Re: Cross-compilation broken on canonical packages.
Hey Ludo, > Do we need to do it for ‘lower-object’? I think that one is (almost) > always called from a gexp compiler where it’s explicitly passed ‘system’ > and ‘target’. Yes because of lower-object call in "system-derivation" procedure of (gnu services). > Also, a test or two would be welcome as it’s easy to break it without > noticing. > > Otherwise LGTM, thank you, and sorry for the delay! Pushed with a few tests on core-updates, thanks for reviewing! Mathieu
Re: Cross-compilation broken on canonical packages.
Hi! Mathieu Othacehe skribis: > I think you're right! Pushed it as > f30d84d32db0f4f6cb84e139868e1727a7dc0a51 on core-updates. > > Now the good news is that we are one patch away from having core-updates > to cross-compile bare-bones.tmpl system! This is the patch attached, > we've been discussing it here[1] without reaching a conclusion. Yay! > From f5d773dc0a627ba6059f1025189d33a9e0d083b5 Mon Sep 17 00:00:00 2001 > From: Mathieu Othacehe > Date: Sat, 28 Dec 2019 21:29:06 +0100 > Subject: [PATCH 1/2] gexp: Default to current target. > > * guix/gexp.scm (lower-object): Set target argument to 'current by default and > look for the current target system at bind time if needed, > (gexp->file): ditto, > (gexp->script): ditto. Do we need to do it for ‘lower-object’? I think that one is (almost) always called from a gexp compiler where it’s explicitly passed ‘system’ and ‘target’. Also, a test or two would be welcome as it’s easy to break it without noticing. Otherwise LGTM, thank you, and sorry for the delay! Ludo’.
Re: Cross-compilation broken on canonical packages.
Hello, > The only downside will be potentially an extra glibc download/build to > build the locale set, but that’s probably OK. > > WDYT? I think you're right! Pushed it as f30d84d32db0f4f6cb84e139868e1727a7dc0a51 on core-updates. Now the good news is that we are one patch away from having core-updates to cross-compile bare-bones.tmpl system! This is the patch attached, we've been discussing it here[1] without reaching a conclusion. WDYT? Thanks, Mathieu [1]: https://lists.gnu.org/archive/html/guix-patches/2019-12/msg00751.html >From f5d773dc0a627ba6059f1025189d33a9e0d083b5 Mon Sep 17 00:00:00 2001 From: Mathieu Othacehe Date: Sat, 28 Dec 2019 21:29:06 +0100 Subject: [PATCH 1/2] gexp: Default to current target. * guix/gexp.scm (lower-object): Set target argument to 'current by default and look for the current target system at bind time if needed, (gexp->file): ditto, (gexp->script): ditto. --- guix/gexp.scm | 87 ++- 1 file changed, 51 insertions(+), 36 deletions(-) diff --git a/guix/gexp.scm b/guix/gexp.scm index 12331052a6..11d4e86037 100644 --- a/guix/gexp.scm +++ b/guix/gexp.scm @@ -215,7 +215,7 @@ procedure to expand it; otherwise return #f." (define* (lower-object obj #:optional (system (%current-system)) - #:key target) + #:key (target 'current)) "Return as a value in %STORE-MONAD the derivation or store item corresponding to OBJ for SYSTEM, cross-compiling for TARGET if TARGET is true. OBJ must be an object that has an associated gexp compiler, such as a @@ -225,7 +225,10 @@ OBJ must be an object that has an associated gexp compiler, such as a (raise (condition (&gexp-input-error (input obj) (lower ;; Cache in STORE the result of lowering OBJ. - (mlet %store-monad ((graft? (grafting?))) + (mlet %store-monad ((target (if (eq? target 'current) + (current-target-system) + (return target))) + (graft? (grafting?))) (mcached (let ((lower (lookup-compiler obj))) (lower obj system target)) obj @@ -1571,16 +1574,19 @@ are searched for in PATH. Return #f when MODULES and EXTENSIONS are empty." #:key (guile (default-guile)) (module-path %load-path) (system (%current-system)) - target) + (target 'current)) "Return an executable script NAME that runs EXP using GUILE, with EXP's imported modules in its search path. Look up EXP's modules in MODULE-PATH." - (mlet %store-monad ((set-load-path - (load-path-expression (gexp-modules exp) - module-path - #:extensions - (gexp-extensions exp) - #:system system - #:target target))) + (mlet* %store-monad ((target (if (eq? target 'current) + (current-target-system) + (return target))) + (set-load-path +(load-path-expression (gexp-modules exp) + module-path + #:extensions + (gexp-extensions exp) + #:system system + #:target target))) (gexp->derivation name (gexp (call-with-output-file (ungexp output) @@ -1609,7 +1615,7 @@ imported modules in its search path. Look up EXP's modules in MODULE-PATH." (module-path %load-path) (splice? #f) (system (%current-system)) - target) + (target 'current)) "Return a derivation that builds a file NAME containing EXP. When SPLICE? is true, EXP is considered to be a list of expressions that will be spliced in the resulting file. @@ -1620,36 +1626,45 @@ Lookup EXP's modules in MODULE-PATH." (define modules (gexp-modules exp)) (define extensions (gexp-extensions exp)) - (if (or (not set-load-path?) - (and (null? modules) (null? extensions))) - (gexp->derivation name -(gexp - (call-with-output-file (ungexp output) - (lambda (port) - (for-each (lambda (exp) - (write exp port)) - '(ungexp (if splice? -exp -
Re: Cross-compilation broken on canonical packages.
