[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[Yann E. MORIN]

This reverts commit 758ea0e518492c703482a323653c3d20520b599b.

758ea0e bumped gettext-tiny to the latest (at the time) head of the
upstream git tree. In so doing, it included an upstream change of
behaviour that eventually broke the build of packages that use
-Werror=format-string.

Even though in Buildroot we usually try to avoid building with -Werror,
it is sometimes hard to go against such flags, especially when the
package uses an expicit, long list of such -Werror=foo flags.

Revert the bump.

Fixes:
    http://autobuild.buildroot.org/results/dec/decbbcad11fbf312f6e5fbe0bf32143bedb6610b/
    [and many more...]

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Giulio Benetti <giulio.benetti@micronovasrl.com>
Cc: Vadim Kochan <vadim4j@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
---
 package/gettext-tiny/gettext-tiny.hash | 2 +-
 package/gettext-tiny/gettext-tiny.mk   | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/package/gettext-tiny/gettext-tiny.hash b/package/gettext-tiny/gettext-tiny.hash
index 7d13054144..672e964186 100644
--- a/package/gettext-tiny/gettext-tiny.hash
+++ b/package/gettext-tiny/gettext-tiny.hash
@@ -1,5 +1,5 @@
 # Locally Computed:
-sha256 00b03b08d909021aa5580fb45b6cf0ec45e9084ef94c26f034f03a64a69942bc  gettext-tiny-c6dcdcdef801127549d3906d153c061880d25a73.tar.gz
+sha256 654dcd52f2650476c8822b60bee89c20a0aa7f6a1bf6001701eeacd71a9e388b  gettext-tiny-v0.3.1.tar.gz
 sha256 b57aa4fdc1c614c28d41c1e2d5c4090935964c5f86291ba7d1c99ffd1d698b34  LICENSE
 sha256 8ceb4b9ee5adedde47b31e975c1d90c73ad27b6b165a1dcd80c7c545eb65b903  extra/COPYING
 # From http://lists.gnu.org/archive/html/bug-gettext/2016-06/msg00008.html
diff --git a/package/gettext-tiny/gettext-tiny.mk b/package/gettext-tiny/gettext-tiny.mk
index e3a6d11399..881edc6f35 100644
--- a/package/gettext-tiny/gettext-tiny.mk
+++ b/package/gettext-tiny/gettext-tiny.mk
@@ -4,7 +4,7 @@
 #
 ################################################################################
 
-GETTEXT_TINY_VERSION = c6dcdcdef801127549d3906d153c061880d25a73
+GETTEXT_TINY_VERSION = v0.3.1
 GETTEXT_TINY_SITE = $(call github,sabotage-linux,gettext-tiny,$(GETTEXT_TINY_VERSION))
 GETTEXT_TINY_LICENSE = MIT, GPL-3.0+ (extra gettext)
 GETTEXT_TINY_INSTALL_STAGING = YES
-- 
2.14.1

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[Vadim Kochan]

