Directory prefixify part 1.
In addition, E/D and E/ROOT has trailing slashes. No need to write
an additional slash.
---
eclass/toolchain.eclass | 29 +++--
1 file changed, 15 insertions(+), 14 deletions(-)
diff --git a/eclass/toolchain.eclass
In Prefix, PATH should also be preserved, resulting in a mouthful of
`env -i PATH=${PATH} ROOT=${ROOT}`. The origin commit introducing
env -i was for "cleanup". Dropping env -i is cleaner.
Reference:
Hi,
This patch series is splitted from the previous one
https://archives.gentoo.org/gentoo-dev/message/8a7ac352cb047567309c70aaf7105305
Note that the splitting is not perfect when different kinds
of updates happen in adjacent lines.
Please review.
Benda
Benda Xu (7):
toolchain.eclass
Mike Gilbert writes:
> Regarding the PATH/gcc-config change, I'm thinking it would make more
> sense to drop the env -i command than to add to he list of special
> variables we pass through to it.
All the env -i has been introduced in this commit by vapier in 2005:
James Le Cuirot writes:
> Why? All the ebuilds using this eclass that I can find are at least
> EAPI 4.
Okay! I will remove this.
Benda
signature.asc
Description: PGP signature
Hi,
Please help me review the patch attached.
This patch has 5 main kinds of modifications to the eclass,
* Define ED and EROOT for EAPI 0, 1 and 2.
* Add ${EPREFIX} to ${PREFIX} and quote the variables.
* Strip ${EPREFIX} if used with a ebuild helper.
* call fix_libtool_files.sh by
gro...@gentoo.org writes:
> I hope gcj will remain in (which I use not very often but regularly). There are no real
> alternatived: pdfshuffle fails on many (otherwise normal) pdf files.
+1
I am a regular pdftk user.
Benda
Hi,
"Walter Dnes" writes:
> Are the gmp, mpc, mpfr, and isl libs included? According to the
> "Support libraries" section at https://gcc.gnu.org/wiki/InstallingGCC
They are managed by their own ebuilds.
>> Alternatively, after extracting the GCC source archive, simply
Rich Freeman writes:
> Well, that would be a different approach, but I imagine you could
> build for prefix-style install using something like Catalyst and
> cross-compile. I have no idea how much tweaking that would require.
> The main issue is that unless you use qemu you're
Hi Michael,
"M. J. Everitt" writes:
>> No SD slot on Nexus. We will stress the internal NAND flash with
>> millions of ebuilds and rsync :)
> Completely out of my depth here, but can you cross-compile and
Ah, I was half-joking. Considering the recent advancement of NAND
Kristian Fiskerstrand writes:
>>> Actually .. given my nosiness in -pr matters lately, anyone done a
>>> write-up for Planet.g.o or suchlike that we can post up to social media?
>>> From my own stand-point it might be something the wider community would
>>> be interested in, and
Francesco Riosa writes:
> Christmas is coming, I'll enjoy this
Haha, good point. Should've called it Halloween release ;)
Benda
Hey MJE,
"M. J. Everitt" writes:
> COOL ! I'll have to give this a try .. got an old(er) Samsung Galaxy
> S4 Active,
From our past experience[1], Samsung sometimes imposed noexec to /data by
kernel hacks. It'll be interesting to find out what it looks like on
the Galaxy
Hi Fellows,
This is an announcement of the latest Gentoo on Android stage3 tarball,
http://distfiles.gentoo.org/experimental/prefix/rap/rap-stage3-armv7a_hardfp-20161026.tar.xz
snapshot from the latest main Gentoo repository, featuring gcc-5.4,
glibc-2.23, and the latest prefix-standalone
Nick Vinson writes:
> arguably gcc should also excluded, under that definition, so the wiki
> might not be 100% correct
This is not true regarding libgcc* runtime libraries.
Benda
signature.asc
Description: PGP signature
Hi Patrice,
Patrice Clement writes:
> [...]
Very enjoyable write-up. I completely agree with you.
This will be an important reference. Please consider adding it into the
wiki after we reach a wider consensus on how to merge pull request on
github.
Benda
signature.asc
Hi Alex,
Do you have a mentor now?
Benda
Hi,
"Anthony G. Basile" writes:
> maybe once we get some names we should get a meeting of
> base-system together and coordinate our efforts.
I am interested in keeping the base packages working on Prefix,
especially bash and baselayout. Please count me in.
Benda
Lei Zhang writes:
> I'm just afraid gcc-config would become too complex that way, so I
> prefer a simpler approach: let eselect-compiler be version-agnostic.
> Then we can have clang-config to handle the versioning of clang,
> icc-config to handle icc, etc.
If we model
Hi anonymous reviewer,
R0b0t1 writes:
> What is it intended to solve?
To simplify ebuilds that need to call eprefixify.
> The current behavior seems to make more sense. Hiding defaults causes
> problems.
I am not sure what you mean by "Hiding defaults". It is documented,
Hi,
This is a patch to extend prefix.eclass.
1. a set of heuristics is added to eprefixify:
"s,([^[:alnum:]}])/(usr|etc|bin|sbin|var|opt)/,\1${EPREFIX}/\2/,g"
2. a function wrapper "fprefixity" to do inplace substitution in ${T}.
Please help review it.
Benda
--- prefix.eclass
Michael Orlitzky writes:
>> # Don't forget to make directory for sysfs
>> -[[ ! -d ${ROOT}sys ]] && kernel_is 2 6 && mkdir ${ROOT}sys
>> +[[ ! -d ${EROOT}sys ]] && kernel_is 2 6 && mkdir "${EROOT}"sys
>>
>
> One more =)
Nice catch!
Jonathan Callen
Michael Orlitzky <m...@gentoo.org> writes:
> On 06/12/2016 05:21 AM, Benda Xu wrote:
>> # let other packages install some of these headers
>> -rm -rf "${D}"/${ddir}/scsi #glibc/uclibc/etc...
>> +rm -rf "${ED}"
Zac Medico writes:
> However, it behaves differently when you have a variable with "a b" as
> its content.
>
> foo="a b"
> [[ -f ${foo} ]] && echo hi
Yeah, that's correct. My mistake.
--- kernel-2.eclass 2016-02-17 22:46:25.235543840 +0900
+++ kernel-2.eclass
Hi Ulrich,
Ulrich Mueller writes:
>> I have added EPREFIX logics for EAPI<3 and improved the trailing
>> slashes and quotes.
>
>> -[[ -f ${ROOT}/usr/include/linux/autoconf.h ]] \
>> +[[ -f "${EROOT}"usr/include/linux/autoconf.h ]] \
>
> Inside [[ ]]
Thanks Zac, Göktürk,
I have added EPREFIX logics for EAPI<3 and improved the trailing slashes
and quotes.
Benda
--- kernel-2.eclass 2016-02-17 22:46:25.235543840 +0900
+++ kernel-2.eclass 2016-06-12 11:48:19.843801138 +0900
@@ -105,6 +105,8 @@
HOMEPAGE="https://www.kernel.org/
Hi Fellows,
This is a trivial patch for kernel-2.eclass to support Prefix. Does it
look good to be commited?
Thanks,
Benda
Bug: 478436
--- kernel-2.eclass 2016-02-17 22:46:25.235543840 +0900
+++ kernel-2.eclass 2016-05-24 01:14:22.246809021 +0900
@@ -539,8 +522,8 @@
Hi Farid,
This is an excellent idea! It is very helpful for Gentoo Prefix/libc,
where we maintain a set of nss databases (passwd, group, shadow, etc.)
inside a directory prefix.
Farid BENAMROUCHE writes:
> Currently there is an old known limitation when using ROOT= option
Hi Anthony,
"Anthony G. Basile" writes:
> different direction: what about building with
> rpath=/usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/ and then making sure that
> portage respects that library file during any --depclean or
> @preserved-rebuild? i'm not sure how we'd
>
Hi,
In libtool.eclass[1], it is mentioned in the comments of elt_patch_dir()
that
# If an overlay has eclass overrides, but doesn't actually override the
# libtool.eclass, we'll have ECLASSDIR pointing to the active overlay's
# eclass/ dir, but libtool.eclass is still in the main Gentoo tree.
101 - 130 of 130 matches
Mail list logo