Hi, Mathieu Othacehe skribis: >>> But I'd like to avoid that. I'll see if I can figure out something >>> better after a good night. >> >> Couldn't find much better, here's a patch, tell my what you think! > > I pushed a patch removing all canonical-packages calls, except the one > for the default-locales-libcs variable. > > About this variable, I don't know if the dirty patch I proposed is > better than accepting that the glibc used for locales is different from > the one programs are linked against. WDYT? On second thought, I think removing ‘canonical-package’ calls from (gnu system locale) should work. What matters there is that we build locales with the matching libc (same version, not necessarily same store item.) The only downside will be potentially an extra glibc download/build to build the locale set, but that’s probably OK. WDYT? Thanks, Ludo’.
Re: Cross-compilation broken on canonical packages.
Hey Ludo, >> But I'd like to avoid that. I'll see if I can figure out something >> better after a good night. > > Couldn't find much better, here's a patch, tell my what you think! I pushed a patch removing all canonical-packages calls, except the one for the default-locales-libcs variable. About this variable, I don't know if the dirty patch I proposed is better than accepting that the glibc used for locales is different from the one programs are linked against. WDYT? Thanks, Mathieu
Re: Cross-compilation broken on canonical packages.
Hello, > We could use a thunk field to write something like: > > --8<---cut here---start->8--- > (define (default-locale-libcs ...) > (if target glibc (canonical-package glibc)) > --8<---cut here---end--->8--- > > But I'd like to avoid that. I'll see if I can figure out something > better after a good night. Couldn't find much better, here's a patch, tell my what you think! Thanks, Mathieu >From e1728d6c79f63ad72bcfb4e45111ccb9f43f1251 Mon Sep 17 00:00:00 2001 From: Mathieu Othacehe Date: Fri, 3 Jan 2020 09:23:51 +0100 Subject: [PATCH] defaut locale libc. --- gnu/system.scm| 3 ++- gnu/system/locale.scm | 10 +- 2 files changed, 7 insertions(+), 6 deletions(-) diff --git a/gnu/system.scm b/gnu/system.scm index 0842bd47c7..faff0320d3 100644 --- a/gnu/system.scm +++ b/gnu/system.scm @@ -211,7 +211,8 @@ (locale-definitions operating-system-locale-definitions ; list of (default %default-locale-definitions)) (locale-libcs operating-system-locale-libcs ; list of -(default %default-locale-libcs)) +(thunked) +(default (default-locale-libcs))) (name-service-switch operating-system-name-service-switch ; (default %default-nss)) diff --git a/gnu/system/locale.scm b/gnu/system/locale.scm index c7a371e9bf..361c024eeb 100644 --- a/gnu/system/locale.scm +++ b/gnu/system/locale.scm @@ -37,7 +37,7 @@ locale-name->definition locale-directory -%default-locale-libcs +default-locale-libcs %default-locale-definitions glibc-supported-locales)) @@ -123,7 +123,7 @@ of LIBC." (computed-file (string-append "locale-" version) build)) (define* (locale-directory locales - #:key (libcs %default-locale-libcs)) + #:key (libcs (default-locale-libcs))) "Return a locale directory containing all of LOCALES for each libc package listed in LIBCS. @@ -145,9 +145,9 @@ data format changes between libc versions." #:options '(#:local-build? #t #:substitutable? #f)) -(define %default-locale-libcs - ;; The libcs for which we build locales by default. - (list glibc)) +(define* (default-locale-libcs + #:key (target (%current-target-system))) + (if target (list glibc) (list (canonical-package glibc (define %default-locale-definitions ;; Arbitrary set of locales that are built by default. They are here mostly -- 2.24.1
Re: Cross-compilation broken on canonical packages.
