GNU bug report logs - #73529
[PATCH core] gnu: patchelf: Build from git sources.

Previous Next

Full log

Message #8 received at 73529 <at> debbugs.gnu.org (full text, mbox):

From: Z572 <zhengjunjie <at> iscas.ac.cn>
To: attila.lendvai <at> gmail.com
Cc: Attila Lendvai <attila <at> lendvai.name>, 73529 <at> debbugs.gnu.org
Subject: Re: [bug#73529] [PATCH core] gnu: patchelf: Build from git sources.
Date: Mon, 30 Sep 2024 00:12:27 +0800

[Message part 1 (text/plain, inline)]

attila.lendvai <at> gmail.com writes:

> From: Attila Lendvai <attila <at> lendvai.name>
>
> Rationale: it makes it much easier to build a pending PR by using
> `(commit "pull/544/head")` in the origin, or in general to experiment
> with different patchelf versions.  It also makes it more resilient
> against attacks hidden in release tarballs.
>
> Change-Id: I94eaedcd826c32b2c90823fe026718213f8d5962
> ---
>
> sadly, this leads to endless rebuilds...
>
>  gnu/packages/elf.scm | 27 +++++++++++++++------------
>  1 file changed, 15 insertions(+), 12 deletions(-)
>
> diff --git a/gnu/packages/elf.scm b/gnu/packages/elf.scm
> index dc2db836321..21228a9d878 100644
> --- a/gnu/packages/elf.scm
> +++ b/gnu/packages/elf.scm
> @@ -278,19 +278,19 @@ (define-public patchelf
>      (name "patchelf")
>      (version "0.18.0")
>      (source (origin
> -             (method url-fetch)
> -             (uri (string-append
> -                   "https://github.com/NixOS/patchelf/releases/download/"
> -                   version
> -                   "/patchelf-" version ".tar.bz2"))
> -             (sha256
> -              (base32
> -               "02s7ap86rx6yagfh9xwp96sgsj0p6hp99vhiq9wn4mxshakv4lhr"))))
> +              (method git-fetch)
> +              (uri (git-reference
> +                    (url "https://github.com/NixOS/patchelf")
> +                    (commit version)))
> +              (file-name (git-file-name "patchelf" version))

i think should "(git-file-name name version)" ?

> +              (sha256
> +               (base32
> +                "1v5mh8wdax4a97hw79nvzy0dixkbs0h487xkskkljkkdamq4rj5z"))))
>      (build-system gnu-build-system)
>      (arguments
>       '(#:phases
>         (modify-phases %standard-phases
> -         (add-after 'unpack 'fix-tests
> +         (add-after 'bootstrap 'fix-tests

Just a question, why do you need to change to bootstrap here?

>             ;; Our GCC code ensures that RUNPATH is never empty, it includes
>             ;; at least glibc/lib and gcc:lib/lib.
>             (lambda* (#:key inputs #:allow-other-keys)
> @@ -298,8 +298,8 @@ (define-public patchelf
>                 ;; Disable checking for an empty runpath:
>                 (("^if test.*") "")
>                 ;; Find libgcc_s.so, which is necessary for the test:
> -               (("/xxxxxxxxxxxxxxx") (string-append (assoc-ref inputs "gcc:lib")
> -                                                    "/lib")))
> +               (("/xxxxxxxxxxxxxxx")
> +                (dirname (search-input-file inputs "/lib/libgcc_s.so"))))
>               (substitute* "tests/replace-needed.sh"
>                 ;; This test assumes that only libc will be linked alongside
>                 ;; libfoo, but we also link libgcc_s.
> @@ -312,7 +312,10 @@ (define-public patchelf
>               (substitute* "tests/Makefile.in"
>                 ((".*shared-rpath\\.sh \\.*") "")))))))
>      (native-inputs
> -     `(("gcc:lib" ,gcc "lib")))
> +     (list
> +      autoconf
> +      automake
> +      (list gcc "lib")))

It would be nice to mention removing labels and adding new dependencies
in the submission, and etc/committer.scm maybe useful.

>      (home-page "https://nixos.org/patchelf.html")
>      (synopsis "Modify the dynamic linker and RPATH of ELF executables")
>      (description
>
> base-commit: 4f86fa20179ded1e6314eeba7da17309d501a32f

[signature.asc (application/pgp-signature, inline)]

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.