Bug#947881: [Pkg-opencl-devel] Bug#947881: pocl: fix build on hurd-i386

2020-01-19 Thread Andreas Beckmann 
On 01/01/2020 22.46, Samuel Thibault wrote:
> Andreas Beckmann, le mer. 01 janv. 2020 22:10:57 +0100, a ecrit:
>> On 01/01/2020 15.19, Samuel Thibault wrote:
>>> pocl currently FTBFS on hurd-i386 only due to a header inclusion
>>> condition bug, could you apply the attached patch?
>>
>> Nice. Thanks.
>> Does the buildlog show any symbols changes?
> 
> This is the diff I am getting

OK, now it only fails due to some tests failing on a (virtual) single
core CPU. Unfortunately my patch only fixed one of the two cases (the
test should just be skipped if insufficient compute units are available)
despite of doing the more or less same thing twice.
But my cmake skills are not that good, I don't see what's different here.


Andreas

Bug#947881: [Pkg-opencl-devel] Bug#947881: pocl: fix build on hurd-i386

2020-01-01 Thread Samuel Thibault 
Andreas Beckmann, le mer. 01 janv. 2020 22:10:57 +0100, a ecrit:
> On 01/01/2020 15.19, Samuel Thibault wrote:
> > pocl currently FTBFS on hurd-i386 only due to a header inclusion
> > condition bug, could you apply the attached patch?
> 
> Nice. Thanks.
> Does the buildlog show any symbols changes?

This is the diff I am getting

Samuel
dpkg-gensymbols: warning: debian/libpocl2/DEBIAN/symbols doesn't match 
completely debian/libpocl2.symbols
--- debian/libpocl2.symbols (libpocl2_1.3-10+hurd.1_hurd-i386)
+++ dpkg-gensymbols5QYPpU   2020-01-01 21:44:24.0 +
@@ -95,7 +95,7 @@
  
(optional=templinst)_ZN4llvm15callDefaultCtorIN4pocl14AllocasToEntryEEEPNS_4PassEv@Base
 0.10
  
(optional=templinst)_ZN4llvm15callDefaultCtorIN4pocl14IsolateRegionsEEEPNS_4PassEv@Base
 0.10
  
(optional=templinst)_ZN4llvm15callDefaultCtorIN4pocl18RemoveBarrierCallsEEEPNS_4PassEv@Base
 0.14
-#MISSING: 0.14-4~llvm3.9# 
(optional=templinst)_ZN4llvm15callDefaultCtorIN4pocl19TargetAddressSpacesEEEPNS_4PassEv@Base
 0.10
+#MISSING: 1.3-10+hurd.1# 
(optional=templinst)_ZN4llvm15callDefaultCtorIN4pocl19TargetAddressSpacesEEEPNS_4PassEv@Base
 0.10
  
(optional=templinst)_ZN4llvm15callDefaultCtorIN4pocl19WorkitemReplicationEEEPNS_4PassEv@Base
 0.10
  
(optional=templinst)_ZN4llvm15callDefaultCtorIN4pocl20CanonicalizeBarriersEEEPNS_4PassEv@Base
 0.10
  
(optional=templinst)_ZN4llvm15callDefaultCtorIN4pocl20ImplicitLoopBarriersEEEPNS_4PassEv@Base
 0.10
@@ -391,9 +391,9 @@
  
_ZNK4pocl27ImplicitConditionalBarriers16getAnalysisUsageERN4llvm13AnalysisUsageE@Base
 0.10
  (arch=!x32)_ZNK4pocl9Workgroup16getAnalysisUsageERN4llvm13AnalysisUsageE@Base 
1.3-5~llvm7
  (optional=templinst|arch=!x32)_ZNKSt5ctypeIcE8do_widenEc@Base 1.3-5~llvm7
-#MISSING: 1.3-3~gcc9# 
(optional=templinst)_ZNKSt8_Rb_treeIPN4llvm10BasicBlockES2_St9_IdentityIS2_ESt4lessIS2_ESaIS2_EE4findERKS2_@Base
 1.1-6~llvm6.0+gcc8
-#MISSING: 1.3-3~gcc9# 
(optional=templinst)_ZNSt11_Deque_baseIPN4llvm6RegionESaIS2_EED1Ev@Base 
1.1-6~llvm6.0+gcc8
-#MISSING: 1.3-3~gcc9# 
(optional=templinst)_ZNSt11_Deque_baseIPN4llvm6RegionESaIS2_EED2Ev@Base 
1.1-6~llvm6.0+gcc8
+#MISSING: 1.3-10+hurd.1# 
(optional=templinst)_ZNKSt8_Rb_treeIPN4llvm10BasicBlockES2_St9_IdentityIS2_ESt4lessIS2_ESaIS2_EE4findERKS2_@Base
 1.1-6~llvm6.0+gcc8
+#MISSING: 1.3-10+hurd.1# 
(optional=templinst)_ZNSt11_Deque_baseIPN4llvm6RegionESaIS2_EED1Ev@Base 
1.1-6~llvm6.0+gcc8
+#MISSING: 1.3-10+hurd.1# 
(optional=templinst)_ZNSt11_Deque_baseIPN4llvm6RegionESaIS2_EED2Ev@Base 
1.1-6~llvm6.0+gcc8
 #MISSING: 1.3-5~llvm7# 
(optional=templinst|arch=armel)_ZNSt23_Sp_counted_ptr_inplaceIN4llvm3sys2fs6detail12DirIterStateESaIS4_ELN9__gnu_cxx12_Lock_policyE1EE10_M_destroyEv@Base
 0.14-7~llvm4.0
 #MISSING: 1.3-5~llvm7# 
(optional=templinst|arch=armel)_ZNSt23_Sp_counted_ptr_inplaceIN4llvm3sys2fs6detail12DirIterStateESaIS4_ELN9__gnu_cxx12_Lock_policyE1EE10_M_disposeEv@Base
 0.14-7~llvm4.0
 #MISSING: 1.3-5~llvm7# 
(optional=templinst|arch=armel)_ZNSt23_Sp_counted_ptr_inplaceIN4llvm3sys2fs6detail12DirIterStateESaIS4_ELN9__gnu_cxx12_Lock_policyE1EE14_M_get_deleterERKSt9type_info@Base
 0.14-7~llvm4.0
@@ -458,7 +458,7 @@
  
(optional=templinst)_ZNSt6vectorIPN4llvm10BasicBlockESaIS2_EE17_M_realloc_insertIJS2_EEEvN9__gnu_cxx17__normal_iteratorIPS2_S4_EEDpOT_@Base
 0.13-9~llvm3.8+gcc7
  
(optional=templinst)_ZNSt6vectorIPN4llvm10BasicBlockESaIS2_EE6insertEN9__gnu_cxx17__normal_iteratorIPKS2_S4_EERS7_@Base
 0.10
  
(optional=templinst)_ZNSt6vectorIPN4llvm11InstructionESaIS2_EE17_M_realloc_insertIJRKS2_EEEvN9__gnu_cxx17__normal_iteratorIPS2_S4_EEDpOT_@Base
 0.13-9~llvm3.8+gcc7
-#MISSING: 0.14-4~llvm3.9# 
(optional=templinst)_ZNSt6vectorIPN4llvm12MemIntrinsicESaIS2_EE17_M_realloc_insertIJS2_EEEvN9__gnu_cxx17__normal_iteratorIPS2_S4_EEDpOT_@Base
 0.13-9~llvm3.8+gcc7
+#MISSING: 1.3-10+hurd.1# 
(optional=templinst)_ZNSt6vectorIPN4llvm12MemIntrinsicESaIS2_EE17_M_realloc_insertIJS2_EEEvN9__gnu_cxx17__normal_iteratorIPS2_S4_EEDpOT_@Base
 0.13-9~llvm3.8+gcc7
  
(optional=templinst)_ZNSt6vectorIPN4llvm4TypeESaIS2_EE12emplace_backIJS2_EEEvDpOT_@Base
 1.0
  
(optional=templinst)_ZNSt6vectorIPN4llvm4TypeESaIS2_EE17_M_realloc_insertIJS2_EEEvN9__gnu_cxx17__normal_iteratorIPS2_S4_EEDpOT_@Base
 0.13-9~llvm3.8+gcc7
  
(optional=templinst)_ZNSt6vectorIPN4llvm5ValueESaIS2_EE12emplace_backIJS2_EEEvDpOT_@Base
 0.10
@@ -471,12 +471,12 @@
  
(optional=templinst)_ZNSt6vectorIPN4llvm8FunctionESaIS2_EE17_M_realloc_insertIJRKS2_EEEvN9__gnu_cxx17__normal_iteratorIPS2_S4_EEDpOT_@Base
 0.14
  
(optional=templinst)_ZNSt6vectorIPN4llvm8FunctionESaIS2_EE17_M_realloc_insertIJS2_EEEvN9__gnu_cxx17__normal_iteratorIPS2_S4_EEDpOT_@Base
 1.3
  
(optional=templinst)_ZNSt7__cxx1110_List_baseINS_12basic_stringIcSt11char_traitsIcESaIcEEESaIS5_EE8_M_clearEv@Base
 1.3
-#MISSING: 0.14-7~llvm4.0# 
(optional=templinst)_ZNSt7__cxx1110_List_baseISt4pairIPvSt10unique_ptrIN4llvm6detail21AnalysisResultConceptINS4_8FunctionEEESt14default_deleteIS8_EEESaISC_EE8_M_clearEv@Base
 0.13-3~llvm3.8

Bug#947881: [Pkg-opencl-devel] Bug#947881: pocl: fix build on hurd-i386

2020-01-01 Thread Andreas Beckmann 
On 01/01/2020 15.19, Samuel Thibault wrote:
> pocl currently FTBFS on hurd-i386 only due to a header inclusion
> condition bug, could you apply the attached patch?

Nice. Thanks.
Does the buildlog show any symbols changes?

Andreas