Hi, > It’s more than I thought but I think it’s OK. (Also, how come > bare-bones takes 1.5 GiB?!) That's one of my next subject of investigation :) >> (define %default-locale-libcs >>;; The libcs for which we build locales by default. >> - (list (canonical-package glibc))) >> + (list glibc)) > > I would leave this bit unchanged if possible, because that way the glibc > used would really match that programs are linked against. operating-system-locale-libcs defaults to %default-locale-libcs, so keeping (canonical-package glibc) prevents cross-compilation :( We could use a thunk field to write something like: --8<---cut here---start->8--- (define (default-locale-libcs ...) (if target glibc (canonical-package glibc)) --8<---cut here---end--->8--- But I'd like to avoid that. I'll see if I can figure out something better after a good night. Thanks, Mathieu
Re: Cross-compilation broken on canonical packages.
Hola! Mathieu Othacehe skribis: >> Two simple solutions here, I think: >> >> 1. Make ‘packages’ a thunked field. >> >> 2. Stop using ‘canonical-package’ altogether in ‘%base-packages’. >> >> I actually have a preference for #2. We’d need to check what impact it >> has on the system closure size, but I suspect it’s going to be minimal. >> >> Thoughts? > > Option #2 seems nicer. There are other canonical-package calls when > building services. I have a patch attached that removes all those calls. > > The image produced when running "guix system disk-image bare-bones.tmpl" > increases by: > > * 30MB when removing the canonical-package of %base-packages > * 30MB for all other calls. OK. (It may be slightly more accurate to look at the output of: guix size $(guix system build …).) > So altogether, that's a 60MB increase on a 1.5GB image (4%). I find it > acceptable, WDYT? It’s more than I thought but I think it’s OK. (Also, how come bare-bones takes 1.5 GiB?!) > From a55eb86975c5f1c2783b0af265364821259e3351 Mon Sep 17 00:00:00 2001 > From: Mathieu Othacehe > Date: Tue, 31 Dec 2019 09:56:51 +0100 > Subject: [PATCH] system: Stop using canonical-package. > > Usage of canonical-package outside of thunked fields breaks cross-compilation, > see: https://lists.gnu.org/archive/html/guix-devel/2019-12/msg00410.html. > > * gnu/installer.scm (installer-program): Remove canonical-package. > * gnu/services/base.scm (): Ditto, > (%base-services): ditto. > * gnu/services/xorg.scm: Remove useless canonical-package import. > * gnu/system.scm (%base-packages): Remove canonical-package. > * gnu/system/install.scm (%installation-services): Ditto, > (installation-os): ditto. > * gnu/system/locale.scm (single-locale-directory): Ditto, > (%default-locale-libcs): ditto. [...] > --- a/gnu/system/locale.scm > +++ b/gnu/system/locale.scm > @@ -86,7 +86,7 @@ or #f on failure." > #f))) > > (define* (single-locale-directory locales > - #:key (libc (canonical-package glibc))) > + #:key (libc glibc)) >"Return a directory containing all of LOCALES for LIBC compiled. > > Because locale data formats are incompatible when switching from one libc to > @@ -147,7 +147,7 @@ data format changes between libc versions." > > (define %default-locale-libcs >;; The libcs for which we build locales by default. > - (list (canonical-package glibc))) > + (list glibc)) I would leave this bit unchanged if possible, because that way the glibc used would really match that programs are linked against. The rest LGTM, thanks! Ludo’.
Re: Cross-compilation broken on canonical packages.
Hola,
> However, the hack doesn’t work for things like the ‘packages’ field of
> because it’s not a thunked field.
I see! Thanks for explaining.
>
> Two simple solutions here, I think:
>
> 1. Make ‘packages’ a thunked field.
