Timeline


and

2022-05-13:

23:37 Changeset in OpenModelica [5b747e9] by Philip Hannebohm <phannebohm@…>
  • 4 edits in /

[janitor] Cleanup events.c

  • make tmpEventList truly static (no malloc)
  • stop returning confusing double from bisection
22:52 Changeset in OpenModelica [1338e35] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

comments added to the code

21:37 Changeset in OpenModelica [520e4d0b]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 3 edits in /

Compile static libs with position independent code (#8954)

  • For the CMake build system set position independent code OpenModelica/ wide instead of just OpenModelica/OMCompiler/
  • For the Makefile build, compile libantlr4 as position independent code.

For the motivation behind this change, see the extended discussion in #8738.

20:23 Changeset in OpenModelica [4c588ed0] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

Some clean-up and correction of multirate calling sequence incl. emit results

19:54 Changeset in OpenModelica [b25d02a4]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 1 edit in OMSimulator

Update OMSimulator (#8944)

16:32 Changeset in OpenModelica [a8f65ec]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 1 edit in OMCompiler/Compiler/runtime/om_curl.c

set curl timeout to 8s (#8952)

16:23 Changeset in OpenModelica [2eb56e2] by Mahder Gebremedhin <mahder.geb@…>
  • 3 edits in /

Compile static libs with position independent code

  • For the CMake build system set position independent code OpenModelica/ wide instead of just OpenModelica/OMCompiler/
  • For the Makefile build, compile libantlr4 as position independent code.

For the motivation behind this change, see the extended discussion in #8738.

16:18 Changeset in OpenModelica [aa8243e0]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 2 edits
    2 adds in /

Dump class comments/annotations in instance API (#8947)

16:14 Changeset in OpenModelica [bedbc3e] by Philip Hannebohm <phannebohm@…>
  • 1 edit in testsuite/openmodelica/dataReconciliation/TSP_Pipe3.mos

Update testsuite

15:55 Changeset in OpenModelica [e8132840] by Henning Kiel <hkiel@…>
  • 1 edit in OMCompiler/Compiler/runtime/om_curl.c

set curl timeout to 8s

15:40 Changeset in OpenModelica [8aa755ac] by bernhardbachmann <bernhard.bachmann@…>
  • 1 edit in OMCompiler/SimulationRuntime/c/simulation/solver/generic_rk_mr.c

Burger workin again

15:39 Changeset in OpenModelica [c79e5191] by AnHeuermann <andreas.heuermann@…>
  • 1 edit in OMCompiler/Compiler/BackEnd/BackendDAEOptimize.mo

Don't replace constants with function calls.

14:16 Changeset in OpenModelica [86c920a9]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 1 edit in OMCompiler/SimulationRuntime/c/util/omc_file.c

Greater equal for MAX_PATH to use long path (#8949)

13:39 Changeset in OpenModelica [b5d94226] by AnHeuermann <andreas.heuermann@…>
  • 1 edit in OMCompiler/SimulationRuntime/c/util/omc_file.c

Greater equal for MAX_PATH to use long path

13:35 Changeset in OpenModelica [a599ab00] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

some clean-up

13:24 Changeset in OpenModelica [2838371]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 1 edit in .CI/cmake/Jenkinsfile.cmake.ubuntu.gcc.test

Fix syntax error, again. (#8946)

13:11 Changeset in OpenModelica [c7ebb16] by kabdelhak <karim.abdelhak@…>
  • 1 edit in OMCompiler/Compiler/BackEnd/BackendEquation.mo

[BE] update nonlinear system handling

  • before: residuals that cannot be created were skipped and were just left out in nonlinear systems
  • now: the function fails to create residuals
  • todo: repair all cases in which we cannot create residuals
12:48 Changeset in OpenModelica [832cf7d1] by Per Östlund <perost86@…>
  • 2 edits
    2 adds in /

Dump class comments/annotations in instance API

11:30 Changeset in OpenModelica [7804a81d] by Lennart Ochel <lennart.ochel@…>
  • 1 edit in OMSimulator

Update OMSimulator

10:10 Changeset in OpenModelica [694be4a4] by Mahder Gebremedhin <mahder.geb@…>
  • 1 edit in .CI/cmake/Jenkinsfile.cmake.ubuntu.gcc.test

Fix syntax error, again.

