Buildroot Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Pascal Mazon <pascal.mazon@6wind.com>
To: buildroot@busybox.net
Subject: [Buildroot] [PATCH 1/1] ncurses: generate libtermcap
Date: Wed, 02 Dec 2015 11:39:32 +0100	[thread overview]
Message-ID: <565ECA64.3070503@6wind.com> (raw)
In-Reply-To: <20151130233049.3ad42bd1@free-electrons.com>

On 11/30/2015 11:30 PM, Thomas Petazzoni wrote:
 > Dear Pascal Mazon,
 >
 > On Fri, 24 Apr 2015 14:54:26 +0200, Pascal Mazon wrote:
 >> There is no libtermcap package in buildroot, but ncurses implements 
termcap
 >> natively. Furthermore, ncurses already provides the termcap.h header 
file.
 >>
 >> With this patch, we fix an issue encountered with some external 
toolchains
 >> that include a libtermcap.a (typically the GNU libtermcap version) 
in their
 >> sysroot folder.
 >> Bash, for instance, would be linking with this libtermcap while using
 >> headers from ncurses.
 >>
 >> In order to be consistent, let's make sure there is only the ncurses'
 >> termcap library. To that effect, we:
 >> - remove any libtermcap.* in the staging dir,
 >> - install a link to libncurses static and/or shared in staging and 
target
 >>    dirs.
 >>
 >> Signed-off-by: Pascal Mazon <pascal.mazon@6wind.com>
 >
 > We finally took a bit of time on IRC today to discuss your patch. From
 > our point of view, the fact that your toolchain provides libtermcap.a
 > is a bug of the toolchain. From Buildroot's point of view, a toolchain
 > should not provide in its sysroot anything but the C library and its
 > headers, and the kernel headers.
 >
 > If we start doing hacks in packages to cope with toolchain sysroot
 > already shipped with some libraries, it's going to be an endless fight
 > against those toolchains. You have the problem with libtermcap.a, but
 > the next person will have it with libz.a, and then lib<something>, and
 > again.
 >
 > Our suggestion is that you fix your toolchain either by asking your
 > toolchain vendor, or by doing a quick hack to remove libtermcap.a
 > before using the toolchain with Buildroot. If you add this toolchain as
 > a known Buildroot profile, you can even automate this and make it part
 > of the build process.
 >
 > Best regards,
 >
 > Thomas
 >

I understand buildroot's point of view.
Indeed, it does not look like the best solution to hack in packages for 
a problem from the toolchain.

I will try your suggestion.
Maybe we can modify copy_toolchain_sysroot in toolchain/helpers.mk to 
add a custom "--exclude" when rsync-ing (with values taken from an newly 
added option)?

Best regards,

-- 
Pascal Mazon
www.6wind.com

  reply	other threads:[~2015-12-02 10:39 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-04-24 12:54 [Buildroot] [PATCH 1/1] ncurses: generate libtermcap Pascal Mazon
2015-04-24 14:18 ` Thomas Petazzoni
2015-04-27  8:11   ` Pascal Mazon
2015-07-19  9:19 ` Thomas Petazzoni
2015-11-30 22:30 ` Thomas Petazzoni
2015-12-02 10:39   ` Pascal Mazon [this message]
2015-12-05 17:39     ` Yann E. MORIN
  -- strict thread matches above, loose matches on Subject: below --
2015-05-20  9:11 Pascal Mazon

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=565ECA64.3070503@6wind.com \
    --to=pascal.mazon@6wind.com \
    --cc=buildroot@busybox.net \
    /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