>
> 2. Stop using ‘canonical-package’ altogether in ‘%base-packages’.
>
> I actually have a preference for #2. We’d need to check what impact it
> has on the system closure size, but I suspect it’s going to be minimal.
>
> Thoughts?
Option #2 seems nicer. There are other canonical-package calls when
building services. I have a patch attached that removes all those calls.
The image produced when running "guix system disk-image bare-bones.tmpl"
increases by:
* 30MB when removing the canonical-package of %base-packages
* 30MB for all other calls.
So altogether, that's a 60MB increase on a 1.5GB image (4%). I find it
acceptable, WDYT?
Thanks,
Mathieu
>From a55eb86975c5f1c2783b0af265364821259e3351 Mon Sep 17 00:00:00 2001
From: Mathieu Othacehe
Date: Tue, 31 Dec 2019 09:56:51 +0100
Subject: [PATCH] system: Stop using canonical-package.
Usage of canonical-package outside of thunked fields breaks cross-compilation,
see: https://lists.gnu.org/archive/html/guix-devel/2019-12/msg00410.html.
* gnu/installer.scm (installer-program): Remove canonical-package.
* gnu/services/base.scm (): Ditto,
(%base-services): ditto.
* gnu/services/xorg.scm: Remove useless canonical-package import.
* gnu/system.scm (%base-packages): Remove canonical-package.
* gnu/system/install.scm (%installation-services): Ditto,
(installation-os): ditto.
* gnu/system/locale.scm (single-locale-directory): Ditto,
(%default-locale-libcs): ditto.
---
gnu/installer.scm | 22 +++---
gnu/services/base.scm | 10 --
gnu/services/xorg.scm | 2 +-
gnu/system.scm | 11 ++-
gnu/system/install.scm | 7 +++
gnu/system/locale.scm | 4 ++--
6 files changed, 23 insertions(+), 33 deletions(-)
diff --git a/gnu/installer.scm b/gnu/installer.scm
index 167653263f..7f12b39ffc 100644
--- a/gnu/installer.scm
+++ b/gnu/installer.scm
@@ -290,17 +290,17 @@ selected keymap."
(define set-installer-path
;; Add the specified binary to PATH for later use by the installer.
#~(let* ((inputs
- '#$(append (list bash ;start subshells
- connman ;call connmanctl
- cryptsetup
- dosfstools ;mkfs.fat
- e2fsprogs ;mkfs.ext4
- btrfs-progs
- kbd ;chvt
- guix ;guix system init call
- util-linux ;mkwap
- shadow)
- (map canonical-package (list coreutils)
+ '#$(list bash ;start subshells
+ connman ;call connmanctl
+ cryptsetup
+ dosfstools ;mkfs.fat
+ e2fsprogs ;mkfs.ext4
+ btrfs-progs
+ kbd ;chvt
+ guix ;guix system init call
+ util-linux ;mkwap
+ shadow
+ coreutils)))
(with-output-to-port (%make-void-port "w")
(lambda ()
(set-path-environment-variable "PATH" '("bin" "sbin") inputs)
diff --git a/gnu/services/base.scm b/gnu/services/base.scm
index b1eff89ecc..b7cde09611 100644
--- a/gnu/services/base.scm
+++ b/gnu/services/base.scm
@@ -44,7 +44,7 @@
#:select (alsa-utils crda eudev e2fsprogs fuse gpm kbd lvm2 rng-tools))
#:use-module (gnu packages bash)
#:use-module ((gnu packages base)
-#:select (canonical-package coreutils glibc glibc-utf8-locales))
+#:select (coreutils glibc glibc-utf8-locales))
#:use-module (gnu packages package-management)
#:use-module ((gnu packages gnupg) #:select (guile-gcrypt))
#:use-module (gnu packages linux)
@@ -1208,7 +1208,7 @@ the tty to run, among other things."
(name-services nscd-configuration-name-services ;list of
(default '()))
(glibc nscd-configuration-glibc;
- (default (canonical-package glibc
+ (default glibc)))
(define-record-type* nscd-cache make-nscd-cache
nscd-cache?
@@ -2441,9 +2441,7 @@ to handle."
