// The death testing framework causes this to have interesting semantics,
// since the sideeffects of the call are only visible in opt mode, and not
// in debug mode.
//
// In practice, this can be used to test functions that utilize the
// LOG(DFATAL) macro using the following style:
//
// int DieInDebugOr12(int* sideeffect) {
//   if (sideeffect) {
//     *sideeffect = 12;
//   }
//   LOG(DFATAL) << "death";
//   return 12;
// }

## Les annotations de type dans **FastAPI**

**FastAPI** utilise ces annotations pour faire différentes choses.

Avec **FastAPI**, vous déclarez des paramètres grâce aux annotations de types et vous obtenez :

* **du support de l'éditeur**
* **de la vérification de types**

...et **FastAPI** utilise ces mêmes déclarations pour :

#ifndef TENSORFLOW_COMPILER_MLIR_LITE_UTILS_UTILS_H_
#define TENSORFLOW_COMPILER_MLIR_LITE_UTILS_UTILS_H_

#include <algorithm>
#include <cstddef>
#include <cstdint>
#include <utility>
#include <vector>

#include "llvm/ADT/ArrayRef.h"
#include "mlir/IR/Attributes.h"  // from @llvm-project
#include "mlir/IR/BuiltinAttributeInterfaces.h"  // from @llvm-project

Webargs est un outil qui a été créé pour fournir cela par-dessus plusieurs frameworks, dont Flask.

Il utilise Marshmallow pour effectuer la validation des données. Et il a été créé par les mêmes développeurs.

C'est un outil formidable et je l'ai beaucoup utilisé aussi, avant d'avoir **FastAPI**.

!!! info
Webargs a été créé par les développeurs de Marshmallow.

!!! check "A inspiré **FastAPI** à"

			cfg.SetUserSupplied(true)
			return cfg, nil
		}
	}
	return nil, nil
}

// fromConfigMap is an utility function, which will load the value of a key of a config map and use h.FromDocumentMap() to perform the parsing
// This is an utility func. Used by the component config support implementations. Don't use it outside of that context.

        checkEclipseWtpComponentForJava()
        checkEclipseWtpFacet([
                new Facet(FacetType.fixed, 'jst.java', null),
                new Facet(FacetType.installed, 'jst.utility', '1.0'),
                new Facet(FacetType.installed, 'jst.java', '6.0')])
    }

    def applyFirstToJavaProject_shouldHaveWebProjectAndClasspathTask() {
        when:
        wtpPlugin.apply(project)

Since our build consists of multiple-subprojects, we want to share build logic and configuration between them.
For this, we utilize so-called _convention plugins_ that are located in the `buildSrc` folder.
Convention plugins in `buildSrc` are an easy way to utilise Gradle's plugin system to write reusable bits of build configuration.

in this sample, we can find three such convention plugins that are based on each other:

#ifndef TENSORFLOW_COMPILER_MLIR_TENSORFLOW_ANALYSIS_SIDE_EFFECT_ANALYSIS_H_
#define TENSORFLOW_COMPILER_MLIR_TENSORFLOW_ANALYSIS_SIDE_EFFECT_ANALYSIS_H_

#include <cstddef>
#include <cstdint>
#include <memory>
#include <utility>

#include "absl/container/flat_hash_map.h"
#include "absl/container/flat_hash_set.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/STLFunctionalExtras.h"
#include "llvm/ADT/SetVector.h"

Excluding a transitive dependency might lead to runtime errors if external libraries do not properly function without them.
If you use excludes, make sure that you do not utilise any code path requiring the excluded dependency by sufficient test coverage.
====

Transitive dependencies can be excluded on the level of a declared dependency.

#define TENSORFLOW_COMPILER_MLIR_QUANTIZATION_COMMON_IR_UNIFORMSUPPORT_H_

#include <algorithm>
#include <cassert>
#include <cmath>
#include <cstdint>
#include <cstring>
#include <utility>

#include "llvm/ADT/APFloat.h"
#include "llvm/ADT/APInt.h"
#include "llvm/ADT/APSInt.h"
#include "mlir/Dialect/Quant/QuantTypes.h"  // from @llvm-project
#include "mlir/IR/Attributes.h"  // from @llvm-project