Decouple OPM Flow From Libecl #1107
Conversation
|
jenkins build this please |
|
jenkins build this opm-material=350 opm-grid=397 opm-models=550 opm-simulators=2060 opm-upscaling=281 please |
|
please amend with diff --git a/python/setup.py b/python/setup.py
index 897a73f81..f5790e913 100644
--- a/python/setup.py
+++ b/python/setup.py
@@ -47,7 +47,7 @@ ext_modules = [
'cxx/well.cpp',
'cxx/export.cpp'
],
- libraries=['opmcommon', 'boost_filesystem', 'boost_regex', 'ecl', 'z'],
+ libraries=['opmcommon', 'boost_filesystem', 'boost_regex'],
language='c++',
undef_macros=["NDEBUG"],
include_dirs=["pybind11/include"]Diff edited by @joakim-hove |
Okay, gives me a chance to try GitHub's editor... |
|
jenkins build this opm-material=350 opm-grid=397 opm-models=559 opm-simulators=2060 opm-upscaling=281 please |
|
The build check including changes to the build system goes through—at least when launched from opm-common. I think this is ready for review now. |
|
Thank you for your comments, @joakim-hove. I'll adjust this PR accordingly. |
bska
force-pushed
the
decouple-libecl-final
branch
from
fcfe6c7
to
052fb8a
Compare
|
I have rebased this onto the current master branch and tried to address review comments in the process. I will rerun the build test now. |
|
jenkins build this opm-material=350 opm-grid=397 opm-models=559 opm-simulators=2060 opm-upscaling=281 please |
|
This PR was approved by module maintainer. I intend to merge this (and downstream PRs) no later 4pm CEST today (Wednesday). |
In particular, use EGrid, ERst and EclFile as appropriate.
Instead, switch to 'int' for the 'before' and 'after' row indices. The 'ssize_t' Posix type alias is not appropriate for this usage since its range is only guaranteed to be [ -1 .. (1<<15)-1 ].
Mostly for converting between std::time_t and broken-down time stamps. Uses UTC and std::chrono::system_clock. May wrap in as little as 292 years, depending on the period of system_clock. Intended to replace various timestamping utility functions from libecl. A comprehensive time-service protocol for Flow is much more work than this, and will likely not be easily realized before we have C++17 and its much expanded time/calendar library.
Note that we have to reduce the year-range in the specific test
createDeckWithDRSDTthenDRVDT
in order not to wrap around for system_clock. This is a deficency of
the new time-service protocol.
Mostly just to provide a simple overload of the utility function
ecl_util_make_date
from libecl. The rest of the test code remains intact.
Only affects the 'first_sim()' helper function.
Suggested by [at]akva2.
bska
force-pushed
the
decouple-libecl-final
branch
from
052fb8a
to
288be7f
Compare
|
jenkins build this opm-material=350 opm-grid=397 opm-models=559 opm-simulators=2060 opm-upscaling=281 please |
|
Final pre-merge build check goes through. I'm merging this into master now per prior approval from module maintainer. |
Decouple OPM Flow From Libecl (Backport of pr #1107)
|
Backported to release via #1118 |
This PR adds a very rudimentary time-service protocol in
opm/common/utility. It is just sufficient to replace various time-related utility functions used from libecl. We get a substantial reduction in available time range (down to about 292 years on my Ubuntu system, although this depends onchrono::system_clock's "period".) I have not looked into what we would need to extend the time range. 292 years is typically not a problem for usual reservoir applications, but will be a concern for CCS.We replace the usage of the time-related utility functions mentioned earlier. As of this PR none of the libraries that constitute OPM Flow reference libecl functions or types.