Failing to build FesAPI v2.0.0.0

[coderAkshay]

I’m trying to build FESAPI v2.0.0.0 using CMake 4.1.0 and Visual Studio 2022. I have the following dependencies installed:

Boost v1.89.0 (including uuid, system, regex)
HDF5 v1.12.0
zlib and minizip
SWIG (for .NET wrapper generation)
However, I’m encountering multiple issues during the CMake configuration, mostly related to Boost — such as missing components and policy warnings (CMP0167).

Can you guide me through the correct setup and configuration steps to resolve these issues and successfully build FESAPI with all dependencies?

[philippeVerney]

Hello,

The schreenshot does not show the CMake configure step but the VS build step. Is your issue really at CMake configuration step? If yes, please provide these issues, not the build ones.

From a more general point of view, the guidelines to build FESAPI are in the README file : https://github.com/F2I-Consulting/fesapi/blob/v2.0.0.0/README.md

Last but not least. FESAPI 2.0.0.0 is now pretty old. If you can, think about building the latest version instead.

[coderAkshay]

Hello, PFB CMake message stacks:

CMake Warning at CMakeLists.txt:56 (MESSAGE):
The HDF5 library (HDF5_C_LIBRARY_DEBUG variable) does not look to be a
valid file. Please modify it.

CMake Warning at CMakeLists.txt:75 (MESSAGE):
The zlib library (ZLIB_LIBRARY_DEBUG variable) does not look to be a valid
file. Please modify it.

CMake Warning at CMakeLists.txt:90 (MESSAGE):
The Minizip library (MINIZIP_LIBRARY_DEBUG variable) does not look to be a
valid file. Please modify it.

Performing Test CAN_USE_STD_REGEX
Performing Test CAN_USE_STD_REGEX - Success
Generating SWIG C# files...
Language subdirectory: csharp
Search paths:
.
.\swig_lib\csharp
C:\Users\AMahajan11\Desktop\fesapi_env\SWIG\swigwin-4.3.1\Lib\csharp
.\swig_lib
C:\Users\AMahajan11\Desktop\fesapi_env\SWIG\swigwin-4.3.1\Lib
Preprocessing...
Starting language-specific parse...

Processing types...
C++ analysis...
Processing nested classes...
Generating wrappers...

C:\Users\AMahajan11\Desktop\fesapi_env\fesapi\fesapi-2.1.0.0\swig\swigResqml2Include.i(507) : Warning 314: 'byte' is a C# keyword, renaming to 'byte_'
C:\Users\AMahajan11\Desktop\fesapi_env\fesapi\fesapi-2.1.0.0\swig\swigResqml2Include.i(913) : Warning 314: 'in' is a C# keyword, renaming to 'in_'
C:\Users\AMahajan11\Desktop\fesapi_env\fesapi\fesapi-2.1.0.0\swig\swigResqml2Include.i(1800) : Warning 314: 'in' is a C# keyword, renaming to 'in_'
C:\Users\AMahajan11\Desktop\fesapi_env\fesapi\fesapi-2.1.0.0\swig\swigResqml2_0_1Include.i(105) : Warning 314: 'string' is a C# keyword, renaming to 'string_'
C:\Users\AMahajan11\Desktop\fesapi_env\fesapi\fesapi-2.1.0.0\swig\swigWitsml2Include.i(308) : Warning 314: 'null' is a C# keyword, renaming to 'null_'
C:\Users\AMahajan11\Desktop\fesapi_env\fesapi\fesapi-2.1.0.0\swig\swigWitsml2Include.i(310) : Warning 314: 'string' is a C# keyword, renaming to 'string_'
C:\Users\AMahajan11\Desktop\fesapi_env\fesapi\fesapi-2.1.0.0\swig\swigWitsml2Include.i(711) : Warning 314:
'byte' is a C# keyword, renaming to 'byte_'
C:\Users\AMahajan11\Desktop\fesapi_env\fesapi\fesapi-2.1.0.0\swig\swigWitsml2Include.i(1117) : Warning 314: 'in' is a C# keyword, renaming to 'in_'
C:\Users\AMahajan11\Desktop\fesapi_env\fesapi\fesapi-2.1.0.0\swig\swigWitsml2Include.i(1999) : Warning 314: 'in' is a C# keyword, renaming to 'in_'

SWIG C# files have been generated.
Configuring done (10.6s)
Generating done (0.1s)

seems like boost is not getting linked.