(rules (list lvm2 fuse alsa-utils crda
(service special-files-service-type
- `(("/bin/sh" ,(file-append (canonical-package bash)
-"/bin/sh"))
- ("/usr/bin/env" ,(file-append (canonical-package coreutils)
- "/bin/env"))
+ `(("/bin/sh" ,(file-append bash "/bin/sh"))
+ ("/usr/bin/env" ,(file-append coreutils "/bin/env"))
;;; base.scm ends her
Re: Cross-compilation broken on canonical packages.
Hi, Mathieu Othacehe skribis: >> This is expected: packages in ‘%final-inputs’ (those returned by >> ‘canonical-package’) are rooted in the bootstrap graph and cannot be >> cross-compiled. > > Looking at canonical-package in (gnu packages commencement), I see that > there's already a switch on (%current-target-system). The given package > is directly returned if (%current-target-system) is set, which appears > to be what we want! > > ;; In general we want CANON, except if we're cross-compiling: CANON > ;; uses explicit inputs, so it is "anchored" in the bootstrapped > ;; process, with dependencies on things that cannot be > ;; cross-compiled. > (if (%current-target-system) > package > canon)) > > > But, this doesn't work as expected. I guess it is because of > (%current-target-system) evaluation time. Ah yeah. The hack works when you use ‘canonical-package’ in a package input, because package input fields are “thunked”, so they’re evaluated in a dynamic context where ‘%current-target-system’ has the right value. However, the hack doesn’t work for things like the ‘packages’ field of because it’s not a thunked field. Two simple solutions here, I think: 1. Make ‘packages’ a thunked field. 2. Stop using ‘canonical-package’ altogether in ‘%base-packages’. I actually have a preference for #2. We’d need to check what impact it has on the system closure size, but I suspect it’s going to be minimal. Thoughts? (The trick you posted should work of course, but it’s even better if we can avoid it!) Thanks, Ludo’.
Re: Cross-compilation broken on canonical packages.
Hello Ludo,
Thanks for your explanation :)
> This is expected: packages in ‘%final-inputs’ (those returned by
> ‘canonical-package’) are rooted in the bootstrap graph and cannot be
> cross-compiled.
Looking at canonical-package in (gnu packages commencement), I see that
there's already a switch on (%current-target-system). The given package
is directly returned if (%current-target-system) is set, which appears
to be what we want!
--8<---cut here---start->8---
;; In general we want CANON, except if we're cross-compiling: CANON
;; uses explicit inputs, so it is "anchored" in the bootstrapped
;; process, with dependencies on things that cannot be
;; cross-compiled.
(if (%current-target-system)
package
canon))
--8<---cut here---end--->8---
But, this doesn't work as expected. I guess it is because of
(%current-target-system) evaluation time.
As I'm not fully understand everything here, I would propose to define a
gexp-compiler for "canonical-packages", but there's maybe a better thing
to do?
Anyway, the snippet below works fine with a gexp-compiler (patch
attached) and doesn't work with the current implementation of
canonical-package.
--8<---cut here---start->8---
(use-modules (guix)
(gnu packages base))
(run-with-store (open-connection)
(mlet* %store-monad
((gexp-> #~(#$(canonical-package grep)))
(drv (gexp->script "test.scm" gexp
#:target "aarch64-linux-gnu"))
(build (built-derivations (list drv
(return #t)))
--8<---cut here---end--->8---
WDYT?
Thanks,
Mathieu
>From 6266c46181d2880684c8519423be8d8a8ea3 Mon Sep 17 00:00:00 2001
From: Mathieu Othacehe
Date: Sun, 22 Dec 2019 17:25:59 +0100
Subject: [PATCH] wip: Add a canonical-package gexp compiler.
---
gnu/packages/base.scm | 6 ++
gnu/packages/commencement.scm | 40 +++
2 files changed, 46 insertions(+)
diff --git a/gnu/packages/base.scm b/gnu/packages/base.scm
index e8150708c0..22b6e05dae 100644
--- a/gnu/packages/base.scm
+++ b/gnu/packages/base.scm
@@ -1337,6 +1337,12 @@ package needs iconv ,@(libiconv-if-needed) should be added."
(proc (module-ref iface 'canonical-package)))
(proc package)))
+(define-public (canonical-package* package)
+ ;; Avoid circular dependency by lazily resolving 'commencement'.
+ (let* ((iface (resolve-interface '(gnu packages commencement)))
+ (proc (module-ref iface 'make-canonical-package)))
+(proc package)))
+
(define-public (%final-inputs)
"Return the list of \"final inputs\"."
