UploadLanguage (EN)Support
BusinessMobileSocial MediaMarketingTechnologyArt & PhotosCareerDesignEducationPresentations & Public SpeakingGovernment & NonprofitHealthcareInternetLawLeadership & ManagementAutomotiveEngineeringSoftwareRecruiting & HRRetailSalesServicesScienceSmall Business & EntrepreneurshipFoodEnvironmentEconomy & FinanceData & AnalyticsInvestor RelationsSportsSpiritualNews & PoliticsTravelSelf ImprovementReal EstateEntertainment & HumorHealth & MedicineDevices & HardwareLifestyle
Upload
Sign in
Germán Diago Gómez, profile picture
Uploaded byGermán Diago Gómez
414 views

Stl design overview

The document provides an overview of the Standard Template Library (STL) in C++, focusing on its fundamental ideas, design choices, and main components, including containers, iterators, and algorithms. It discusses the importance of generic programming and concepts, emphasizing customization, reusability, and performance without penalty. The presentation aims to introduce key elements of STL without delving into detailed examples due to time constraints.

Software◦
Downloaded 13 times
1 / 132
2 / 132
3 / 132
4 / 132
5 / 132
6 / 132
7 / 132
8 / 132
9 / 132
10 / 132
11 / 132
12 / 132
13 / 132
14 / 132
15 / 132
16 / 132
17 / 132
18 / 132
19 / 132
20 / 132
The STL design: an overview HCM City C++ users meetup Germán Diago Gómez May 22nd, 2016 Germán Diago Gómez The STL design: an overview May 22nd, 2016 1 / 32
Overview Introduction Goals of this talk introduce STL fundamental ideas Germán Diago Gómez The STL design: an overview May 22nd, 2016 2 / 32
Overview Introduction Goals of this talk introduce STL fundamental ideas understand the design choices that drove its current design Germán Diago Gómez The STL design: an overview May 22nd, 2016 2 / 32
Overview Introduction Goals of this talk introduce STL fundamental ideas understand the design choices that drove its current design light introduction to each of its fundamental parts Germán Diago Gómez The STL design: an overview May 22nd, 2016 2 / 32
Overview Introduction Goals of this talk introduce STL fundamental ideas understand the design choices that drove its current design light introduction to each of its fundamental parts Germán Diago Gómez The STL design: an overview May 22nd, 2016 2 / 32
Overview Introduction Goals of this talk introduce STL fundamental ideas understand the design choices that drove its current design light introduction to each of its fundamental parts Germán Diago Gómez The STL design: an overview May 22nd, 2016 2 / 32
Overview Introduction Non-goals not a detailed, fine-grained overview Germán Diago Gómez The STL design: an overview May 22nd, 2016 3 / 32
Overview Introduction Non-goals not a detailed, fine-grained overview not lots of examples Germán Diago Gómez The STL design: an overview May 22nd, 2016 3 / 32
Overview Introduction Non-goals not a detailed, fine-grained overview not lots of examples due to time constraints Germán Diago Gómez The STL design: an overview May 22nd, 2016 3 / 32
Fundamentals STL STL stands for Standard Template Library Germán Diago Gómez The STL design: an overview May 22nd, 2016 4 / 32
Fundamentals STL STL stands for Standard Template Library it is a subset of the C++ Standard Library Germán Diago Gómez The STL design: an overview May 22nd, 2016 4 / 32
Fundamentals STL STL stands for Standard Template Library it is a subset of the C++ Standard Library its main author is Alexander Stepanov Germán Diago Gómez The STL design: an overview May 22nd, 2016 4 / 32
Fundamentals STL STL stands for Standard Template Library it is a subset of the C++ Standard Library its main author is Alexander Stepanov got accepted into ISO C++98 standard library Germán Diago Gómez The STL design: an overview May 22nd, 2016 4 / 32
Fundamentals Design general Germán Diago Gómez The STL design: an overview May 22nd, 2016 5 / 32
Fundamentals Design general allows a high degree of customization through composition Germán Diago Gómez The STL design: an overview May 22nd, 2016 5 / 32
Fundamentals Design general allows a high degree of customization through composition extensible: you can add new components to the framework Germán Diago Gómez The STL design: an overview May 22nd, 2016 5 / 32
Fundamentals Design general allows a high degree of customization through composition extensible: you can add new components to the framework reusable without loss of performace compared to hand-written code Germán Diago Gómez The STL design: an overview May 22nd, 2016 5 / 32
Fundamentals Design general allows a high degree of customization through composition extensible: you can add new components to the framework reusable without loss of performace compared to hand-written code intensive use of generic programming Germán Diago Gómez The STL design: an overview May 22nd, 2016 5 / 32
Fundamentals STL main components containers Germán Diago Gómez The STL design: an overview May 22nd, 2016 6 / 32