Hi Yann, All

On Sun, Apr 07, 2019 at 09:37:08AM +0200, Yann E. MORIN wrote:
> This reverts commit 758ea0e518492c703482a323653c3d20520b599b.
> 
> 758ea0e bumped gettext-tiny to the latest (at the time) head of the
> upstream git tree. In so doing, it included an upstream change of
> behaviour that eventually broke the build of packages that use
> -Werror=format-string.
> 
> Even though in Buildroot we usually try to avoid building with -Werror,
> it is sometimes hard to go against such flags, especially when the
> package uses an expicit, long list of such -Werror=foo flags.
> 
> Revert the bump.
> 
> Fixes:
>     http://autobuild.buildroot.org/results/dec/decbbcad11fbf312f6e5fbe0bf32143bedb6610b/
>     [and many more...]
> 
> Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
> Cc: Giulio Benetti <giulio.benetti@micronovasrl.com>
> Cc: Vadim Kochan <vadim4j@gmail.com>
> Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
> Cc: Peter Korsgaard <peter@korsgaard.com>
> ---
>  package/gettext-tiny/gettext-tiny.hash | 2 +-
>  package/gettext-tiny/gettext-tiny.mk   | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/package/gettext-tiny/gettext-tiny.hash b/package/gettext-tiny/gettext-tiny.hash
> index 7d13054144..672e964186 100644
> --- a/package/gettext-tiny/gettext-tiny.hash
> +++ b/package/gettext-tiny/gettext-tiny.hash
> @@ -1,5 +1,5 @@
>  # Locally Computed:
> -sha256 00b03b08d909021aa5580fb45b6cf0ec45e9084ef94c26f034f03a64a69942bc  gettext-tiny-c6dcdcdef801127549d3906d153c061880d25a73.tar.gz
> +sha256 654dcd52f2650476c8822b60bee89c20a0aa7f6a1bf6001701eeacd71a9e388b  gettext-tiny-v0.3.1.tar.gz
>  sha256 b57aa4fdc1c614c28d41c1e2d5c4090935964c5f86291ba7d1c99ffd1d698b34  LICENSE
>  sha256 8ceb4b9ee5adedde47b31e975c1d90c73ad27b6b165a1dcd80c7c545eb65b903  extra/COPYING
>  # From http://lists.gnu.org/archive/html/bug-gettext/2016-06/msg00008.html
> diff --git a/package/gettext-tiny/gettext-tiny.mk b/package/gettext-tiny/gettext-tiny.mk
> index e3a6d11399..881edc6f35 100644
> --- a/package/gettext-tiny/gettext-tiny.mk
> +++ b/package/gettext-tiny/gettext-tiny.mk
> @@ -4,7 +4,7 @@
>  #
>  ################################################################################
>  
> -GETTEXT_TINY_VERSION = c6dcdcdef801127549d3906d153c061880d25a73
> +GETTEXT_TINY_VERSION = v0.3.1
>  GETTEXT_TINY_SITE = $(call github,sabotage-linux,gettext-tiny,$(GETTEXT_TINY_VERSION))
>  GETTEXT_TINY_LICENSE = MIT, GPL-3.0+ (extra gettext)
>  GETTEXT_TINY_INSTALL_STAGING = YES

I am testing a fix with passing LIBINTL=NONE in gettext-tiny.mk which
does not install libintl.h at all, I think this should be ok as each
kind of toolchain should provide gettext stubs already.

Regards,
Vadim Kochan

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[James Hilliard]

On Sun, Apr 7, 2019 at 9:37 AM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
>
> This reverts commit 758ea0e518492c703482a323653c3d20520b599b.
>
> 758ea0e bumped gettext-tiny to the latest (at the time) head of the
> upstream git tree. In so doing, it included an upstream change of
> behaviour that eventually broke the build of packages that use
> -Werror=format-string.
I'm assuming the glib2 build error was caused by this change.
I have a potential fix for that here:
https://gitlab.gnome.org/GNOME/glib/merge_requests/770/diffs
>
> Even though in Buildroot we usually try to avoid building with -Werror,
> it is sometimes hard to go against such flags, especially when the
> package uses an expicit, long list of such -Werror=foo flags.
>
> Revert the bump.
>
> Fixes:
>     http://autobuild.buildroot.org/results/dec/decbbcad11fbf312f6e5fbe0bf32143bedb6610b/
>     [and many more...]
>
> Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
> Cc: Giulio Benetti <giulio.benetti@micronovasrl.com>
> Cc: Vadim Kochan <vadim4j@gmail.com>
> Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
> Cc: Peter Korsgaard <peter@korsgaard.com>
> ---
>  package/gettext-tiny/gettext-tiny.hash | 2 +-
>  package/gettext-tiny/gettext-tiny.mk   | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/package/gettext-tiny/gettext-tiny.hash b/package/gettext-tiny/gettext-tiny.hash
> index 7d13054144..672e964186 100644
> --- a/package/gettext-tiny/gettext-tiny.hash
> +++ b/package/gettext-tiny/gettext-tiny.hash
> @@ -1,5 +1,5 @@
>  # Locally Computed:
> -sha256 00b03b08d909021aa5580fb45b6cf0ec45e9084ef94c26f034f03a64a69942bc  gettext-tiny-c6dcdcdef801127549d3906d153c061880d25a73.tar.gz
> +sha256 654dcd52f2650476c8822b60bee89c20a0aa7f6a1bf6001701eeacd71a9e388b  gettext-tiny-v0.3.1.tar.gz
>  sha256 b57aa4fdc1c614c28d41c1e2d5c4090935964c5f86291ba7d1c99ffd1d698b34  LICENSE
>  sha256 8ceb4b9ee5adedde47b31e975c1d90c73ad27b6b165a1dcd80c7c545eb65b903  extra/COPYING
>  # From http://lists.gnu.org/archive/html/bug-gettext/2016-06/msg00008.html
> diff --git a/package/gettext-tiny/gettext-tiny.mk b/package/gettext-tiny/gettext-tiny.mk
> index e3a6d11399..881edc6f35 100644
> --- a/package/gettext-tiny/gettext-tiny.mk
> +++ b/package/gettext-tiny/gettext-tiny.mk
> @@ -4,7 +4,7 @@
>  #
>  ################################################################################
>
> -GETTEXT_TINY_VERSION = c6dcdcdef801127549d3906d153c061880d25a73
> +GETTEXT_TINY_VERSION = v0.3.1
>  GETTEXT_TINY_SITE = $(call github,sabotage-linux,gettext-tiny,$(GETTEXT_TINY_VERSION))
>  GETTEXT_TINY_LICENSE = MIT, GPL-3.0+ (extra gettext)
>  GETTEXT_TINY_INSTALL_STAGING = YES
> --
> 2.14.1
>
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[Yann E. MORIN]

