Skip to contentSkip to conversation

package/gnmupg2: build fixes

GIT PULL New Yann E. MORIN 3 participants 5 messages Dec 16, 2015

lore
Share & tools
lore
#0 lore raw

Hello All!

This series fixes two build issues with gnupg2:
- build failure with musl (missing #include),
- really not running tests in cross-compilation.

Regards,
Yann E. MORIN.

The following changes since commit 4cb010a5f1773d11c0dadcb3ac5ffa6e154fa113:

package/libhdhomerun: new package (2015-12-16 22:11:33 +0100)

are available in the git repository at:

git://git.busybox.net/~ymorin/git/buildroot yem/gnupg2

for you to fetch changes up to c7dd2e8bb5db5e326a7a4e7cb3611b8a0eea0e68:

package/gnupg2: really do not run tests in cross-compilation (2015-12-16 22:42:36 +0100)

----------------------------------------------------------------
Yann E. MORIN (2):
package/gnupg2: fix build with musl package/gnupg2: really do not run tests in cross-compilation

 .../0000-silence-git-when-not-in-git-tree.patch    | 26 +++++++++++++++
 package/gnupg2/0001-fix-pth-config-usage.patch     | 29 ++++++++--------
 package/gnupg2/0002-missing-include.patch          | 15 +++++++++
 package/gnupg2/0003-dont-run-tests.patch           | 39 ++++++++++++++++++++++
 package/gnupg2/gnupg2.mk                           |  4 +++
 5 files changed, 97 insertions(+), 16 deletions(-)
 create mode 100644 package/gnupg2/0000-silence-git-when-not-in-git-tree.patch
 create mode 100644 package/gnupg2/0002-missing-include.patch
 create mode 100644 package/gnupg2/0003-dont-run-tests.patch
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#1 lore raw

PATCH 1/2 package/gnupg2: fix build with musl

Add a missing include to get select().

Fixes:

    http://autobuild.buildroot.org/results/8b2/8b282f546dc36f4bf9c25721cdf922c7a7ba340b/
    http://autobuild.buildroot.org/results/dc1/dc192255b1fcae18db8b98e65309e5c1493bdbc1/
    http://autobuild.buildroot.org/results/4c0/4c0ab2b28500a28945da3f4df742d8c2e2f21cd0/
    ...
---
 package/gnupg2/0002-missing-include.patch | 15 +++++++++++++++
 1 file changed, 15 insertions(+)
 create mode 100644 package/gnupg2/0002-missing-include.patch
package/gnupg2/0002-missing-include.patch new file +15 -0
new file mode 100644
index 0000000..f20994c
@@ -0,0 +1,15 @@
+tools/watchgnupg: select() is from sys/select.h
+
+Signed-off-by: "Yann E. MORIN" <[email protected]>
+
+diff -durN gnupg2-2.0.29.orig/tools/watchgnupg.c gnupg2-2.0.29/tools/watchgnupg.c
+--- gnupg2-2.0.29.orig/tools/watchgnupg.c	2015-09-08 14:39:24.000000000 +0200
++++ gnupg2-2.0.29/tools/watchgnupg.c	2015-12-16 18:17:17.758057787 +0100
+@@ -32,6 +32,7 @@
+ #include <sys/un.h>
+ #include <fcntl.h>
+ #include <time.h>
++#include <sys/select.h>
+ 
+ #define PGM "watchgnupg"
+ 
Signature · 1 line
1.9.1
#2 lore raw

PATCH 2/2 package/gnupg2: really do not run tests in cross-compilation

In reply to Yann E. MORIN

gnupg2 still tries to run test even if it tries to avoid it in cross-compilation.

Really disable running the tests.

Since that requires a complete autoreconf, the existing patch against configure is turned into a patch against the m4 macro.

since we autoreconf, we slightly patch configure.ac to not emit git errors on stderr because it is not in a git tree.