Note: If I use FesAPI (2.13.0.0) along with HDF (1.12.0) and other dependencies my project is getting build successfully. However, I need to build 2.1.0.0 as per our business needs

[philippeVerney]

I don't see any relevant errors at the cmake configure step.
Looking at line 336 of AbstractObject.cpp which is the only FESAPI line in error in your screenshot, I can find a call to boost :

fesapi/src/common/AbstractObject.cpp

Line 336 in 2872026

const std::string uuidStr = boost::uuids::to_string(gen());

I remember that Boost 1.86 includes significant internal changes to Boost.Uuid. Could you try with a 1.85 or less version of boost please?

[coderAkshay]

Thanks for your support. I was able to successfully build the library after downgrading Boost to version 1.81.

In my project, I had previously used FESAPI version 0.16.0.1, and now I'm working on resolving compilation issues after upgrading to version 2.0.0.0. One issue I'm facing is that I can no longer find the enum resqml2__MdReference, which I had used in the older version. Could you help me understand what happened to it or what the equivalent is in the new version?

[philippeVerney]

MdDatum I guess is the replacement :

fesapi/src/common/DataObjectRepository.h

Line 1043 in 2872026

DLL_IMPORT_OR_EXPORT RESQML2_NS::MdDatum* createMdDatum(const std::string & guid, const std::string & title,

From a general point of view, look at the stereotype of the new function and see what is the new parameter datatype required.

Moving from version 0 to 2 is quite hard. This is generally done by incremental more little upgrade or by paying a dedicated support.

[coderAkshay]

Hi, thanks for all the support. I am using FesApi (2.0.0.0) and getting below status while epcFile.deserializeInto(dataObjectRepository,DataObjectRepository.openingMode.READ_WRITE), can you please tell me what is wrong ?

The continuous property fdd1274b-c67d-47c4-86e2-29e954b36ab4 cannot be associated to a resqml property kind "permeability rock" which is abstract. This property will be assumed to be a partial one.
The continuous property e7a574a6-fabe-494e-868f-29e9eaf797f9 cannot be associated to a resqml property kind "dimensionless" which is abstract. This property will be assumed to be a partial one.
The continuous property 1300eaf0-c5d9-472a-ab67-ddccd1950668 cannot be associated to a resqml property kind "dimensionless" which is abstract. This property will be assumed to be a partial one.
The continuous property 619a71bc-4032-409f-a6d0-d6c7abaefc4f cannot be associated to a resqml property kind "permeability rock" which is abstract. This property will be assumed to be a partial one.
The continuous property d56fae6c-0c46-40d4-a59d-20af37a4e746 cannot be associated to a resqml property kind "permeability rock" which is abstract. This property will be assumed to be a partial one.
The continuous property 0b30d052-c5c0-4b22-bcc2-26f3b2ce9741 cannot be associated to a resqml property kind "permeability rock" which is abstract. This property will be assumed to be a partial one.
The continuous property b3db8e3d-e96a-4b4c-9144-d6ec10293fa0 cannot be associated to a resqml property kind "permeability rock" which is abstract. This property will be assumed to be a partial one.
The continuous property 6bd4b9fc-5b54-4157-b64d-9b5ad26e910e cannot be associated to a resqml property kind "dimensionless" which is abstract. This property will be assumed to be a partial one.

[philippeVerney]

A RESQML property cannot be associated to an "abstract" RESQML proparty kind per the standard rule : see isAbstract property in https://docs.energistics.org/#RESQML/RESQML_TOPICS/RESQML-500-152-0-R-sv2010.html
The file you are reading is consequently invalid and FESAPI warns you about the fact it cannot read the RESQML invalid dataobjects contained in your epc file.
A partial property is similar to a dataobject which would just know its name and its title.

I just want to repeat that this is not a FESAPI rule but a RESQML one. FESAPI just checks it (maybe from v2.0.0.0).

[coderAkshay]

Got it.

Can we use some kind of Concrete Kind to this property? May be like custom property
Note: I don't have a prior knowledge of working with either RESQMLv2 or FesAPI

[philippeVerney]

If you are the writer of this file, sure you should!!!!
You should read carfully the RESQML spec : at the very least https://docs.energistics.org/#RESQML/RESQML_TOPICS/RESQML-000-173-0-C-sv2010.html and follow the spec.

For instance, solving the error on "permeability rock" is an easy one, just use "rock permeability" kind (which is not abstract) instead.
Solving the dimensionless one is much harder and it all depends on what the writer really wants to export....

[coderAkshay]

[philippeVerney]

Please stop using a single thread for all your questions.
One question/topic should go in one single thread. Otherwise, it would be hard for other readers to find back similar issues.