2022-05-12:

17:05 Changeset in OpenModelica [c2f8344c] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

Bug fix in event handling (use findRoot function)

15:58 Changeset in OpenModelica [ef01ade]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 1 edit in OMCompiler/Compiler/FrontEnd/Inline.mo

[FE] update inline functions (#8942)

  • inlining empty functions returns empty arrays
15:24 Changeset in OpenModelica [fa31907e] by kabdelhak <karim.abdelhak@…>
  • 1 edit in OMCompiler/Compiler/FrontEnd/Inline.mo

[FE] update inline functions

  • inlining empty functions returns empty arrays
14:45 Changeset in OpenModelica [39406937] by Per Östlund <perost86@…>
  • 1 edit in OMCompiler/Compiler/NFFrontEnd/NFTypeCheck.mo

Add type casting for records

13:35 Changeset in OpenModelica [169c016] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

Added Root Finding to the integrator

13:10 Changeset in OpenModelica [132048b3]maintenance/v1.19v1.19.0-dev.beta3 by Adeel Asghar <adeel.asghar@…>
  • 1 edit in libraries

update libraries submodules (#8934)

13:10 Changeset in OpenModelica [36fce63]maintenance/v1.19 by Adeel Asghar <adeel.asghar@…>
  • 1 edit in OMCompiler/Compiler/runtime/systemimpl.c

Ignore build metadata in System.getLoadModelPath (#8897)

  • Ignore build metadata when determining which library to load, since it shouldn't be used to determine version precedence according to semantic versioning. Otherwise we fail to load the library if the user asks for e.g. MSL 3.2.3+maint.om but has MSL 3.2.3 installed.
12:41 Changeset in OpenModelica [07cc4f30]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by Mahder Gebremedhin <mahge@…>
  • 1 edit in .CI/cmake/Jenkinsfile.cmake.ubuntu.gcc.test

Fix syntax error.

12:33 Changeset in OpenModelica [45cb1317] by Adeel Asghar <adeel.asghar@…>
  • 1 edit in libraries

update libraries submodules (#8934)

12:32 Changeset in OpenModelica [76b54d61] by Adeel Asghar <adeel.asghar@…>
  • 1 edit in OMCompiler/Compiler/runtime/systemimpl.c

Ignore build metadata in System.getLoadModelPath (#8897)

  • Ignore build metadata when determining which library to load, since it shouldn't be used to determine version precedence according to semantic versioning. Otherwise we fail to load the library if the user asks for e.g. MSL 3.2.3+maint.om but has MSL 3.2.3 installed.
12:10 Changeset in OpenModelica [8774fc2] by Mahder Gebremedhin <mahder.geb@…>
  • 1 edit in .CI/cmake/Jenkinsfile.cmake.ubuntu.gcc.test

Fix syntax error.

11:49 Changeset in OpenModelica [3533d33] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

Fixed interpolation problem again, Burger equation is running again

10:30 Changeset in OpenModelica [e259d16] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

problems with Interpolation, try to come back to this version

2022-05-11:

23:42 Changeset in OpenModelica [0e39c773]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 36 edits
    23 adds
    1 delete in /

Add checks for some specialized class restrictions (#8930)

23:09 Changeset in OpenModelica [6980ca00] by Per Östlund <perost86@…>
  • 36 edits
    23 adds
    1 delete in /

Add checks for some specialized class restrictions

20:25 Changeset in OpenModelica [fc054030]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 1 edit in libraries

update libraries submodules (#8934)

19:27 Changeset in OpenModelica [7001daf] by Adrian Pop <adrian.pop@…>
  • 1 edit in libraries

update libraries submodules

19:22 Changeset in OpenModelica [7227538]maintenance/v1.20maintenance/v1.21maintenance/v1.22 by GitHub <noreply@…>
  • 2 edits in OMCompiler/Compiler

[BE] creating array parameter bindings correctly (#8933)

  • correctly creates array parameter bindings during initialization
19:07 Changeset in OpenModelica [eca07f31]omlib-staging by hudson <openmodelica@…>
  • 1 edit in libraries

Updated libraries

  • libraries 6990a88...1413a34 (1):

    update Makefiles.libs

18:51 Changeset in OpenModelica [e6de4bea] by kabdelhak <karim.abdelhak@…>
  • 2 edits in OMCompiler/Compiler

[BE] creating array parameter bindings correctly

  • correctly creates array parameter bindings during initialization
  • fixes ticket #8772
16:51 Changeset in OpenModelica [080c3974] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

improved initial start values for NLS

14:40 Changeset in OpenModelica [62c78e68] by bernhardbachmann <bernhard.bachmann@…>
  • 5 edits in /

Falg for percentage added

14:20 Changeset in OpenModelica [e37efd4f]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit
    2 adds in .CI

Add CI jobs for extra CMake related builds. (#8932)

  • A new job is added that builds OpenModelica on Windows with OMDev and gcc using CMake.
  • A new job is added that builds OpenModelica on Ubuntu with gcc and then runs the testsuite with it. For this job a normal (autoconf) build is also performed to build some things that the CMake build does not build yet, e.g. OMSI runtimes. The relevant files from the autoconf build are then stashed and used together with the stash from the CMake build.
13:43 Changeset in OpenModelica [9c2124a] by Mahder Gebremedhin <mahder.geb@…>
  • 1 edit
    2 adds in .CI

Add CI jobs for extra CMake related builds.