James, All,

On 2019-04-07 10:23 +0200, James Hilliard spake thusly:
> On Sun, Apr 7, 2019 at 9:37 AM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> >
> > This reverts commit 758ea0e518492c703482a323653c3d20520b599b.
> >
> > 758ea0e bumped gettext-tiny to the latest (at the time) head of the
> > upstream git tree. In so doing, it included an upstream change of
> > behaviour that eventually broke the build of packages that use
> > -Werror=format-string.
> I'm assuming the glib2 build error was caused by this change.

Yes, see the build failure I pointed to below...

> I have a potential fix for that here:
> https://gitlab.gnome.org/GNOME/glib/merge_requests/770/diffs

I've ahd a quick look, but since I don;t know the glib2 code, I have no
idea if that is the correct solution.

However, you're probably going to have to provide more explanations
about your change to upstream, because libglib2 does build with the
non-tiny  implementations (either from glibc or from GNU gettext).

Regards,
Yann E. MORIN.

> >
> > Even though in Buildroot we usually try to avoid building with -Werror,
> > it is sometimes hard to go against such flags, especially when the
> > package uses an expicit, long list of such -Werror=foo flags.
> >
> > Revert the bump.
> >
> > Fixes:
> >     http://autobuild.buildroot.org/results/dec/decbbcad11fbf312f6e5fbe0bf32143bedb6610b/
> >     [and many more...]
> >
> > Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
> > Cc: Giulio Benetti <giulio.benetti@micronovasrl.com>
> > Cc: Vadim Kochan <vadim4j@gmail.com>
> > Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
> > Cc: Peter Korsgaard <peter@korsgaard.com>
> > ---
> >  package/gettext-tiny/gettext-tiny.hash | 2 +-
> >  package/gettext-tiny/gettext-tiny.mk   | 2 +-
> >  2 files changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/package/gettext-tiny/gettext-tiny.hash b/package/gettext-tiny/gettext-tiny.hash
> > index 7d13054144..672e964186 100644
> > --- a/package/gettext-tiny/gettext-tiny.hash
> > +++ b/package/gettext-tiny/gettext-tiny.hash
> > @@ -1,5 +1,5 @@
> >  # Locally Computed:
> > -sha256 00b03b08d909021aa5580fb45b6cf0ec45e9084ef94c26f034f03a64a69942bc  gettext-tiny-c6dcdcdef801127549d3906d153c061880d25a73.tar.gz
> > +sha256 654dcd52f2650476c8822b60bee89c20a0aa7f6a1bf6001701eeacd71a9e388b  gettext-tiny-v0.3.1.tar.gz
> >  sha256 b57aa4fdc1c614c28d41c1e2d5c4090935964c5f86291ba7d1c99ffd1d698b34  LICENSE
> >  sha256 8ceb4b9ee5adedde47b31e975c1d90c73ad27b6b165a1dcd80c7c545eb65b903  extra/COPYING
> >  # From http://lists.gnu.org/archive/html/bug-gettext/2016-06/msg00008.html
> > diff --git a/package/gettext-tiny/gettext-tiny.mk b/package/gettext-tiny/gettext-tiny.mk
> > index e3a6d11399..881edc6f35 100644
> > --- a/package/gettext-tiny/gettext-tiny.mk
> > +++ b/package/gettext-tiny/gettext-tiny.mk
> > @@ -4,7 +4,7 @@
> >  #
> >  ################################################################################
> >
> > -GETTEXT_TINY_VERSION = c6dcdcdef801127549d3906d153c061880d25a73
> > +GETTEXT_TINY_VERSION = v0.3.1
> >  GETTEXT_TINY_SITE = $(call github,sabotage-linux,gettext-tiny,$(GETTEXT_TINY_VERSION))
> >  GETTEXT_TINY_LICENSE = MIT, GPL-3.0+ (extra gettext)
> >  GETTEXT_TINY_INSTALL_STAGING = YES
> > --
> > 2.14.1
> >
> > _______________________________________________
> > buildroot mailing list
> > buildroot at busybox.net
> > http://lists.busybox.net/mailman/listinfo/buildroot

