Buildroot Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Korsgaard <peter@korsgaard.com>
To: buildroot@busybox.net
Subject: [Buildroot] [PATCH] gutenprint: fix host build when libusb found
Date: Sat, 16 Apr 2016 00:07:59 +0200	[thread overview]
Message-ID: <87potqfruo.fsf@dell.be.48ers.dk> (raw)
In-Reply-To: <20160415175035.GQ7454@waldemar-brodkorb.de> (Waldemar Brodkorb's message of "Fri, 15 Apr 2016 19:50:35 +0200")

>>>>> "Waldemar" == Waldemar Brodkorb <wbx@openadk.org> writes:

 > Hi Peter,
 > Peter Korsgaard wrote,

 >> >>>>> "Waldemar" == Waldemar Brodkorb <wbx@openadk.org> writes:
 >> 
 >> > Allow to disable USB support, otherwise host build breaks when
 >> > libusb is found as installation tries to install to /usr.
 >> > Fixes following autobuild error:
 >> > http://autobuild.buildroot.net/results/081b3be918ac1eaa8cfbc5919e00bc1ea267c1df/
 >> 
 >> Hmm, but why does that end up with /usr and not $(HOST_DIR)/usr? Is it
 >> picking up a libusb.pc from your host or is our host-libusb .pc file
 >> wrong?

 > It is picking up libusb.pc from the host and I don't think
 > gutenprint-print will build host libusb at all. As cups support
 > is disabled anyway for the host, I don't think installing any USB
 > backends for the hosttools is strictly necessary and might be better
 > avoided.

I had a closer look, and it actually isn't libusb.pc from the build host
(or directly anything to do with libusb).

The problem is really that the --without-cups option doesn't really
work. We already have a related workaround for this where we
ac_cv_path_CUPS_CONFIG='', but more is needed. When we pass
--without-cups the configure script hardcodes cups_prefix to "/usr".

host-gutenprint itself doesn't depend on host-libusb, but other packages
do so it might be built before host-gutenprint. If the configure script
detects libusb support (using pkg-config), then it will build a cups
backend using libusb (even though cups support is disabled) and intall
it into $cups_prefix/share/cups/usb, breaking the build :/

So either we fix the configure script to really disable the cups support
when configured with --without-cups, we add your patch as a workaround
or we simply add PKG_CONFIG=false to HOST_GUTENPRINT_CONF_ENV to make
sure the configure script doesn't find libusb. We can do that as
pkg-config is only used to check for libusb and gtk2, neither of which
we need for host-gutenprint.

-- 
Venlig hilsen,
Peter Korsgaard 

      reply	other threads:[~2016-04-15 22:07 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-04-02 18:47 [Buildroot] [PATCH] gutenprint: fix host build when libusb found Waldemar Brodkorb
2016-04-15  4:49 ` [Buildroot] " Baruch Siach
2016-04-15  9:31 ` [Buildroot] [PATCH] " Peter Korsgaard
2016-04-15 17:50   ` Waldemar Brodkorb
2016-04-15 22:07     ` Peter Korsgaard [this message]

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=87potqfruo.fsf@dell.be.48ers.dk \
    --to=peter@korsgaard.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