  • A new job is added that builds OpenModelica on Windows with OMDev and gcc using CMake.
  • A new job is added that builds OpenModelica on Ubuntu with gcc and then runs the testsuite with it. For this job a normal (autoconf) build is also performed to build some things that the CMake build does not build yet, e.g. OMSI runtimes. The relevant files from the autoconf build are then stashed and used together with the stash from the CMake build.
13:15 Changeset in OpenModelica [78818a8]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit
    1 delete in /

Disable compliance-newinst job (#8931)

  • Disable the compliance-newinst job, since it's now identical to the regular compliance job.
12:40 Changeset in OpenModelica [fdd116b] by Per Östlund <perost86@…>
  • 1 edit
    1 delete in /

Disable compliance-newinst job

  • Disable the compliance-newinst job, since it's now identical to the regular compliance job.
11:30 Changeset in OpenModelica [a2596f1c]maintenance/v1.19 by GitHub <noreply@…>
  • 5 edits in OMEdit/OMEditLIB

Improved settings of MODELICAPATH (#8927) (#8929)

Do not save the MODELICAPATH in settings until user explicitly modifies it. Otherwise we had issues when the OM version is changed.

11:18 Changeset in OpenModelica [7c1d44bc] by Philip Hannebohm <phannebohm@…>
  • 3 edits in OMCompiler/Compiler

move to old cpp runtime

11:15 Changeset in OpenModelica [2b778e1] by Philip Hannebohm <phannebohm@…>
  • 1 edit in testsuite/openmodelica/cppruntime/testVectorizedBlocks.mos

update testsuite

11:15 Changeset in OpenModelica [d5735762] by Philip Hannebohm <phannebohm@…>
  • 1 edit in OMCompiler/Compiler/Template/CodegenFMU.tpl

[Codegen] scalarize variables for FMU

11:15 Changeset in OpenModelica [ffec91f] by Philip Hannebohm <phannebohm@…>
  • 1 edit in OMCompiler/Compiler/BackEnd/RemoveSimpleEquations.mo

[BE] keep arrays when removing simple equations

11:15 Changeset in OpenModelica [7116593] by Philip Hannebohm <phannebohm@…>
  • 2 edits in OMCompiler/Compiler/Template

add vector mapping on array assignment

11:15 Changeset in OpenModelica [cbf7beed] by Philip Hannebohm <phannebohm@…>
  • 1 edit in OMCompiler/Compiler/SimCode/SimCodeUtil.mo

stop recursing

11:15 Changeset in OpenModelica [a947343] by Philip Hannebohm <phannebohm@…>
  • 11 edits
    2 adds in /

add for-equation support in symbolic jacobian

should make for faster symbolic manipulations and smaller target code.

  • needs -d=-nfScalarize to take effect
  • works only for one dimensional arrays and only for target cpp yet
  • in codegen: use #define to map cref names onto jacobian variables
  • in SimCode: increment simvar index not by 1 but by dimension size
11:11 Changeset in OpenModelica [9b000e6] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

improved initial start values for NLS

11:06 Changeset in OpenModelica [ab6291a7] by Adeel Asghar <adeel.asghar@…>
  • 5 edits in OMEdit/OMEditLIB

Improved settings of MODELICAPATH (#8927)

Do not save the MODELICAPATH in settings until user explicitly modifies it. Otherwise we had issues when the OM version is changed.

11:04 Changeset in OpenModelica [489bbd3]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 5 edits in OMEdit/OMEditLIB

Improved settings of MODELICAPATH (#8927)

Do not save the MODELICAPATH in settings until user explicitly modifies it. Otherwise we had issues when the OM version is changed.

11:02 Changeset in OpenModelica [c1f7c952]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 29 edits in /

Sanity check for sparsity pattern (#8907)

  • Check if the sparsity pattern can represent a regular matrix (no zero rows or columns, at least n non-zero elements)
  • Use dense numeric Jacobian, disable Jacobian scaling and issue a warning when sanity check failed.

Co-authored-by: Philip Hannebohm <phannebohm@…>

10:35 Changeset in OpenModelica [9911ac78] by Adeel Asghar <adeel.asghar@…>
  • 5 edits in OMEdit/OMEditLIB

Improved settings of MODELICAPATH

Do not save the MODELICAPATH in settings until user explicitly modifies it. Otherwise we had issues when the OM version is changed.