;; Avoid circular dependency by lazily resolving 'commencement'.
diff --git a/gnu/packages/commencement.scm b/gnu/packages/commencement.scm
index 14ecf246d4..41ab430388 100644
--- a/gnu/packages/commencement.scm
+++ b/gnu/packages/commencement.scm
@@ -54,8 +54,11 @@
#:use-module (guix build-system gnu)
#:use-module (guix build-system trivial)
#:use-module (guix memoization)
+ #:use-module (guix gexp)
+ #:use-module (guix records)
#:use-module (guix utils)
#:use-module (srfi srfi-1)
+ #:use-module (srfi srfi-9)
#:use-module (srfi srfi-26)
#:use-module (ice-9 vlist)
#:use-module (ice-9 match)
@@ -2613,5 +2616,42 @@ Fortran development to be installed in user profiles. This includes
gfortran, as well as libc (headers and binaries, plus debugging symbols
in the @code{debug} output), and binutils.")))
+(define-record-type
+ (%make-canonical-package name)
+ canonical-package?
+ (name canonical-package-name))
+
+(define-public (make-canonical-package package)
+ (%make-canonical-package package))
+
+(define-gexp-compiler (canonical-package-compiler
+ (canonical-package ) system target)
+ ;; Return the 'canonical' variant of PACKAGE---i.e., if PACKAGE is one of
+ ;; the implicit inputs of 'gnu-build-system', return that one, otherwise
+ ;; return PACKAGE.
+ ;;
+ ;; The goal is to avoid duplication in cases like GUILE-FINAL vs. GUILE-2.2,
+ ;; COREUTILS-FINAL vs. COREUTILS, etc.
+ (let ((name->package (fold (lambda (input result)
+ (match input
+ ((_ package . outputs)
+ (vhash-cons (package-full-name package)
+ package result
+ vlist-null
+ `(("guile" ,guile-final)
+ ,@%final-inputs)))
+(package (canonical-package-name canonical-package)))
+;; XXX: This doesn't handle dependencies of the final inputs, such as
+;; libunistring, GMP, etc.
+(match (vhash-assoc (package-full-name package) name->package)
+ ((_ . canon)
+ ;; In general we want CANON,
Re: Cross-compilation broken on canonical packages.
Hi Mathieu, Mathieu Othacehe skribis: > Small mistake sorry. This fails: > > guix build --target=aarch64-linux-gnu -e "((@ (gnu packages base) > canonical-package) (@ (gnu packages base) grep))" > > > while this succeeds: > > guix build -e "((@ (gnu packages base) canonical-package) (@ (gnu packages > base) grep))" This is expected: packages in ‘%final-inputs’ (those returned by ‘canonical-package’) are rooted in the bootstrap graph and cannot be cross-compiled. Perhaps ‘%base-packages’ should use packages form (gnu packages base) instead? The reason it uses ‘canonical-package’ is just to avoid introducing redundant packages in the system, but I don’t think it makes that much sense. Thanks, Ludo’.
Re: Cross-compilation broken on canonical packages.
Small mistake sorry. This fails: --8<---cut here---start->8--- guix build --target=aarch64-linux-gnu -e "((@ (gnu packages base) canonical-package) (@ (gnu packages base) grep))" --8<---cut here---end--->8--- while this succeeds: --8<---cut here---start->8--- guix build -e "((@ (gnu packages base) canonical-package) (@ (gnu packages base) grep))" --8<---cut here---end--->8--- Mathieu
Cross-compilation broken on canonical packages.