Cc Thomas Petazzoni <[email protected]>
---
 .../0000-silence-git-when-not-in-git-tree.patch    | 26 +++++++++++++++
 package/gnupg2/0001-fix-pth-config-usage.patch     | 29 ++++++++--------
 package/gnupg2/0003-dont-run-tests.patch           | 39 ++++++++++++++++++++++
 package/gnupg2/gnupg2.mk                           |  4 +++
 4 files changed, 82 insertions(+), 16 deletions(-)
 create mode 100644 package/gnupg2/0000-silence-git-when-not-in-git-tree.patch
 create mode 100644 package/gnupg2/0003-dont-run-tests.patch
package/gnupg2/0000-silence-git-when-not-in-git-tree.patch new file +26 -0
new file mode 100644
index 0000000..5eaefe4
@@ -0,0 +1,26 @@
+configure: silence autoreconf when not in a git tree
+
+When autoreconfiguring ourside of a git tree, the output is verbose
+with git errors.
+
+Silence that by consigning stderr to oblivion.
+
+Signed-off-by: "Yann E. MORIN" <[email protected]>
+
+diff -durN gnupg2-2.0.29.orig/configure.ac gnupg2-2.0.29/configure.ac
+--- gnupg2-2.0.29.orig/configure.ac	2015-09-08 14:39:24.000000000 +0200
++++ gnupg2-2.0.29/configure.ac	2015-12-16 18:34:47.099493863 +0100
+@@ -33,11 +33,11 @@
+ # flag indicating a development version (mym4_isgit).  Note that the
+ # m4 processing is done by autoconf and not during the configure run.
+ m4_define([mym4_revision],
+-          m4_esyscmd([git rev-parse --short HEAD | tr -d '\n\r']))
++          m4_esyscmd([git rev-parse --short HEAD 2>/dev/null | tr -d '\n\r']))
+ m4_define([mym4_revision_dec],
+           m4_esyscmd_s([echo $((0x$(echo ]mym4_revision[|head -c 4)))]))
+ m4_define([mym4_betastring],
+-          m4_esyscmd_s([git describe --match 'gnupg-2.[0-9].*[0-9]' --long|\
++          m4_esyscmd_s([git describe --match 'gnupg-2.[0-9].*[0-9]' --long 2>/dev/null |\
+                         awk -F- '$3!=0{print"-beta"$3}']))
+ m4_define([mym4_isgit],m4_if(mym4_betastring,[],[no],[yes]))
+ m4_define([mym4_full_version],[mym4_version[]mym4_betastring])
package/gnupg2/0001-fix-pth-config-usage.patch +13 -16
index e58aad3..756d351 100644
@@ -6,22 +6,19 @@ original one when reporting the version number. This patch to gnupg2's
 configure script adjusts the version checking to support this
 difference, since Buildroot uses pth-config from pthsem.
 
-We patch directly the configure script, because triggering the entire
-autoreconf dance for just a one byte change in the configure script
-seems a bit silly.
-
 Signed-off-by: Thomas Petazzoni <[email protected]>
+[yann.morin.1998 at free.fr: adapt to patch m4 macro for autoreconfiguring]
+Signed-off-by: "Yann E. MORIN" <[email protected]>
 
-Index: b/configure
-===================================================================
---- a/configure
-+++ b/configure
-@@ -8346,7 +8346,7 @@
-   tmp=1.3.7
-   if test "$PTH_CONFIG" != "no"; then
+diff -durN gnupg2-2.0.29.orig/m4/gnupg-pth.m4 gnupg2-2.0.29/m4/gnupg-pth.m4
+--- gnupg2-2.0.29.orig/m4/gnupg-pth.m4	2015-09-08 14:39:24.000000000 +0200
++++ gnupg2-2.0.29/m4/gnupg-pth.m4	2015-12-16 18:30:54.336513493 +0100
+@@ -17,7 +17,7 @@
+ # Taken and modified from the m4 macros which come with Pth.
+ AC_DEFUN([GNUPG_PTH_VERSION_CHECK],
+   [
+-    _pth_version=`$PTH_CONFIG --version | awk 'NR==1 {print [$]3}'`
++    _pth_version=`$PTH_CONFIG --version | awk 'NR==1 {print [$]2}'`
+     _req_version="ifelse([$1],,1.2.0,$1)"
  
--    _pth_version=`$PTH_CONFIG --version | awk 'NR==1 {print $3}'`
-+    _pth_version=`$PTH_CONFIG --version | awk 'NR==1 {print $2}'`
-     _req_version="$tmp"
- 
-     { $as_echo "$as_me:${as_lineno-$LINENO}: checking for PTH - version >= $_req_version" >&5
+     AC_MSG_CHECKING(for PTH - version >= $_req_version)
package/gnupg2/0003-dont-run-tests.patch new file +39 -0
new file mode 100644
index 0000000..1858a0d
@@ -0,0 +1,39 @@
+tests: really don't run tests in cross-compilation
+
+Signed-off-by: "Yann E. MORIN" <[email protected]>
+
+diff -durN gnupg2-2.0.29.orig/tests/Makefile.am gnupg2-2.0.29/tests/Makefile.am
+--- gnupg2-2.0.29.orig/tests/Makefile.am	2015-09-08 14:39:24.000000000 +0200
++++ gnupg2-2.0.29/tests/Makefile.am	2015-12-16 18:27:55.778227183 +0100
+@@ -64,6 +64,7 @@
+ asschk_SOURCES = asschk.c
+ 
+ 
++if RUN_GPG_TESTS
+ all-local: inittests.stamp
+ 
+ clean-local:
+@@ -72,4 +73,5 @@
+ inittests.stamp: inittests
+ 	srcdir=$(srcdir) $(TESTS_ENVIRONMENT) $(srcdir)/inittests
+ 	echo timestamp >./inittests.stamp
++endif # RUN_GPG_TESTS
+ 
+diff -durN gnupg2-2.0.29.orig/tests/pkits/Makefile.am gnupg2-2.0.29/tests/pkits/Makefile.am
+--- gnupg2-2.0.29.orig/tests/pkits/Makefile.am	2015-09-01 08:52:21.000000000 +0200
++++ gnupg2-2.0.29/tests/pkits/Makefile.am	2015-12-16 18:44:06.422655594 +0100
+@@ -52,6 +52,7 @@
+ 
+ DISTCLEANFILES = pubring.kbx~ random_seed
+ 
++if RUN_GPG_TESTS
+ all-local: inittests.stamp
+ 
+ clean-local:
+@@ -71,5 +72,5 @@
+            elif test $$? -eq 77; then echo "- SKIP $$tst"; \
+ 	   fi; \
+          done
+-
++endif # RUN_GPG_TESTS
+ 
package/gnupg2/gnupg2.mk +4 -0
index 6d279fd..2e1f067 100644
@@ -11,6 +11,10 @@ GNUPG2_LICENSE = GPLv3+
 GNUPG2_LICENSE_FILES = COPYING
 GNUPG2_DEPENDENCIES = zlib libgpg-error libgcrypt libassuan libksba libpthsem \
 	$(if $(BR2_PACKAGE_LIBICONV),libiconv)
+
+# Patching configure.ac and m4 macros, as well as Makefile.am
+GNUPG2_AUTORECONF = YES
+
 GNUPG2_CONF_OPTS = \
 	--disable-rpath --disable-regex --disable-doc \
 	--with-libgpg-error-prefix=$(STAGING_DIR)/usr \
Signature · 1 line
1.9.1
#3 lore raw

PATCH 2/2 package/gnupg2: really do not run tests in cross-compilation

Quoted 5 lines · Yann E. MORIN

gnupg2 still tries to run test even if it tries to avoid it in cross-compilation.

Really disable running the tests.

Quoted 8 lines

Since that requires a complete autoreconf, the existing patch against configure is turned into a patch against the m4 macro.

since we autoreconf, we slightly patch configure.ac to not emit git errors on stderr because it is not in a git tree.

Cc Thomas Petazzoni <[email protected]>

but minor nits.

Quoted 2 lines
---
 .../0000-silence-git-when-not-in-git-tree.patch    | 26 +++++++++++++++

I would have prefered to renumber the patches rather than adding a 0000. And
anyway it's not necessary to add it as the first patch, you can add it as a 0004 as well.

Quoted 13 lines
 package/gnupg2/0001-fix-pth-config-usage.patch     | 29 ++++++++--------
 package/gnupg2/0003-dont-run-tests.patch           | 39 ++++++++++++++++++++++
 package/gnupg2/gnupg2.mk                           |  4 +++
 4 files changed, 82 insertions(+), 16 deletions(-)
 create mode 100644 package/gnupg2/0000-silence-git-when-not-in-git-tree.patch
 create mode 100644 package/gnupg2/0003-dont-run-tests.patch
package/gnupg2/0000-silence-git-when-not-in-git-tree.patch new file +0 -0
new file mode 100644
index 0000000..5eaefe4

package/gnupg2/0000-silence-git-when-not-in-git-tree.patch line 8

Quoted 8 lines
package/gnupg2/0000-silence-git-when-not-in-git-tree.patch new file +8 -0
@@ -0,0 +1,26 @@
+configure: silence autoreconf when not in a git tree
+
+When autoreconfiguring ourside of a git tree, the output is verbose
+with git errors.
+
+Silence that by consigning stderr to oblivion.
+
+Signed-off-by: "Yann E. MORIN" <[email protected]>

Is this upstreamable?

Regards,
Arnout

Quoted 18 lines
package/gnupg2/0000-silence-git-when-not-in-git-tree.patch new file +18 -0
+
+diff -durN gnupg2-2.0.29.orig/configure.ac gnupg2-2.0.29/configure.ac
+--- gnupg2-2.0.29.orig/configure.ac	2015-09-08 14:39:24.000000000 +0200
++++ gnupg2-2.0.29/configure.ac	2015-12-16 18:34:47.099493863 +0100
+@@ -33,11 +33,11 @@
+ # flag indicating a development version (mym4_isgit).  Note that the
+ # m4 processing is done by autoconf and not during the configure run.
+ m4_define([mym4_revision],
+-          m4_esyscmd([git rev-parse --short HEAD | tr -d '\n\r']))
++          m4_esyscmd([git rev-parse --short HEAD 2>/dev/null | tr -d '\n\r']))
+ m4_define([mym4_revision_dec],
+           m4_esyscmd_s([echo $((0x$(echo ]mym4_revision[|head -c 4)))]))
+ m4_define([mym4_betastring],
+-          m4_esyscmd_s([git describe --match 'gnupg-2.[0-9].*[0-9]' --long|\
++          m4_esyscmd_s([git describe --match 'gnupg-2.[0-9].*[0-9]' --long 2>/dev/null |\
+                         awk -F- '$3!=0{print"-beta"$3}']))
+ m4_define([mym4_isgit],m4_if(mym4_betastring,[],[no],[yes]))
+ m4_define([mym4_full_version],[mym4_version[]mym4_betastring])

[snip]

Signature · 6 lines
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF
#4 lore raw
In reply to Yann E. MORIN

Yann,

Quoted 3 lines · Yann E. MORIN

Yann E. MORIN (2):
package/gnupg2: fix build with musl package/gnupg2: really do not run tests in cross-compilation

Both applied. On patch 2, as suggested by Arnout, I have renamed patch 0000 to 0004.

As suggested by Arnout as well, please submit the patches upstream :-)

Thanks!

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com