10:26 Changeset in OpenModelica [5ba9fd07] by Philip Hannebohm <phannebohm@…>
  • 73 edits in /

Only test if pattern is there

10:14 Changeset in OpenModelica [d7ab8540]maintenance/v1.19 by GitHub <noreply@…>
  • 2 edits in OMCompiler/Compiler/Script

Reset Global.packageIndexCacheIndex when ModelicaPath is changed (#8898) (#8926)

The package manager use the path that ends with / so when checking it against ModelicaPath then check with and without /

10:13 Changeset in OpenModelica [1abfdbd2] by bernhardbachmann <bernhard.bachmann@…>
  • 3 edits in OMCompiler/SimulationRuntime/c/simulation/solver

checkForEvents added to the integrators

08:57 Changeset in OpenModelica [4b57280f] by Adeel Asghar <adeel.asghar@…>
  • 2 edits in OMCompiler/Compiler/Script

Reset Global.packageIndexCacheIndex when ModelicaPath is changed (#8898)

The package manager use the path that ends with / so when checking it against ModelicaPath then check with and without /

2022-05-10:

23:00 Changeset in OpenModelica [ae85cd9] by bernhardbachmann <bernhard.bachmann@…>
  • 2 edits in OMCompiler/SimulationRuntime/c/simulation/solver

Found work-around for event handling

22:13 Changeset in OpenModelica [f11277c4]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by Mahder Gebremedhin <mahge@…>
  • 4 edits in OMCompiler/SimulationRuntime/c/util

Use winsock2.h instead of windows.h.

  • winsock2.h and winsock.h can not be used simultaneously. Which is okay since winsock2.h should provide everything from winsock.h in a backward compatible way. So far so good except for the fact that windows.h includes winsock.h by default. This can be avoided by defining WIN32_LEAN_AND_MEAN before including it everywhere.

It can also be avoided by not including windows.h att all and including
winsock2.h instead which will include windows.h. We have gone for
the latter option for our code which can possibly include winsock2.h
at a latter point for actual socket related purposes.

This is all kind of messy but it should be good enough for now until
someone incldues windows.h and suddenly gets duplicate declration errors.

22:13 Changeset in OpenModelica [df7454a1]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by Mahder Gebremedhin <mahge@…>
  • 1 edit in OMCompiler/SimulationRuntime/c/cmake_3.14.cmake

Link SimulationRuntimeC with Lapack explicitly.

  • It was including the headers from 3rdParty/dgesv and linking to lapack transitively from sunlinsollapackdense. Remove the dependency on 3rdParty/dgesv and use the package Lapack only.

Note: 3rdParty/dgesv is not meant to be used by any library in
OpenModelica itself!
It is there for FMUs only. It is there to be packaged with source
code FMUs and provides only the functions that are needed by our FMUs.

20:54 Changeset in OpenModelica [22e71a92]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit in libraries

Get rid of the warning when loading MSL3.2.3 #8882 (#8924)

Update submodule libraries

  • update MSL3.2.3 and MSL4.0.x hashes to the latest
  • this get us ModelicaServices that has no uses annotations
20:17 Changeset in OpenModelica [40f8375d] by Philip Hannebohm <phannebohm@…>
  • 410 edits in /

Update testsuite

20:03 Changeset in OpenModelica [bfc35dbf] by Mahder Gebremedhin <mahder.geb@…>
  • 4 edits in OMCompiler/SimulationRuntime/c/util

Use winsock2.h instead of windows.h.

  • winsock2.h and winsock.h can not be used simultaneously. Which is okay since winsock2.h should provide everything from winsock.h in a backward compatible way. So far so good except for the fact that windows.h includes winsock.h by default. This can be avoided by defining WIN32_LEAN_AND_MEAN before including it everywhere.

It can also be avoided by not including windows.h att all and including
winsock2.h instead which will include windows.h. We have gone for
the latter option for our code which can possibly include winsock2.h
at a latter point for actual socket related purposes.

This is all kind of messy but it should be good enough for now until
someone incldues windows.h and suddenly gets duplicate declration errors.

20:03 Changeset in OpenModelica [7104831] by Mahder Gebremedhin <mahder.geb@…>
  • 1 edit in OMCompiler/SimulationRuntime/c/cmake_3.14.cmake

Link SimulationRuntimeC with Lapack explicitly.

  • It was including the headers from 3rdParty/dgesv and linking to lapack transitively from sunlinsollapackdense. Remove the dependency on 3rdParty/dgesv and use the package Lapack only.

Note: 3rdParty/dgesv is not meant to be used by any library in
OpenModelica itself!
It is there for FMUs only. It is there to be packaged with source
code FMUs and provides only the functions that are needed by our FMUs.