Hello, This command fails: --8<---cut here---start->8--- guix build -e "((@ (gnu packages base) canonical-package) (@ (gnu packages base) grep))" --8<---cut here---end--->8--- with this output: --8<---cut here---start->8--- building /gnu/store/nlq0iw749hnyy7nk3cq6pjwqkxv437ca-make-boot0-4.2.1.drv... starting phase `set-SOURCE-DATE-EPOCH' phase `set-SOURCE-DATE-EPOCH' succeeded after 0.0 seconds starting phase `set-paths' environment variable `PATH' set to `/gnu/store/36zgy21hfy185nmdgj87ll6q4hj5riiw-gcc-cross-aarch64-linux-gnu-7.4.0/bin:/gnu/store/4d68mhjjwqc95lyrm63yrsnrlqzkbpyr-binutils-cross-aarch64-linux-gnu-2.32/bin' environment variable `CROSS_LIBRARY_PATH' set to `/gnu/store/pgg0nxh1l39dybb5pwqyskb9qxqz57r9-glibc-mesboot-2.16.0/lib:/gnu/store/p0s4wym402gsnqzicr87zrli43dvfbms-binutils-mesboot-2.20.1a/lib:/gnu/store/bb6s2y6121pxvhwyr1n6dg7pwj88bsbj-gcc-mesboot-4.9.4/lib:/gnu/store/p5r7gprzg774vq8l94924m598i86ryya-glibc-cross-aarch64-linux-gnu-2.29/lib:/gnu/store/z86g7zwj2mdl1328qpgmim262sppjxqp-glibc-cross-aarch64-linux-gnu-2.29-static/lib' environment variable `CROSS_CPATH' set to `/gnu/store/pgg0nxh1l39dybb5pwqyskb9qxqz57r9-glibc-mesboot-2.16.0/include:/gnu/store/p0s4wym402gsnqzicr87zrli43dvfbms-binutils-mesboot-2.20.1a/include:/gnu/store/bb6s2y6121pxvhwyr1n6dg7pwj88bsbj-gcc-mesboot-4.9.4/include:/gnu/store/bp1ap2whzmwvh2myj31hfmg9pi4sly0r-linux-libre-headers-bootstrap-0/include:/gnu/store/p5r7gprzg774vq8l94924m598i86ryya-glibc-cross-aarch64-linux-gnu-2.29/include:/gnu/store/f3izsng5fzwh6zf2z7aaj5v04jp5hmcy-linux-libre-headers-cross-aarch64-linux-gnu-4.19.56/include' environment variable `GUIX_LOCPATH' unset environment variable `C_INCLUDE_PATH' set to `/gnu/store/36zgy21hfy185nmdgj87ll6q4hj5riiw-gcc-cross-aarch64-linux-gnu-7.4.0/include' environment variable `CPLUS_INCLUDE_PATH' set to `/gnu/store/36zgy21hfy185nmdgj87ll6q4hj5riiw-gcc-cross-aarch64-linux-gnu-7.4.0/include' environment variable `LIBRARY_PATH' set to `/gnu/store/36zgy21hfy185nmdgj87ll6q4hj5riiw-gcc-cross-aarch64-linux-gnu-7.4.0/lib' environment variable `C_INCLUDE_PATH' set to `/gnu/store/36zgy21hfy185nmdgj87ll6q4hj5riiw-gcc-cross-aarch64-linux-gnu-7.4.0/include' phase `set-paths' succeeded after 0.0 seconds starting phase `install-locale' warning: failed to install 'en_US.utf8' locale: Invalid argument phase `install-locale' succeeded after 0.0 seconds starting phase `unpack' In execvp of tar: No such file or directory command "tar" "xvf" "/gnu/store/cr04i9xi5nbkn7lqb1nnxkqvpwy92m78-make-4.2.1.tar.xz" failed with status 127 builder for `/gnu/store/nlq0iw749hnyy7nk3cq6pjwqkxv437ca-make-boot0-4.2.1.drv' failed with exit code 1 build of /gnu/store/nlq0iw749hnyy7nk3cq6pjwqkxv437ca-make-boot0-4.2.1.drv failed View build log at '/var/log/guix/drvs/nl/q0iw749hnyy7nk3cq6pjwqkxv437ca-make-boot0-4.2.1.drv.gz'. cannot build derivation `/gnu/store/xpb7jhgjghxy624gz7rfy5zb8rsdri3i-grep-3.3.drv': 1 dependencies couldn't be built guix build: error: build of `/gnu/store/xpb7jhgjghxy624gz7rfy5zb8rsdri3i-grep-3.3.drv' failed --8<---cut here---end--->8--- This is topic has already been discussed here[1][2]. Usually I work around it by passing "--no-grafts" or avoiding canonical-packages. Now that cross compiling a system is possible, avoiding canonical packages is hard (used in %base-packages and "system" derivation). Any ideas where to start on this problem? Thanks, Mathieu [1]: https://guix-devel.gnu.narkive.com/LsByvXrK/cross-building-bootstrap-binaries-fail-in-current-master#post5 [2]: https://lists.gnu.org/archive/html/bug-guix/2017-10/msg00101.html