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[James Hilliard]

On Sun, Apr 7, 2019 at 10:30 AM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
>
> James, All,
>
> On 2019-04-07 10:23 +0200, James Hilliard spake thusly:
> > On Sun, Apr 7, 2019 at 9:37 AM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> > >
> > > This reverts commit 758ea0e518492c703482a323653c3d20520b599b.
> > >
> > > 758ea0e bumped gettext-tiny to the latest (at the time) head of the
> > > upstream git tree. In so doing, it included an upstream change of
> > > behaviour that eventually broke the build of packages that use
> > > -Werror=format-string.
> > I'm assuming the glib2 build error was caused by this change.
>
> Yes, see the build failure I pointed to below...
Ah, I was a little confused exactly how -Werror=format-nonliteral was related to
-Werror=format-string.
>
> > I have a potential fix for that here:
> > https://gitlab.gnome.org/GNOME/glib/merge_requests/770/diffs
>
> I've ahd a quick look, but since I don;t know the glib2 code, I have no
> idea if that is the correct solution.
>
> However, you're probably going to have to provide more explanations
> about your change to upstream, because libglib2 does build with the
> non-tiny  implementations (either from glibc or from GNU gettext).
Oh, I wasn't aware this was limited to a specific getttext implementation.
I'll mention that in the pull request.
>
> Regards,
> Yann E. MORIN.
>
> > >
> > > Even though in Buildroot we usually try to avoid building with -Werror,
> > > it is sometimes hard to go against such flags, especially when the
> > > package uses an expicit, long list of such -Werror=foo flags.
> > >
> > > Revert the bump.
> > >
> > > Fixes:
> > >     http://autobuild.buildroot.org/results/dec/decbbcad11fbf312f6e5fbe0bf32143bedb6610b/
> > >     [and many more...]
> > >
> > > Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
> > > Cc: Giulio Benetti <giulio.benetti@micronovasrl.com>
> > > Cc: Vadim Kochan <vadim4j@gmail.com>
> > > Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
> > > Cc: Peter Korsgaard <peter@korsgaard.com>
> > > ---
> > >  package/gettext-tiny/gettext-tiny.hash | 2 +-
> > >  package/gettext-tiny/gettext-tiny.mk   | 2 +-
> > >  2 files changed, 2 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/package/gettext-tiny/gettext-tiny.hash b/package/gettext-tiny/gettext-tiny.hash
> > > index 7d13054144..672e964186 100644
> > > --- a/package/gettext-tiny/gettext-tiny.hash
> > > +++ b/package/gettext-tiny/gettext-tiny.hash
> > > @@ -1,5 +1,5 @@
> > >  # Locally Computed:
> > > -sha256 00b03b08d909021aa5580fb45b6cf0ec45e9084ef94c26f034f03a64a69942bc  gettext-tiny-c6dcdcdef801127549d3906d153c061880d25a73.tar.gz
> > > +sha256 654dcd52f2650476c8822b60bee89c20a0aa7f6a1bf6001701eeacd71a9e388b  gettext-tiny-v0.3.1.tar.gz
> > >  sha256 b57aa4fdc1c614c28d41c1e2d5c4090935964c5f86291ba7d1c99ffd1d698b34  LICENSE
> > >  sha256 8ceb4b9ee5adedde47b31e975c1d90c73ad27b6b165a1dcd80c7c545eb65b903  extra/COPYING
> > >  # From http://lists.gnu.org/archive/html/bug-gettext/2016-06/msg00008.html
> > > diff --git a/package/gettext-tiny/gettext-tiny.mk b/package/gettext-tiny/gettext-tiny.mk
> > > index e3a6d11399..881edc6f35 100644
> > > --- a/package/gettext-tiny/gettext-tiny.mk
> > > +++ b/package/gettext-tiny/gettext-tiny.mk
> > > @@ -4,7 +4,7 @@
> > >  #
> > >  ################################################################################
> > >
> > > -GETTEXT_TINY_VERSION = c6dcdcdef801127549d3906d153c061880d25a73
> > > +GETTEXT_TINY_VERSION = v0.3.1
> > >  GETTEXT_TINY_SITE = $(call github,sabotage-linux,gettext-tiny,$(GETTEXT_TINY_VERSION))
> > >  GETTEXT_TINY_LICENSE = MIT, GPL-3.0+ (extra gettext)
> > >  GETTEXT_TINY_INSTALL_STAGING = YES
> > > --
> > > 2.14.1
> > >
> > > _______________________________________________
> > > buildroot mailing list
> > > buildroot at busybox.net
> > > http://lists.busybox.net/mailman/listinfo/buildroot
>
> --
> .-----------------.--------------------.------------------.--------------------.
> |  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
> | +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
> | +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
> | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
> '------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[Yann E. MORIN]