19:59 Changeset in OpenModelica [5dfe6db] by Adrian Pop <adrian.pop@…>
  • 1 edit in libraries

Get rid of the warning when loading MSL3.2.3 #8882
Update submodule libraries

  • update MSL3.2.3 and MSL4.0.x hashes to the latest
  • this get us ModelicaServices that has no uses annotations
19:44 Changeset in OpenModelica [9d480f73]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit in OMCompiler/3rdParty

Enable sunlinsollapackdense even with no Fortran support. (#8923)

  • The only reason sundails wants to have Fortran support to enable sunlinsollapackdense is that it wants to figure out Fortran the name mangling scheme so that it can use the Lapack libraries needed.

As it happens we can actually manually specify the name mangling scheme
to avoid the requirement for a Fortran compiler.

Thank you Sundials for being all around very well designed.

19:15 Changeset in OpenModelica [cfb41997] by Mahder Gebremedhin <mahder.geb@…>
  • 1 edit in OMCompiler/3rdParty

Enable sunlinsollapackdense even with no Fortran support.

  • The only reason sundails wants to have Fortran support to enable sunlinsollapackdense is that it wants to figure out Fortran the name mangling scheme so that it can use the Lapack libraries needed.

As it happens we can actually manually specify the name mangling scheme
to avoid the requirement for a Fortran compiler.

Thank you Sundails for being all around very well designed.

18:56 Changeset in OpenModelica [50a1946]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit in OMCompiler/Compiler/FrontEnd/ClassLoader.mo

Revert #8921 (#8922)

18:28 Changeset in OpenModelica [3d8e982] by Per Östlund <perost86@…>
  • 1 edit in OMCompiler/Compiler/FrontEnd/ClassLoader.mo

Revert #8921

18:15 Changeset in OpenModelica [62c4f136] by bernhardbachmann <bernhard.bachmann@…>
  • 4 edits in /

many bug-fixes, but still not good enough

17:41 Changeset in OpenModelica [f441a33]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit in OMCompiler/Compiler/runtime/systemimpl.c

Ignore build metadata in System.getLoadModelPath (#8897)

  • Ignore build metadata when determining which library to load, since it shouldn't be used to determine version precedence according to semantic versioning. Otherwise we fail to load the library if the user asks for e.g. MSL 3.2.3+maint.om but has MSL 3.2.3 installed.
17:41 Changeset in OpenModelica [fd6693ae]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit in OMCompiler/Compiler/FrontEnd/ClassLoader.mo

Improve library loading (#8921)

  • Move exactly matching versions to the head of the list when using the package manager to get compatible versions, to avoid loading other versions of libraries when the requested version exists.
17:06 Changeset in OpenModelica [b51676d] by Per Östlund <perost86@…>
  • 1 edit in OMCompiler/Compiler/FrontEnd/ClassLoader.mo

Improve library loading

  • Move exactly matching versions to the head of the list when using the package manager to get compatible versions, to avoid loading other versions of libraries when the requested version exists.
13:49 Changeset in OpenModelica [b34f6bbf] by AnHeuermann <andreas.heuermann@…>
  • 29 edits
    8 adds in /

Merge remote-tracking branch 'origin/master' into sparsity-sanity-check

13:05 Changeset in OpenModelica [6cbe437]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 26 edits in /

Use long path in omc_file functions on Windows (#8848)

  • Make omc_file compatible with long paths on Windows
  • Use omc_file functions in Compiler and C runtime
13:04 Changeset in OpenModelica [9cc8609]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 4 edits
    1 add in /

Fix the check for outer elements in functions (#8919)

  • Make sure we don't fail due to not instantiated outer components before we get to the error check.
12:35 Changeset in OpenModelica [0744f5a7]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 4 edits
    1 add in /

Add an interface PThreads library for readability (#8917)

  • A new library OMCPThreads provides and interface to a PThreads implementation on Linux as well as on Windows (if one is available)

This is to avoid having to do repeated checks for MSVC, Linux ...
in cases where we always want to use PThreads, e.g. 3rdParty/GC and
our runtime libraries want to use only PThreads and not Win32 Threads.

This allows us to have cleaner setup (I think?)

12:25 Changeset in OpenModelica [925c048] by Per Östlund <perost86@…>
  • 4 edits
    1 add in /

Fix the check for outer elements in functions

  • Make sure we don't fail due to not instantiated outer components before we get to the error check.
12:16 Changeset in OpenModelica [8584315c]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit in OMCompiler/Compiler/Util/UnorderedSet.mo

Rename UnorderedSet.addNoUpdCheck to addNew (#8918)

  • Rename UnorderedSet.addNoUpdCheck to addNew in order to use the same naming convention as UnorderedMap.
11:41 Changeset in OpenModelica [1b1158f] by Per Östlund <perost86@…>
  • 1 edit in OMCompiler/Compiler/Util/UnorderedSet.mo

