petteri Räty wrote:
> We should link this info to the devmanual.

Yeah that was v. instructive. Since there's only 3 ebuilds left with the old
syntax, the obvious question is: is there anything else holding up impl of
the GLEP?

Also, would the preferred syntax be open to usage in eg recommended deps as
opposed to mandatory ones? I appreciate that it's not the same GLEP or
anything. I'm also curious about impl date of ranged deps and use deps,
without wanting to start a flamewar about the merits or lack thereof.



-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] Deps (was Re: Empty DEPEND strings in virtuals)

[Zac Medico]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Steve Long wrote:
> Petteri Räty wrote:
>> We should link this info to the devmanual.
> 
> Yeah that was v. instructive. Since there's only 3 ebuilds left with the old
> syntax, the obvious question is: is there anything else holding up impl of
> the GLEP?

package.prefer remains unimplemented, but I'd like to implement it
or something similar.

> Also, would the preferred syntax be open to usage in eg recommended deps as
> opposed to mandatory ones? I appreciate that it's not the same GLEP or
> anything.

I guess "recommended deps" can be represented by the existing USE
flag conditional syntax [1], can't they?

> I'm also curious about impl date of ranged deps and use deps,
> without wanting to start a flamewar about the merits or lack thereof.

I'm working on a new resolver design right now and I hope to have
something usable within 1 or 2 months (sooner if we're lucky).  The
new design should make it relatively easy to implement all of the
dependencies of bugs 144480 [2] and 155723 [3].

Zac

[1]
http://devmanual.gentoo.org/general-concepts/dependencies/index.html#use-conditional-dependencies
[2] http://bugs.gentoo.org/show_bug.cgi?id=144480
[3] http://bugs.gentoo.org/show_bug.cgi?id=155723
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.3 (GNU/Linux)

iD8DBQFGIIQv/ejvha5XGaMRAvQjAKDLZG+ko6QTK1Be64xp8W5hoElvgwCg3auv
DluAGZKZiOOEeQ8Ltw/A/8c=
=ygOk
-----END PGP SIGNATURE-----
-- 
gentoo-dev@gentoo.org mailing list