Skip to contentSkip to conversation

package/cmake: revert the bump to 3.7

PATCH New Yann E. MORIN 5 participants 17 messages Feb 25, 2017 – Feb 26, 2017

lore
Share & tools
lore
#0 lore raw

cover package/cmake: revert the bump to 3.7

Hello All!

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

See for example:

- domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7abf973691b39a0ca1bb4e07d749593a/
- freerdp : http://autobuild.buildroot.org/results/5d4/5d429d0e288754a541ee5d8be515454c5fccd28b/
- libcec : http://autobuild.buildroot.org/results/3f3/3f3593bab7734dd274faf5b5690895e9424cbb89/

(and many others)

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk having to hunt down packages one by one.

Regards,
Yann E. MORIN.

The following changes since commit a3c07d08eae55deeb1b1731174309750bb0ea628

util-linux: re-add <pkg>_AUTORECONF = YES (2017-02-24 12:00:02 +0100)

are available in the git repository at:

git://git.buildroot.org/~ymorin/git/buildroot.git

for you to fetch changes up to df3990e80996775534d5d823ea9bb772836b945d

Revert "cmake: bump version to 3.7.0" (2017-02-25 19:40:13 +0100)

----------------------------------------------------------------
Yann E. MORIN (4):
Revert "cmake: bump version to 3.7.2" Revert "cmake: adjust dependencies added for libuv" Revert "cmake: bump version to 3.7.1" Revert "cmake: bump version to 3.7.0"

 package/cmake/Config.in  | 9 ++-------
 package/cmake/cmake.hash | 4 ++--
 package/cmake/cmake.mk   | 6 +++---
 3 files changed, 7 insertions(+), 12 deletions(-)
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 1/4 Revert "cmake: bump version to 3.7.2"

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

This reverts commit f8a6990c92451dab2ff9e14f3e94fdecbd79b604.

Cc Vicente Olivert Riera <[email protected]>
Cc Samuel Martin <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc J?rg Krause <[email protected]>
Cc Baruch Siach <[email protected]>
---
 package/cmake/cmake.hash | 4 ++--
 package/cmake/cmake.mk   | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)
package/cmake/cmake.hash +2 -2
index 13c0237..7a22f21 100644
@@ -1,2 +1,2 @@
-# From http://www.cmake.org/files/v3.7/cmake-3.7.2-SHA-256.txt
-sha256 dc1246c4e6d168ea4d6e042cfba577c1acd65feea27e56f5ff37df920c30cae0  cmake-3.7.2.tar.gz
+# From http://www.cmake.org/files/v3.7/cmake-3.7.1-SHA-256.txt
+sha256 449a5bce64dbd4d5b9517ebd1a1248ed197add6ad27934478976fd5f1f9330e1  cmake-3.7.1.tar.gz
package/cmake/cmake.mk +1 -1
index 0fd7139..107715c 100644
@@ -5,7 +5,7 @@
 ################################################################################
 
 CMAKE_VERSION_MAJOR = 3.7
-CMAKE_VERSION = $(CMAKE_VERSION_MAJOR).2
+CMAKE_VERSION = $(CMAKE_VERSION_MAJOR).1
 CMAKE_SITE = https://cmake.org/files/v$(CMAKE_VERSION_MAJOR)
 CMAKE_LICENSE = BSD-3c
 CMAKE_LICENSE_FILES = Copyright.txt
Signature · 1 line
2.7.4
#2 lore raw

PATCH 2/4 Revert "cmake: adjust dependencies added for libuv"

In reply to Yann E. MORIN

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

This reverts commit b754237520a4795eadc0edec768b267dd598ab0e.

Cc Vicente Olivert Riera <[email protected]>
Cc Samuel Martin <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc J?rg Krause <[email protected]>
Cc Baruch Siach <[email protected]>
---
 package/cmake/Config.in | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
package/cmake/Config.in +3 -3
index 2706db5..716d80e 100644
@@ -26,7 +26,7 @@ config BR2_PACKAGE_CMAKE_CTEST
 	depends on BR2_USE_WCHAR # libarchive
 	depends on BR2_INSTALL_LIBSTDCPP
 	depends on !BR2_STATIC_LIBS
-	depends on BR2_TOOLCHAIN_HAS_THREADS_NPTL # libuv
+	depends on BR2_TOOLCHAIN_HAS_THREADS # libuv
 	depends on BR2_USE_MMU # libuv
 	help
 	  CTest is a testing tool distributed as a part of CMake. It
@@ -37,9 +37,9 @@ config BR2_PACKAGE_CMAKE_CTEST
 
 	  http://www.cmake.org/
 
-comment "ctest needs a toolchain w/ C++, wchar, dynamic library, gcc >= 4.7, NPTL"
+comment "ctest needs a toolchain w/ C++, wchar, dynamic library, gcc >= 4.7, threads"
 	depends on BR2_PACKAGE_CMAKE_ARCH_SUPPORTS
 	depends on BR2_USE_MMU
 	depends on !BR2_INSTALL_LIBSTDCPP || !BR2_USE_WCHAR || \
 		BR2_STATIC_LIBS || !BR2_TOOLCHAIN_GCC_AT_LEAST_4_7 || \
-		!BR2_TOOLCHAIN_HAS_THREADS_NPTL
+		!BR2_TOOLCHAIN_HAS_THREADS
Signature · 1 line
2.7.4
#3 lore raw

PATCH 3/4 Revert "cmake: bump version to 3.7.1"

In reply to Yann E. MORIN

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

This reverts commit d96fafc3d346af8790df45557b16069d6ce81e79.

Cc Vicente Olivert Riera <[email protected]>
Cc Samuel Martin <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc J?rg Krause <[email protected]>
Cc Baruch Siach <[email protected]>
---
 package/cmake/cmake.hash | 4 ++--
 package/cmake/cmake.mk   | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)
package/cmake/cmake.hash +2 -2
index 7a22f21..687e5a4 100644
@@ -1,2 +1,2 @@
-# From http://www.cmake.org/files/v3.7/cmake-3.7.1-SHA-256.txt
-sha256 449a5bce64dbd4d5b9517ebd1a1248ed197add6ad27934478976fd5f1f9330e1  cmake-3.7.1.tar.gz
+# From http://www.cmake.org/files/v3.7/cmake-3.7.0-SHA-256.txt
+sha256 ed63e05c41aeb6c036e503114ab15847f29c312f9f21f5f1a7060a4b4ec2fb31  cmake-3.7.0.tar.gz
package/cmake/cmake.mk +1 -1
index 107715c..e4904fc 100644
@@ -5,7 +5,7 @@
 ################################################################################
 
 CMAKE_VERSION_MAJOR = 3.7
-CMAKE_VERSION = $(CMAKE_VERSION_MAJOR).1
+CMAKE_VERSION = $(CMAKE_VERSION_MAJOR).0
 CMAKE_SITE = https://cmake.org/files/v$(CMAKE_VERSION_MAJOR)
 CMAKE_LICENSE = BSD-3c
 CMAKE_LICENSE_FILES = Copyright.txt
Signature · 1 line
2.7.4
#4 lore raw

PATCH 4/4 Revert "cmake: bump version to 3.7.0"

In reply to Yann E. MORIN

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

This reverts commit 1ffcf364b6e9894a876dc581a090f87685945412.

Cc Vicente Olivert Riera <[email protected]>
Cc Samuel Martin <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc J?rg Krause <[email protected]>
Cc Baruch Siach <[email protected]>
---
 package/cmake/Config.in  | 9 ++-------
 package/cmake/cmake.hash | 4 ++--
 package/cmake/cmake.mk   | 6 +++---
 3 files changed, 7 insertions(+), 12 deletions(-)
package/cmake/Config.in +2 -7
index 716d80e..6f04a4a 100644
@@ -20,14 +20,11 @@ config BR2_PACKAGE_CMAKE_CTEST
 	select BR2_PACKAGE_EXPAT
 	select BR2_PACKAGE_BZIP2
 	select BR2_PACKAGE_XZ
-	select BR2_PACKAGE_LIBUV
 	depends on BR2_PACKAGE_CMAKE_ARCH_SUPPORTS
 	depends on BR2_TOOLCHAIN_GCC_AT_LEAST_4_7 # from jsoncpp
 	depends on BR2_USE_WCHAR # libarchive
 	depends on BR2_INSTALL_LIBSTDCPP
 	depends on !BR2_STATIC_LIBS
-	depends on BR2_TOOLCHAIN_HAS_THREADS # libuv
-	depends on BR2_USE_MMU # libuv
 	help
 	  CTest is a testing tool distributed as a part of CMake. It
 	  can be used to automate updating (using CVS for example),
@@ -37,9 +34,7 @@ config BR2_PACKAGE_CMAKE_CTEST
 
 	  http://www.cmake.org/
 
-comment "ctest needs a toolchain w/ C++, wchar, dynamic library, gcc >= 4.7, threads"
+comment "ctest needs a toolchain w/ C++, wchar, dynamic library, gcc >= 4.7"
 	depends on BR2_PACKAGE_CMAKE_ARCH_SUPPORTS
-	depends on BR2_USE_MMU
 	depends on !BR2_INSTALL_LIBSTDCPP || !BR2_USE_WCHAR || \
-		BR2_STATIC_LIBS || !BR2_TOOLCHAIN_GCC_AT_LEAST_4_7 || \
-		!BR2_TOOLCHAIN_HAS_THREADS
+		BR2_STATIC_LIBS || !BR2_TOOLCHAIN_GCC_AT_LEAST_4_7
package/cmake/cmake.hash +2 -2
index 687e5a4..f846dc9 100644
@@ -1,2 +1,2 @@
-# From http://www.cmake.org/files/v3.7/cmake-3.7.0-SHA-256.txt
-sha256 ed63e05c41aeb6c036e503114ab15847f29c312f9f21f5f1a7060a4b4ec2fb31  cmake-3.7.0.tar.gz
+# From http://www.cmake.org/files/v3.6/cmake-3.6.3-SHA-256.txt
+sha256 7d73ee4fae572eb2d7cd3feb48971aea903bb30a20ea5ae8b4da826d8ccad5fe  cmake-3.6.3.tar.gz
package/cmake/cmake.mk +3 -3
index e4904fc..bc55509 100644
@@ -4,8 +4,8 @@
 #
 ################################################################################
 
-CMAKE_VERSION_MAJOR = 3.7
-CMAKE_VERSION = $(CMAKE_VERSION_MAJOR).0
+CMAKE_VERSION_MAJOR = 3.6
+CMAKE_VERSION = $(CMAKE_VERSION_MAJOR).3
 CMAKE_SITE = https://cmake.org/files/v$(CMAKE_VERSION_MAJOR)
 CMAKE_LICENSE = BSD-3c
 CMAKE_LICENSE_FILES = Copyright.txt
@@ -22,7 +22,7 @@ CMAKE_LICENSE_FILES = Copyright.txt
 #   the system-wide libraries instead of rebuilding and statically
 #   linking with the ones bundled into the CMake sources.
 
-CMAKE_DEPENDENCIES = zlib jsoncpp libcurl libarchive expat bzip2 xz libuv
+CMAKE_DEPENDENCIES = zlib jsoncpp libcurl libarchive expat bzip2 xz
 
 CMAKE_CONF_OPTS = \
 	-DKWSYS_LFS_WORKS=TRUE \
Signature · 1 line
2.7.4
#5 lore raw

cover package/cmake: revert the bump to 3.7

In reply to Yann E. MORIN

Hi Yann,

Quoted 9 lines · Yann E. MORIN

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

See for example:

- domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7abf973691b39a0ca1bb4e07d749593a/
- freerdp : http://autobuild.buildroot.org/results/5d4/5d429d0e288754a541ee5d8be515454c5fccd28b/
- libcec : http://autobuild.buildroot.org/results/3f3/3f3593bab7734dd274faf5b5690895e9424cbb89/

Quoted 7 lines

(and many others)

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk having to hunt down packages one by one.

What about host installed cmake? Maybe do something like the following (untested) to blacklist version 3.7.x?

support/dependencies/check-host-cmake.sh +4 -0
index 9b63b0648d97..1c4fec8cae41 100755
@@ -27,6 +27,10 @@ version="$(${cmake} --version \
 major="${version%.*}"
 minor="${version#*.}"
 
+if [ ${version} = 3.7 ]; then
+	exit 1
+fi
+
 if [ ${major} -gt ${major_min} ]; then
     echo "${cmake}"
 else

baruch

Signature · 3 lines
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -
#6 lore raw

cover package/cmake: revert the bump to 3.7

Baruch, All,

On 2017-02-25 22:44 +0200, Baruch Siach spake thusly:

Quoted 12 lines · Yann E. MORIN
Quoted 11 lines

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

See for example:

- domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7abf973691b39a0ca1bb4e07d749593a/
- freerdp : http://autobuild.buildroot.org/results/5d4/5d429d0e288754a541ee5d8be515454c5fccd28b/
- libcec : http://autobuild.buildroot.org/results/3f3/3f3593bab7734dd274faf5b5690895e9424cbb89/

(and many others)

Quoted 8 lines
Quoted 5 lines

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk having to hunt down packages one by one.

What about host installed cmake? Maybe do something like the following (untested) to blacklist version 3.7.x?

Good catch, indeed.

support/dependencies/check-host-cmake.sh version="$(${cmake} --version \ · line 31

Quoted 10 lines
support/dependencies/check-host-cmake.sh +2 -0
index 9b63b0648d97..1c4fec8cae41 100755
@@ -27,6 +27,10 @@ version="$(${cmake} --version \
 major="${version%.*}"
 minor="${version#*.}"
 
+if [ ${version} = 3.7 ]; then
+	exit 1

Minor nit: the rest of the script uses leading spaces.

Also, we should probably catch >= 3.7, no?

Regards,
Yann E. MORIN.

Quoted 12 lines
support/dependencies/check-host-cmake.sh +2 -0
+fi
+
 if [ ${major} -gt ${major_min} ]; then
     echo "${cmake}"
 else

baruch

Signature · 3 lines
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -
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.  |
'------------------------------^-------^------------------^--------------------'
#7 lore raw

cover package/cmake: revert the bump to 3.7

Hi Yann,

Quoted 32 lines · Yann E. MORIN

On 2017-02-25 22:44 +0200, Baruch Siach spake thusly:

Quoted 12 lines · Yann E. MORIN
Quoted 11 lines

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

See for example:

- domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7abf973691b39a0ca1bb4e07d749593a/
- freerdp : http://autobuild.buildroot.org/results/5d4/5d429d0e288754a541ee5d8be515454c5fccd28b/
- libcec : http://autobuild.buildroot.org/results/3f3/3f3593bab7734dd274faf5b5690895e9424cbb89/

(and many others)

Quoted 8 lines
Quoted 5 lines

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk having to hunt down packages one by one.

What about host installed cmake? Maybe do something like the following (untested) to blacklist version 3.7.x?

Good catch, indeed.

Quoted 8 lines
support/dependencies/check-host-cmake.sh +0 -0
index 9b63b0648d97..1c4fec8cae41 100755
@@ -27,6 +27,10 @@ version="$(${cmake} --version \
 major="${version%.*}"
 minor="${version#*.}"
 
Quoted 6 lines

support/dependencies/check-host-cmake.sh version="$(${cmake} --version \ · line 31

Quoted 2 lines
support/dependencies/check-host-cmake.sh +2 -0
+if [ ${version} = 3.7 ]; then
+	exit 1

Minor nit: the rest of the script uses leading spaces.

Also, we should probably catch >= 3.7, no?

The latest cmake release is 3.7.2. Let's hope that this issue gets fixed before 3.8. If not, we can always add more blacklisted versions.

baruch

Signature · 3 lines
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -
#8 lore raw

cover package/cmake: revert the bump to 3.7

In reply to Yann E. MORIN

Hi,

Quoted 4 lines · Yann E. MORIN

Hello All!

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

For an explanation what's going wrong please have a look at [1].

Quoted 10 lines

See for example:

? - domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7ab
f973691b39a0ca1bb4e07d749593a/
? - freerdp??: http://autobuild.buildroot.org/results/5d4/5d429d0e288
754a541ee5d8be515454c5fccd28b/
? - libcec???: http://autobuild.buildroot.org/results/3f3/3f3593bab77
34dd274faf5b5690895e9424cbb89/

(and many others)

For example bctoolbox. Although the build errors are fixed, CMake does not find the optionally mbedTLS package if zlib is available as it links against the host zlib when using check_symbol_exists().

Quoted 6 lines

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk
having to hunt down packages one by one.

A fix which worked for me was to remove the line set(CMAKE_SYSTEM_NAME Linux) from the toolchainfile to be able to set the flag FIND_LIBRARY_USE_LIB32_PATHS property to FALSE. Of course, we also have to set the necessary flags which are set depending on CMAKE_SYSTEM_NAME. However, reverting CMake is probably much cleaner.

[1] http://lists.busybox.net/pipermail/buildroot/2017-February/183579.h
tml
[2] https://git.buildroot.net/buildroot/tree/support/misc/toolchainfile
.cmake.in#n13

#9 lore raw

cover package/cmake: revert the bump to 3.7

J?rg, All,

On 2017-02-26 09:44 +0100, J?rg Krause spake thusly:

Quoted 8 lines · Yann E. MORIN
Quoted 4 lines

Hello All!

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

For an explanation what's going wrong please have a look at [1].
[1] http://lists.busybox.net/pipermail/buildroot/2017-February/183579.html

Are you sure this is related? The code you point to in this mail should not be executed in cross-compilation:

if(NOT CMAKE_CROSSCOMPILING AND EXISTS "/etc/debian_version")
set_property(GLOBAL PROPERTY FIND_LIBRARY_USE_LIB32_PATHS FALSE) set_property(GLOBAL PROPERTY FIND_LIBRARY_USE_LIB64_PATHS FALSE)
endif()

So, if this code is indeed the culprit, then it eans that CMAKE_CROSSCOMPILING is not set, and that in turn is probably the real issue.

(But I only had one coffee since I woke up, so I may have missed something...)

Regards,
Yann E. MORIN.

Quoted 32 lines
Quoted 10 lines

See for example:

? - domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7ab
f973691b39a0ca1bb4e07d749593a/
? - freerdp??: http://autobuild.buildroot.org/results/5d4/5d429d0e288
754a541ee5d8be515454c5fccd28b/
? - libcec???: http://autobuild.buildroot.org/results/3f3/3f3593bab77
34dd274faf5b5690895e9424cbb89/

(and many others)

For example bctoolbox. Although the build errors are fixed, CMake does not find the optionally mbedTLS package if zlib is available as it links against the host zlib when using check_symbol_exists().

Quoted 6 lines

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk
having to hunt down packages one by one.

A fix which worked for me was to remove the line set(CMAKE_SYSTEM_NAME Linux) from the toolchainfile to be able to set the flag FIND_LIBRARY_USE_LIB32_PATHS property to FALSE. Of course, we also have to set the necessary flags which are set depending on CMAKE_SYSTEM_NAME. However, reverting CMake is probably much cleaner.

[1] http://lists.busybox.net/pipermail/buildroot/2017-February/183579.h
tml
[2] https://git.buildroot.net/buildroot/tree/support/misc/toolchainfile
.cmake.in#n13

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.  |
'------------------------------^-------^------------------^--------------------'
#10 lore raw

cover package/cmake: revert the bump to 3.7

Hi Yann,

Quoted 25 lines · Yann E. MORIN

J?rg, All,

On 2017-02-26 09:44 +0100, J?rg Krause spake thusly:

Quoted 10 lines · Yann E. MORIN
Quoted 5 lines

Hello All!

cmake 3.7 causes serious regressions in some cmake-based packages,
related to how RPATH is handled.

For an explanation what's going wrong please have a look at [1].
[1] http://lists.busybox.net/pipermail/buildroot/2017-February/1835
79.html

Are you sure this is related? The code you point to in this mail should
not be executed in cross-compilation:

????if(NOT CMAKE_CROSSCOMPILING AND EXISTS "/etc/debian_version") ??????set_property(GLOBAL PROPERTY FIND_LIBRARY_USE_LIB32_PATHS FALSE)
??????set_property(GLOBAL PROPERTY FIND_LIBRARY_USE_LIB64_PATHS FALSE)
????endif()

Quoted 4 lines

So, if this code is indeed the culprit, then it eans that CMAKE_CROSSCOMPILING is not set, and that in turn is probably the real
issue.

By default FIND_LIBRARY_USE_LIB32_PATHS is set to TRUE [1], but is set to FALSE if NOT cross-compiling AND building on a Debian host.

Quoted 2 lines

(But I only had one coffee since I woke up, so I may have missed something...)

Me, too :-)

[1] https://cmake.org/gitweb?p=cmake.git;a=blob;f=Modules/Platform/Unix
Paths.cmake;h=d6c3d41f3112207ee11b42be18db438103e771f7;hb=896ad251#l97

#11 lore raw

cover package/cmake: revert the bump to 3.7

In reply to Yann E. MORIN
Quoted 9 lines · Yann E. MORIN

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

See for example:

- domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7abf973691b39a0ca1bb4e07d749593a/
- freerdp : http://autobuild.buildroot.org/results/5d4/5d429d0e288754a541ee5d8be515454c5fccd28b/
- libcec : http://autobuild.buildroot.org/results/3f3/3f3593bab7734dd274faf5b5690895e9424cbb89/

Quoted 7 lines

(and many others)

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk having to hunt down packages one by one.

Could you please submit a report to CMake's issue tracker? Regressions are taken seriously and it'd be nice to have this fixed for 3.8.0 (which is currently in its RC cycle).

https://gitlab.kitware.com/cmake/cmake/issues

Thanks,

--Ben

#12 lore raw

cover package/cmake: revert the bump to 3.7

Ben, All,

On 2017-02-26 16:32 +0000, Ben Boeckel spake thusly:

Quoted 15 lines · Yann E. MORIN
Quoted 14 lines

cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

See for example:

- domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7abf973691b39a0ca1bb4e07d749593a/
- freerdp : http://autobuild.buildroot.org/results/5d4/5d429d0e288754a541ee5d8be515454c5fccd28b/
- libcec : http://autobuild.buildroot.org/results/3f3/3f3593bab7734dd274faf5b5690895e9424cbb89/

(and many others)

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Quoted 7 lines
Quoted 2 lines

Properly fixing this so close to the release is problematic; we'd risk having to hunt down packages one by one.

Could you please submit a report to CMake's issue tracker? Regressions are taken seriously and it'd be nice to have this fixed for 3.8.0 (which is currently in its RC cycle).
https://gitlab.kitware.com/cmake/cmake/issues

Yes, that would be awesome.

However, I don't know how to caracterise the bug with a minimal reproducer. Nor am I able to explain exactly what happens in details...

Maybe J?rg (which has already made an extensive analyis of the problem) or you (if you are knowledgeable about cmake) could do it? ;-)

Thanks!

Regards,
Yann E. MORIN.

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.  |
'------------------------------^-------^------------------^--------------------'
#13 lore raw

cover package/cmake: revert the bump to 3.7

Quoted 2 lines · Yann E. MORIN

However, I don't know how to caracterise the bug with a minimal reproducer. Nor am I able to explain exactly what happens in details...

That's fine for an initial report; getting it on the radar (with buildroot devs in the loop) would be a good start though.

Quoted 2 lines

Maybe J?rg (which has already made an extensive analyis of the problem) or you (if you are knowledgeable about cmake) could do it? ;-)

I do work on CMake upstream, but I'm not familiar with this functionality specifically at the moment.

--Ben

#14 lore raw

cover package/cmake: revert the bump to 3.7

In reply to Yann E. MORIN

Hi Ben, hi Yann,

Quoted 36 lines · Yann E. MORIN

Ben, All,

On 2017-02-26 16:32 +0000, Ben Boeckel spake thusly:

Quoted 24 lines · Yann E. MORIN
Quoted 15 lines

cmake 3.7 causes serious regressions in some cmake-based packages,
related to how RPATH is handled.

See for example:

? - domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54
c7abf973691b39a0ca1bb4e07d749593a/
? - freerdp??: http://autobuild.buildroot.org/results/5d4/5d429d0
e288754a541ee5d8be515454c5fccd28b/
? - libcec???: http://autobuild.buildroot.org/results/3f3/3f3593b
ab7734dd274faf5b5690895e9424cbb89/

(and many others)

Quoted 7 lines

All causes the link to be attemped against host libraries, which is
definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk
having to hunt down packages one by one.

Quoted 6 lines

Could you please submit a report to CMake's issue tracker? Regressions
are taken seriously and it'd be nice to have this fixed for 3.8.0 (which
is currently in its RC cycle). ????https://gitlab.kitware.com/cmake/cmake/issues

Yes, that would be awesome.

Quoted 7 lines

However, I don't know how to caracterise the bug with a minimal reproducer.
Nor am I able to explain exactly what happens in details...

Maybe J?rg (which has already made an extensive analyis of the problem)
or you (if you are knowledgeable about cmake) could do it? ;-)

In fact, I already submitted a report [1] to the CMake mailing list some weeks ago. I was not sure, if this is a bug or a lack of understanding at that moment.

Ray Donnelly replied in that thread he has a PR [2] in the loop "[..] that asks the linker (via the compiler) what its implicit search directories are instead.?It is the right way to do it IMHO, but I need to find time to finish it unfortunately."

I did not checked the PR, yet.

[1] http://public.kitware.com/pipermail/cmake/2017-February/064970.html
[2] https://gitlab.kitware.com/mingwandroid/cmake/commit/b937ff949d8fda
ab7d8b812d503f67f8cef69532

J?rg

#15 lore raw

cover package/cmake: revert the bump to 3.7

In reply to Yann E. MORIN
Quoted 12 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Yann" == Yann E MORIN <[email protected]> writes:

Hello All!
cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

See for example:

- domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7abf973691b39a0ca1bb4e07d749593a/
- freerdp : http://autobuild.buildroot.org/results/5d4/5d429d0e288754a541ee5d8be515454c5fccd28b/
- libcec : http://autobuild.buildroot.org/results/3f3/3f3593bab7734dd274faf5b5690895e9424cbb89/

Quoted 7 lines

(and many others)

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk having to hunt down packages one by one.

Committed, thanks.

Will you send an update to check-host-cmake.sh as well?

Signature · 1 line
Bye, Peter Korsgaard
#16 lore raw

cover package/cmake: revert the bump to 3.7

Peter, All,

On 2017-02-26 22:41 +0100, Peter Korsgaard spake thusly:

Quoted 14 lines
Quoted 13 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Yann" == Yann E MORIN <[email protected]> writes:

Hello All!
cmake 3.7 causes serious regressions in some cmake-based packages, related to how RPATH is handled.

See for example:

- domoticz : http://autobuild.buildroot.org/results/fd0/fd0ba54c7abf973691b39a0ca1bb4e07d749593a/
- freerdp : http://autobuild.buildroot.org/results/5d4/5d429d0e288754a541ee5d8be515454c5fccd28b/
- libcec : http://autobuild.buildroot.org/results/3f3/3f3593bab7734dd274faf5b5690895e9424cbb89/

(and many others)

Quoted 7 lines
Quoted 5 lines

All causes the link to be attemped against host libraries, which is definitely not appropriate...

Properly fixing this so close to the release is problematic; we'd risk having to hunt down packages one by one.

Committed, thanks.

Thanks! :-)

Quoted 1 line

Will you send an update to check-host-cmake.sh as well?

Yes.

Regards,
Yann E. MORIN.

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.  |
'------------------------------^-------^------------------^--------------------'