Rename UnorderedSet.addNoUpdCheck to addNew

  • Rename UnorderedSet.addNoUpdCheck to addNew in order to use the same naming convention as UnorderedMap.
11:02 Changeset in OpenModelica [3e62499] by AnHeuermann <andreas.heuermann@…>
  • 2 edits in OMCompiler

Using omc_fopen instead of fopen

10:49 Changeset in OpenModelica [f0135c8] by Mahder Gebremedhin <mahder.geb@…>
  • 4 edits
    1 add in /

Add an interface PThreads library for readability

  • A new library OMCPThreads provides and interface to a PThreads implementation on Linux as well as on Windows (if one is available)

This is to avoid having to do repeated checks for MSVC, Linux ...
in cases where we always want to use PThreads, e.g. 3rdParty/GC and
our runtime libraries want to use only PThreads and not Win32 Threads.

This allows us to have cleaner setup (I think?)

2022-05-09:

18:33 Changeset in OpenModelica [8529da7]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 8 edits in OMCompiler

Make Ipopt an optional dependency (#8916)

  • This only works for the CMake build right now. As in you can only disable Ipopt successfully link everything if you are using the CMake build. For the autoconf+Makefiles build we need to update the Makefiles a bit more to tell them when Ipopt is not available. However, there is not much of a need for that now. The main reason for disabling Ipopt is systems that do not have a Fortran compiler readily available. On those systems we do not plan to use the autoconf build anyway.
  • The macros and configure variables have been renamed to be more unique.
18:04 Changeset in OpenModelica [d0df3132] by Mahder Gebremedhin <mahder.geb@…>
  • 1 edit in OMCompiler/3rdParty

Update 3rdParty.

17:59 Changeset in OpenModelica [d420345a] by Mahder Gebremedhin <mahder.geb@…>
  • 7 edits in OMCompiler

Make Ipopt an optional dependency.

  • This only works for the CMake build right now. As in you can only disable Ipopt successfully link everything if you are using the CMake build. For the autoconf+Makefiles build we need to update the Makefiles a bit more to tell them when Ipopt is not available. However, there is not much of a need for that now. The main reason for disabling Ipopt is systems that do not have a Fortran compiler readily available. On those systems we do not plan to use the autoconf build anyway.
  • The macros and configure variables have been renamed to be more unique.
17:48 Changeset in OpenModelica [f92dd685]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 5 edits in OMCompiler

Memory leak for static LS/NLS data (#8915)

  • Only update min, max, nominal in updateStaticDataOfNonlinearSystems / updateStaticDataOfLinearSystems
17:22 Changeset in OpenModelica [647c905] by AnHeuermann <andreas.heuermann@…>
  • 4 edits in OMCompiler

Only update min, max, nominal in updateStaticDataOfNonlinearSystems / updateStaticDataOfLinearSystems

17:12 Changeset in OpenModelica [7e8b5a4b] by Philip Hannebohm <phannebohm@…>
  • 3 edits in OMCompiler/SimulationRuntime/c/simulation/solver

Perform zero check on rows and columns

16:25 Changeset in OpenModelica [e3ade60] by AnHeuermann <andreas.heuermann@…>
  • 3 edits in OMCompiler/SimulationRuntime/c/simulation/solver

Memroy leak for static LS/NLS data solved

15:52 Changeset in OpenModelica [a20fdfdf]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 3 edits in OMCompiler/Compiler

Minor fixes and improvements. (#8914)

  • Tell multi-config generator not to add the config specific folders, e.g., Debug/, Release/ ... inside the RUNTIME_OUTPUT_DIRECTORY we have specified for the bootstrapped omc.
  • Disable some warnings on libOpenModelicaCompiler with MSVC. These warning should be fixed eventually by fixing C code generation.
  • Include dirent.h with MSVC as well. We have ways of getting the header. It is just a lone one file implementation. It can be dropped in.

Hiding it in a !defined(_MSC_VER) macro is pointless since the code
will not compile without it anyway.

15:20 Changeset in OpenModelica [815b3b20]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 3 edits in OMCompiler/Compiler

[BE] update evalFunction (#8913)

  • updates the evaluation of function bodies to remove empty outputs (wild cards)
    • allows correct detection of unability to differentiate functions
    • fixes ticket #8381
15:10 Changeset in OpenModelica [5633fd5] by Mahder Gebremedhin <mahder.geb@…>
  • 3 edits in OMCompiler/Compiler

Minor fixes and improvements.

  • Tell multi-config generator not to add the config specific folders, e.g., Debug/, Release/ ... inside the RUNTIME_OUTPUT_DIRECTORY we have specified for the bootstrapped omc.
  • Disable some warnings on libOpenModelicaCompiler with MSVC. These warning should be fixed eventually by fixing C code generatation.
  • Include dirent.h with MSVC as well. We have ways of getting the header. It is just a lone one file implementation. It can be dropped in.

Hiding it in a !defined(_MSC_VER) macro is pointless since the code
will not compile without it anyway.

14:45 Changeset in OpenModelica [ea75048] by AnHeuermann <andreas.heuermann@…>
  • 39 edits in testsuite

Updating testsuite 2nd try

14:41 Changeset in OpenModelica [e418c61f] by kabdelhak <karim.abdelhak@…>
  • 3 edits in OMCompiler/Compiler

[BE] update evalFunction

  • updates the evaluation of function bodies to remove empty outputs (wild cards)
  • allows correct detection of unability to differentiate functions
  • fixes ticket #8381
13:54 Changeset in OpenModelica [c337a5b0] by AnHeuermann <andreas.heuermann@…>
  • 367 edits in testsuite

updating testsuite 1st try

12:49 Changeset in OpenModelica [56f7f2b]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit in OMCompiler/3rdParty

Improve 3rdParty/GC configuration. (#8910)

  • Use target_compile_definitions(...) instead of just add_definitions(). This allows us to add defines in a way that can apply transitively on targets that link to omcgc. For example, GC_DLL or GC_NOT_DLL need to be defined accordingly on anything that links to omcgc and includes the header gc.h. Otherwise we will have inconsistencies on the attributes used to declare functions when GC is built versus when GC is used.
  • It is probably a good idea to fix all the other defines this way. We just need to figure out which ones need to actually be PUBLIC (transitive) and which ones are only needed to compiling omcgc.
  • Disable build of GC's cord library. We do not use it.
12:43 Changeset in OpenModelica [2484be92]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 2 edits
    6 adds in /

Improve typing of min/max (#8911)

  • Restrict the argument type of min/max to only the types allowed by the specification.
  • Improve the error message for wrong arguments to min/max to better reflect the allowed types.
12:14 Changeset in OpenModelica [3371c90] by Per Östlund <perost86@…>
  • 2 edits
    6 adds in /

Improve typing of min/max

  • Restrict the argument type of min/max to only the types allowed by the specification.
  • Improve the error message for wrong arguments to min/max to better reflect the allowed types.
11:54 Changeset in OpenModelica [ab648c2]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 2 edits
    2 adds in /

Add simplification rules for min({})/max({}) (#8909)

11:51 Changeset in OpenModelica [01769016]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 2 edits in OMParser/3rdParty/antlr4/runtime/Cpp/runtime

Improve OMParser antlr4 configuration. (#8908)

  • Add transitive include directories for the shared version of libantlr4

They were added for the static version and overlooked for the shared
version.

  • Remove unnecessary OUTPUT_DIRECOTRY specifiers on both shared and static libraries. There is no needed to have them as far as I can see. It is confusing to output libraries in unexpected locations.
  • Remove attributes on a forward declaration of a class. The class seems to already be defined at that forward declaration point (the header where the class is defined in is probably included indirectly). Avoids repeated warnings about it.
11:28 Changeset in OpenModelica [7e593ddc] by Mahder Gebremedhin <mahder.geb@…>
  • 1 edit in OMCompiler/3rdParty

Improve 3rdParty/GC configuration.

  • Use target_compile_definitions(...) instead of just add_definitions(). This allows us to add defines in a way that can apply transitively on targets that link to omcgc. For example, GC_DLL or GC_NOT_DLL need to be defined accordingly on anything that links to omcgc and includes the header gc.h. Otherwise we will have inconsistencies on the attributes used to declare functions when GC is built versus when GC is used.
  • It is probably a good idea to fix all the other defines this way. We just need to figure out which ones need to actually be PUBLIC (transitive) and which ones are only needed to compiling omcgc.
  • Disable build of GC's cord library. We do not use it.
11:19 Changeset in OpenModelica [4ba4bd7] by Per Östlund <perost86@…>
  • 2 edits
    2 adds in /

Add simplification rules for min({})/max({})

11:18 Changeset in OpenModelica [10ec3ac] by Mahder Gebremedhin <mahder.geb@…>
  • 2 edits in OMParser/3rdParty/antlr4/runtime/Cpp/runtime

Improve OMParser antlr4 configuration.

  • Add transitive include directories for the shared version of libantlr4 They were added for the static version and overlooked for the shared version.
  • Remove unnecessary OUTPUT_DIRECOTRY specifiers on both shared and static libraries. There is no needed to have them as far as I can see. It is confusing to output libraries in unexpected locations.
  • Remove attributes on a forward declration of a class. The class seems to already be defined at that forward declration point (the header where the class is defined in is probably included indirectly). Avoids repeated warnings about it.
10:16 Changeset in OpenModelica [eb1d468] by AnHeuermann <andreas.heuermann@…>
  • 1 edit in OMCompiler/SimulationRuntime/c/simulation/solver/nonlinearSystem.c

Updating test message

09:55 Changeset in OpenModelica [51be51d] by AnHeuermann <andreas.heuermann@…>
  • 9 edits in OMCompiler/SimulationRuntime

No omc_fopen in OMSI and ModelicaExternalC

2022-05-06:

18:27 Changeset in OpenModelica [9ccf941]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 1 edit in OMCompiler/Compiler/boot/bootstrap-sources/build/Static.c

Remove mistakenly added debug printout (#8905)

18:27 Changeset in OpenModelica [6c013e7f] by AnHeuermann <andreas.heuermann@…>
  • 5 edits in OMCompiler/SimulationRuntime/c/simulation/solver

Sanity check for sparsity pattern

18:24 Changeset in OpenModelica [8bb4472]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by GitHub <noreply@…>
  • 3 edits in /

Evaluate min/max of empty arrays (#8904)

17:58 Changeset in OpenModelica [699e4e05] by Per Östlund <perost86@…>
  • 1 edit in OMCompiler/Compiler/boot/bootstrap-sources/build/Static.c

Remove mistakenly added debug printout

17:51 Changeset in OpenModelica [13f52452] by Per Östlund <perost86@…>
  • 3 edits in /

Evaluate min/max of empty arrays

17:38 Changeset in OpenModelica [5d89d50b]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by Mahder Gebremedhin <mahge@…>
  • 1 edit in OMCompiler/3rdParty

Minor fixes for portability.

  • Sundails: Do not enable sundails Fortran if OpenModelica has disabled Fortran.
  • FMIL: Link to the correct zlib instance in FMIL's minizip compilation.

For the autoconf based OpenModelica build the location of zlib library
is passed through the variable OMC_ZLIB_LIBRARY.

For the CMake based OpenModelica build this variable was empty and
minizip was not actually linking to zlib. minizip is built as a static
library so the linking has no impact in its own compilation excpet for
the fact that it was including the system's zlib headers because those
are the ones it finds first if it is not linked to our zlib (which
transitively provides the headers).

  • LibFFi: Only add the dll direction markups if we are compiling it as a dll, i.e., FFI_BUILDING_DLL is defined.
  • Metis/GKLib: Remove conflicting definition rint(x). No need to define it since MSVC has it now.
17:38 Changeset in OpenModelica [bd1fbfe6]maintenance/v1.20maintenance/v1.21maintenance/v1.22omlib-staging by Mahder Gebremedhin <mahge@…>
  • 1 edit in OMCompiler/3rdParty

Improve GC compilation portability.

  • Fix the configuration so that it can be compiled fine on MSVC. It needs special treatment since we want it to use PThreads on all platoforms including Windows with MSVC.
  • On Windows with MSVC it now expects to have the PThreads4W library.
  • Built-in atomic intrinsics are not available with MSVC. Make sure the configuration reflects that for MSVC.
  • Cover cases where it was only expecting MinGW on Windows and not MSVC.
15:43 Changeset in OpenModelica [40d9168] by Mahder Gebremedhin <mahder.geb@…>
  • 1 edit in OMCompiler/3rdParty

Minor fixes for portability.

  • Sundails: Do not enable sundails Fortran if OpenModelica has disabled Fortran.
  • FMIL: Link to the correct zlib instance in FMIL's minizip compilation.

For the autoconf based OpenModelica build the location of zlib library
is passed through the variable OMC_ZLIB_LIBRARY.

For the CMake based OpenModelica build this variable was empty and
minizip was not actually linking to zlib. minizip is built as a static
library so the linking has no impact in its own compilation excpet for
the fact that it was including the system's zlib headers because those
are the ones it finds first if it is not linked to our zlib (which
transitively provides the headers).

  • LibFFi: Only add the dll direction markups if we are compiling it as a dll, i.e., FFI_BUILDING_DLL is defined.
  • Metis/GKLib: Remove conflicting definition rint(x). No need to define it since MSVC has it now.
15:41 Changeset in OpenModelica [d58658f9] by Mahder Gebremedhin <mahder.geb@…>
  • 1 edit in OMCompiler/3rdParty

Improve GC compilation portability.

  • Fix the configuration so that it can be compiled fine on MSVC. It needs special treatment since we want it to use PThreads on all platoforms including Windows with MSVC.
  • On Windows with MSVC it now expects to have the PThreads4W library.
  • Built-in atomic intrinsics are not available with MSVC. Make sure the configuration reflects that for MSVC.
  • Cover cases where it was only expecting MinGW on Windows and not MSVC.
07:37 Changeset in OMSimulator [0fbbb17] by arun3688 <rain100falls@…>
  • in /

trigger build

Note: See TracTimeline for information about the timeline view.