James, All,

On 2019-04-07 10:45 +0200, James Hilliard spake thusly:
> On Sun, Apr 7, 2019 at 10:30 AM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> > On 2019-04-07 10:23 +0200, James Hilliard spake thusly:
> > > On Sun, Apr 7, 2019 at 9:37 AM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> > > >
> > > > This reverts commit 758ea0e518492c703482a323653c3d20520b599b.
> > > >
> > > > 758ea0e bumped gettext-tiny to the latest (at the time) head of the
> > > > upstream git tree. In so doing, it included an upstream change of
> > > > behaviour that eventually broke the build of packages that use
> > > > -Werror=format-string.
> > > I'm assuming the glib2 build error was caused by this change.
> >
> > Yes, see the build failure I pointed to below...
> Ah, I was a little confused exactly how -Werror=format-nonliteral was related to
> -Werror=format-string.

That's me again confused with the other patch from Vadim that was
applied to gettext-tiny...

> > > I have a potential fix for that here:
> > > https://gitlab.gnome.org/GNOME/glib/merge_requests/770/diffs
> >
> > I've ahd a quick look, but since I don;t know the glib2 code, I have no
> > idea if that is the correct solution.
> >
> > However, you're probably going to have to provide more explanations
> > about your change to upstream, because libglib2 does build with the
> > non-tiny  implementations (either from glibc or from GNU gettext).
> Oh, I wasn't aware this was limited to a specific getttext implementation.
> I'll mention that in the pull request.

Be sure to check and test locally before hand! ;-)

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[Vadim Kochan]

Hi Yann, James, All

On Sun, Apr 07, 2019 at 10:49:26AM +0200, Yann E. MORIN wrote:
> James, All,
> 
> On 2019-04-07 10:45 +0200, James Hilliard spake thusly:
> > On Sun, Apr 7, 2019 at 10:30 AM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> > > On 2019-04-07 10:23 +0200, James Hilliard spake thusly:
> > > > On Sun, Apr 7, 2019 at 9:37 AM Yann E. MORIN <yann.morin.1998@free.fr> wrote:
> > > > >
> > > > > This reverts commit 758ea0e518492c703482a323653c3d20520b599b.
> > > > >
> > > > > 758ea0e bumped gettext-tiny to the latest (at the time) head of the
> > > > > upstream git tree. In so doing, it included an upstream change of
> > > > > behaviour that eventually broke the build of packages that use
> > > > > -Werror=format-string.
> > > > I'm assuming the glib2 build error was caused by this change.
> > >
> > > Yes, see the build failure I pointed to below...
> > Ah, I was a little confused exactly how -Werror=format-nonliteral was related to
> > -Werror=format-string.
> 
> That's me again confused with the other patch from Vadim that was
> applied to gettext-tiny...
> 
> > > > I have a potential fix for that here:
> > > > https://gitlab.gnome.org/GNOME/glib/merge_requests/770/diffs
> > >
> > > I've ahd a quick look, but since I don;t know the glib2 code, I have no
> > > idea if that is the correct solution.
> > >
> > > However, you're probably going to have to provide more explanations
> > > about your change to upstream, because libglib2 does build with the
> > > non-tiny  implementations (either from glibc or from GNU gettext).
> > Oh, I wasn't aware this was limited to a specific getttext implementation.
> > I'll mention that in the pull request.
> 
> Be sure to check and test locally before hand! ;-)
> 

I sent yet-another-patch in:

	https://patchwork.ozlabs.org/patch/1079937/

but do you think it is better to fix issue per-package ?

Regards,
Vadim Kochan

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[Thomas Petazzoni]

Hello,

On Sun,  7 Apr 2019 09:37:08 +0200
"Yann E. MORIN" <yann.morin.1998@free.fr> wrote:

> This reverts commit 758ea0e518492c703482a323653c3d20520b599b.
> 
> 758ea0e bumped gettext-tiny to the latest (at the time) head of the
> upstream git tree. In so doing, it included an upstream change of
> behaviour that eventually broke the build of packages that use
> -Werror=format-string.
> 
> Even though in Buildroot we usually try to avoid building with -Werror,
> it is sometimes hard to go against such flags, especially when the
> package uses an expicit, long list of such -Werror=foo flags.
> 
> Revert the bump.
> 
> Fixes:
>     http://autobuild.buildroot.org/results/dec/decbbcad11fbf312f6e5fbe0bf32143bedb6610b/
>     [and many more...]
> 
> Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
> Cc: Giulio Benetti <giulio.benetti@micronovasrl.com>
> Cc: Vadim Kochan <vadim4j@gmail.com>
> Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
> Cc: Peter Korsgaard <peter@korsgaard.com>
> ---
>  package/gettext-tiny/gettext-tiny.hash | 2 +-
>  package/gettext-tiny/gettext-tiny.mk   | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)

I applied Vadim's patch instead, in order to give a chance to solve
both the rygel issue and the libglib2 issue. Let's see how it goes.

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[Yann E. MORIN]

Thomas, All,

On 2019-04-07 15:11 +0200, Thomas Petazzoni spake thusly:
> On Sun,  7 Apr 2019 09:37:08 +0200
> "Yann E. MORIN" <yann.morin.1998@free.fr> wrote:
> > This reverts commit 758ea0e518492c703482a323653c3d20520b599b.
[--SNIP--]
> I applied Vadim's patch instead, in order to give a chance to solve
> both the rygel issue and the libglib2 issue. Let's see how it goes.

I'm fine with that, too.

But I would like to emphasise that, so far, we did not manage to have a
gettext-tiny that *really* impersonates GNU gettext.

I think that, ideally, we would like to have:
  - libintl headers that expose the same API as those from GNU gettext,
  - a libintl (probably a static one) that exposes just stub functions.

Additionally, for pre-built blobs, we might want to have a shared
libintl as well, that exposes the same stub functions (but I would not
miss such a lib.)

As I see it, gettext-tiny does not readily and easily provides for this.

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH] Revert "package/gettext-tiny: bump version to add --template option"

[Thomas Petazzoni]

On Sun, 7 Apr 2019 15:55:43 +0200
"Yann E. MORIN" <yann.morin.1998@free.fr> wrote:

> I think that, ideally, we would like to have:
>   - libintl headers that expose the same API as those from GNU gettext,
>   - a libintl (probably a static one) that exposes just stub functions.

Is this actually needed ? All of glibc, musl and uclibc provide
libintl.h, and a dummy implementation of libintl functions.

We only need the full-blown gettext on the target to provide libintl
with real functionality when NLS=y and uclibc/musl is used.

But other than that, when !NLS, the stub implementation in musl/uclibc
is fine, no ?

Thomas
-- 
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com