Buildroot Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Arnout Vandecappelle <arnout@mind.be>
To: "Yann E. MORIN" <yann.morin.1998@free.fr>, buildroot@buildroot.org
Cc: Will Newton <will.newton@gmail.com>,
	Johan Oudinet <johan.oudinet@gmail.com>
Subject: Re: [Buildroot] [PATCH] package/erlang: do not hard-code the Erlang Interface Version (EI_VSN)
Date: Tue, 14 Feb 2023 22:30:52 +0100	[thread overview]
Message-ID: <4f51a196-e922-1e9e-6fbb-e81b35dc9f73@mind.be> (raw)
In-Reply-To: <20230210215548.1050371-1-yann.morin.1998@free.fr>



On 10/02/2023 22:55, Yann E. MORIN wrote:
> The note above the erlang version instructs to refer to another note
> further down the file. However, even if it is not too difficult to find,
> it is still located a bit too far away, and the reference is not very
> explicit what note we should look at.
> 
> When we introduced that variable in 6c1d128844c5 (package/erlang: export
> EI_VSN so other packages can use it), the rationale for hard-coding it
> was "to avoid spawning a shell every time the variable is dereferenced".
> 
> However, that can get a bit confusing and hard to follow. Also, that in
> fact spawns a shell only once for each rebar-packages, so the overhead
> is far from being too high.
> 
> The EI_VSN is only used by rebar-package packages, is derefrenced from
> the rebar-infra and not the packages themselves, and is not needed by
> erlang itself (it knows its own EI_VSN), so we can de-hard-code it, and
> rely on build-time detection, by looking in the appropriate file.
> 
> We have two files where we could look:
>    - lib/erl_interface/vsn.mk in the erlang source tree, but it is not
>      installed,
>    - .../lib/erlang/releases/$(ERLANG_RELASE)/installed_application_versions
>      as installed by erlang.
> 
> We use the second one, as it is cleaner, for a package, to look into
> installed files, rather than to look in the source tree of another
> package.
> 
> Although both the host and target erlang are the same, we still look
> into the corresponding file to extract the version. This is so that it
> would be easier if in the future we ever manage to rely on a
> system-installed erlang that could have a EI_VSN different from the
> target one.
> 
> We can't re-use the variable ERLANG_EI_VSN, because it now needs to be
> $(call)-ed with a parameter. Hopefully, external packages that use it
> directly rather than through the rebar infra, are not legion...
> 
> Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
> Cc: Will Newton <will.newton@gmail.com>
> Cc: Johan Oudinet <johan.oudinet@gmail.com>
> ---
>   package/erlang/erlang.mk | 8 ++++----
>   package/pkg-rebar.mk     | 4 ++--
>   2 files changed, 6 insertions(+), 6 deletions(-)
> 
> diff --git a/package/erlang/erlang.mk b/package/erlang/erlang.mk
> index ac2c204a49..5d6b69057a 100644
> --- a/package/erlang/erlang.mk
> +++ b/package/erlang/erlang.mk
> @@ -4,8 +4,8 @@
>   #
>   ################################################################################
>   
> -# See note below when updating Erlang
>   ERLANG_VERSION = 22.3.4.22
> +ERLANG_RELEASE = $(firstword $(subst ., ,$(ERLANG_VERSION)))
>   ERLANG_SITE = \
>   	https://github.com/erlang/otp/releases/download/OTP-$(ERLANG_VERSION)
>   ERLANG_SOURCE = otp_src_$(ERLANG_VERSION).tar.gz
> @@ -36,9 +36,9 @@ ERLANG_PRE_CONFIGURE_HOOKS += ERLANG_RUN_AUTOCONF
>   HOST_ERLANG_DEPENDENCIES += host-autoconf
>   HOST_ERLANG_PRE_CONFIGURE_HOOKS += ERLANG_RUN_AUTOCONF
>   
> -# Whenever updating Erlang, this value should be updated as well, to the
> -# value of EI_VSN in the file lib/erl_interface/vsn.mk
> -ERLANG_EI_VSN = 3.13.2.2
> +# Return the EIV (Erlang Interface Version, EI_VSN)
> +# $(1): base directory, i.e. either $(HOST_DIR) or $(STAGING_DIR)/usr
> +erlang_ei_vsn = `sed -r -e '/^erl_interface-(.+)/!d; s//\1/' $(1)/lib/erlang/releases/$(ERLANG_RELEASE)/installed_application_versions`

  Since this is only used by rebar-package, I think it should move to pkg-rebar.mk.

>   
>   # The configure checks for these functions fail incorrectly
>   ERLANG_CONF_ENV = ac_cv_func_isnan=yes ac_cv_func_isinf=yes
> diff --git a/package/pkg-rebar.mk b/package/pkg-rebar.mk
> index e4e3f3bb6c..4b993dd970 100644
> --- a/package/pkg-rebar.mk
> +++ b/package/pkg-rebar.mk
> @@ -32,10 +32,10 @@ REBAR_TARGET_DEPS_DIR = $(STAGING_DIR)/usr/share/rebar/deps
>   #
>   REBAR_HOST_DEPS_ENV = \
>   	ERL_COMPILER_OPTIONS='{i, "$(REBAR_HOST_DEPS_DIR)"}' \
> -	ERL_EI_LIBDIR=$(HOST_DIR)/lib/erlang/lib/erl_interface-$(ERLANG_EI_VSN)/lib
> +	ERL_EI_LIBDIR=$(HOST_DIR)/lib/erlang/lib/erl_interface-$(call erlang_ei_vsn,$(HOST_DIR))/lib

  Since we support only one erlang version, we could just as well put

$(wildcard $(HOST_DIR)/lib/erlang/lib/erl_interface-*/lib)

or does that directory not necessarily exist yet after erlang has been installed?

  Regards,
  Arnout


>   REBAR_TARGET_DEPS_ENV = \
>   	ERL_COMPILER_OPTIONS='{i, "$(REBAR_TARGET_DEPS_DIR)"}' \
> -	ERL_EI_LIBDIR=$(STAGING_DIR)/usr/lib/erlang/lib/erl_interface-$(ERLANG_EI_VSN)/lib
> +	ERL_EI_LIBDIR=$(STAGING_DIR)/usr/lib/erlang/lib/erl_interface-$(call erlang_ei_vsn,$(STAGING_DIR)/usr)/lib
>   
>   ################################################################################
>   # Helper functions
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

  reply	other threads:[~2023-02-14 21:31 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-02-10 21:55 [Buildroot] [PATCH] package/erlang: do not hard-code the Erlang Interface Version (EI_VSN) Yann E. MORIN
2023-02-14 21:30 ` Arnout Vandecappelle [this message]
2023-02-15  6:29   ` Yann E. MORIN
2023-08-10 21:16 ` Thomas Petazzoni via buildroot

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4f51a196-e922-1e9e-6fbb-e81b35dc9f73@mind.be \
    --to=arnout@mind.be \
    --cc=buildroot@buildroot.org \
    --cc=johan.oudinet@gmail.com \
    --cc=will.newton@gmail.com \
    --cc=yann.morin.1998@free.fr \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox