Public bug reported:

Affecting Artful:

According to the publishing history, the package Simbody for Artful was
directly copied from Zesty:
https://launchpad.net/ubuntu/+source/simbody/3.5.4+dfsg-
1ubuntu1/+publishinghistory

The problem is that a couple of dependencies of Simbody (libblas and
liblapack) changed the location of their libraries to be arch dependent
(from /usr/lib to /usr/lib/$arch-triplet/) and these paths are hardcoded
by Simbody in the cmake module at building time.

[Impact]

* Any software using Simbody cmake module won't compile because of non existant 
lib paths in Simbody exported libraries.
  - See 
https://build.osrfoundation.org/job/gazebo-ci-default-artful-amd64-gpu-none/23/consoleFull#-14256201906ea37b8d-a6d4-40ae-8431-d90c018842af

* Rebuilding the package (binNMU) without any modifications to the
source code would make cmake to get the new paths and hardcode them in
the cmake module.

[Test Case]

* To check that there are non-existant paths in the current cmake
module:

 - apt-get install libsimbody-dev
 - if [[ -f $(sed -n 's:.*\(/usr/.*libblas.so\).*:\1:p' 
/usr/lib/x86_64-linux-gnu/cmake/simbody/SimbodyConfig.cmake | uniq) ]]; then 
echo "fixed!"; else echo "Wrong path in cmake module"; fi

* If you run the oneliner in any other release different than Artful it
will return a success message.

[Regression Potential]

 * There is no change in code or in Debian metadata, there is a change
in the building environment. Changing the build environment from Zesty
to Artful could produce some changes in the binary interface but given
that it should be its 'native enviroment' I don't expect any
incompatibility introduced by the change in the compiler of build
flags/libraries.

* The lack of changes in code should keep the API stable.

* The lack of change in debian metadata should keep the packaging
stable.

* To minimize the possible problems in the binary interface or any other thing 
related to the change I created a PPA and rebuild the current version:
  - https://launchpad.net/~j-rivero/+archive/ubuntu/simbody-artful

* I've used Gazebo simulator (which uses Simbody) large test suite for trying 
to detect possible problems. All seems fine and problems are not related to 
this bug:
  - 
https://build.osrfoundation.org/job/gazebo-ci-default-artful-amd64-gpu-none/25/console

* This both shows the SRU team that the risks have been considered, and
provides guidance to testers in regression-testing the SRU.

[Other Info]

* I'm currently the maintainer of Simbody in Debian

** Affects: simbody (Ubuntu)
     Importance: Undecided
         Status: New


** Tags: artful

** Description changed:

  Affecting Artful:
  
- According to the publishing history, the package Simbody was directly copied 
from Zesty:
- 
https://launchpad.net/ubuntu/+source/simbody/3.5.4+dfsg-1ubuntu1/+publishinghistory
+ According to the publishing history, the package Simbody for Artful was
+ directly copied from Zesty:
+ https://launchpad.net/ubuntu/+source/simbody/3.5.4+dfsg-
+ 1ubuntu1/+publishinghistory
  
- The problem is that a couple of dependencies of Simbody (libblas and 
liblapack) changed the
- location of their libraries to be arch dependent (from /usr/lib to 
/usr/lib/$arch-triplet/)
- and these paths are hardcoded by Simbody in the cmake module at building time.
+ The problem is that a couple of dependencies of Simbody (libblas and
+ liblapack) changed the location of their libraries to be arch dependent
+ (from /usr/lib to /usr/lib/$arch-triplet/) and these paths are hardcoded
+ by Simbody in the cmake module at building time.
  
  [Impact]
  
-  * Any software using Simbody cmake module won't compile because of non 
existant lib
-    paths in Simbody exported libraries.
-    - See 
https://build.osrfoundation.org/job/gazebo-ci-default-artful-amd64-gpu-none/23/consoleFull#-14256201906ea37b8d-a6d4-40ae-8431-d90c018842af
+ * Any software using Simbody cmake module won't compile because of non 
existant lib paths in Simbody exported libraries.
+   - See 
https://build.osrfoundation.org/job/gazebo-ci-default-artful-amd64-gpu-none/23/consoleFull#-14256201906ea37b8d-a6d4-40ae-8431-d90c018842af
  
-  * Rebuilding the package (binNMU) without any modifications to the source 
code would make
-    cmake to get the new paths and hardcode them in the cmake module.
+ * Rebuilding the package (binNMU) without any modifications to the
+ source code would make cmake to get the new paths and hardcode them in
+ the cmake module.
  
  [Test Case]
  
-  * To check that there are non-existant paths in the current cmake
+ * To check that there are non-existant paths in the current cmake
  module:
  
-     - apt-get install libsimbody-dev
-     - if [[ -f $(sed -n 's:.*\(/usr/.*libblas.so\).*:\1:p' 
/usr/lib/x86_64-linux-gnu/cmake/simbody/SimbodyConfig.cmake | uniq) ]]; then 
echo "fixed!"; else echo "Wrong path in cmake module"; fi
+  - apt-get install libsimbody-dev
+  - if [[ -f $(sed -n 's:.*\(/usr/.*libblas.so\).*:\1:p' 
/usr/lib/x86_64-linux-gnu/cmake/simbody/SimbodyConfig.cmake | uniq) ]]; then 
echo "fixed!"; else echo "Wrong path in cmake module"; fi
  
-  * If you run the oneliner in any other release different than Artful it will 
return a 
-    success message.
+ * If you run the oneliner in any other release different than Artful it
+ will return a success message.
  
  [Regression Potential]
  
-  * There is no change in code or in Debian metadata, there is a change in the 
building
-    environment. Changing the build environment from Zesty to Artful could 
produce some
-    changes in the binary interface but given that it should be its 'native 
enviroment' I
-    don't expect any incompatibility introduced by the change in the compiler 
of build
-    flags/libraries.
+  * There is no change in code or in Debian metadata, there is a change
+ in the building environment. Changing the build environment from Zesty
+ to Artful could produce some changes in the binary interface but given
+ that it should be its 'native enviroment' I don't expect any
+ incompatibility introduced by the change in the compiler of build
+ flags/libraries.
  
-  * The lack of changes in code should keep the API stable.
+ * The lack of changes in code should keep the API stable.
  
-  * The lack of change in debian metadata should keep the packaging stable.
-  
-  * To minimize the possible problems in the binary interface or any other 
thing related
-    to the change I created a PPA and rebuild the current version:
-      - https://launchpad.net/~j-rivero/+archive/ubuntu/simbody-artful
+ * The lack of change in debian metadata should keep the packaging
+ stable.
  
-   * I've used Gazebo simulator (which uses Simbody) large test suite for 
trying to
-     detect possible problems. All seems fine and problems are not related to 
this bug:
-       - 
https://build.osrfoundation.org/job/gazebo-ci-default-artful-amd64-gpu-none/25/console
+ * To minimize the possible problems in the binary interface or any other 
thing related to the change I created a PPA and rebuild the current version:
+   - https://launchpad.net/~j-rivero/+archive/ubuntu/simbody-artful
  
-  * This both shows the SRU team that the risks have been considered,
-    and provides guidance to testers in regression-testing the SRU.
+ * I've used Gazebo simulator (which uses Simbody) large test suite for trying 
to detect possible problems. All seems fine and problems are not related to 
this bug:
+   - 
https://build.osrfoundation.org/job/gazebo-ci-default-artful-amd64-gpu-none/25/console
+ 
+ * This both shows the SRU team that the risks have been considered, and
+ provides guidance to testers in regression-testing the SRU.
  
  [Other Info]
-  
-  * I'm currently the maintainer of Simbody in Debian
+ 
+ * I'm currently the maintainer of Simbody in Debian

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1770678

Title:
  [SRU] Simbody package ships erroneous paths for blas/lapack in the
  cmake module

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/simbody/+bug/1770678/+subscriptions

-- 
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to