Fundamentals STL main components containers iterators Germán Diago Gómez The STL design: an overview May 22nd, 2016 6 / 32
Fundamentals STL main components containers iterators algorithms Germán Diago Gómez The STL design: an overview May 22nd, 2016 6 / 32
Fundamentals STL main components containers iterators algorithms utilities (not the focus of this talk) Germán Diago Gómez The STL design: an overview May 22nd, 2016 6 / 32
Fundamentals Why generic programming allows compile-time polymorphism combined with overloading Germán Diago Gómez The STL design: an overview May 22nd, 2016 7 / 32
Fundamentals Why generic programming allows compile-time polymorphism combined with overloading allows customized code generation at compile-time Germán Diago Gómez The STL design: an overview May 22nd, 2016 7 / 32
Fundamentals Why generic programming allows compile-time polymorphism combined with overloading allows customized code generation at compile-time allows choose best optimization based on overloading, which is a compile-time mechanism Germán Diago Gómez The STL design: an overview May 22nd, 2016 7 / 32
Fundamentals Why generic programming allows compile-time polymorphism combined with overloading allows customized code generation at compile-time allows choose best optimization based on overloading, which is a compile-time mechanism allows high degree of customization, for both algorithms and data structures without loss of performance Germán Diago Gómez The STL design: an overview May 22nd, 2016 7 / 32
Fundamentals Why generic programming (2) types can retroactively model a concept Germán Diago Gómez The STL design: an overview May 22nd, 2016 8 / 32
Fundamentals Why generic programming (2) types can retroactively model a concept example: array pointers are iterators Germán Diago Gómez The STL design: an overview May 22nd, 2016 8 / 32
Fundamentals Why generic programming (2) types can retroactively model a concept example: array pointers are iterators OO alternative: uses run-time polymorphism and dispatch Germán Diago Gómez The STL design: an overview May 22nd, 2016 8 / 32
Fundamentals Why generic programming (2) types can retroactively model a concept example: array pointers are iterators OO alternative: uses run-time polymorphism and dispatch performance penalty Germán Diago Gómez The STL design: an overview May 22nd, 2016 8 / 32
Fundamentals Concepts briefly explained Concepts are the foundation of generic programming Germán Diago Gómez The STL design: an overview May 22nd, 2016 9 / 32
Fundamentals Concepts briefly explained Concepts are the foundation of generic programming a Concept is a set of requirements that a type must meet Germán Diago Gómez The STL design: an overview May 22nd, 2016 9 / 32
Fundamentals Concepts briefly explained Concepts are the foundation of generic programming a Concept is a set of requirements that a type must meet the syntactic requirements that a type must meet Germán Diago Gómez The STL design: an overview May 22nd, 2016 9 / 32
Fundamentals Concepts briefly explained Concepts are the foundation of generic programming a Concept is a set of requirements that a type must meet the syntactic requirements that a type must meet the semantic requirements that these syntactic requirements must meet Germán Diago Gómez The STL design: an overview May 22nd, 2016 9 / 32
Fundamentals Concepts briefly explained Concepts are the foundation of generic programming a Concept is a set of requirements that a type must meet the syntactic requirements that a type must meet the semantic requirements that these syntactic requirements must meet optionally, associated types for that concept Germán Diago Gómez The STL design: an overview May 22nd, 2016 9 / 32
Fundamentals Concepts briefly explained Concepts are the foundation of generic programming a Concept is a set of requirements that a type must meet the syntactic requirements that a type must meet the semantic requirements that these syntactic requirements must meet optionally, associated types for that concept algorithms: include the big O cost associated to them Germán Diago Gómez The STL design: an overview May 22nd, 2016 9 / 32
Fundamentals Concepts briefly explained Concepts are the foundation of generic programming a Concept is a set of requirements that a type must meet the syntactic requirements that a type must meet the semantic requirements that these syntactic requirements must meet optionally, associated types for that concept algorithms: include the big O cost associated to them In C++ these are the requirements that your template parameters need to fullfill. They are stated in the documentation, since there is no language support (yet) for Concepts Germán Diago Gómez The STL design: an overview May 22nd, 2016 9 / 32
Fundamentals Concepts briefly explained (2) similar to the way a virtual function sets requirements on derived types Germán Diago Gómez The STL design: an overview May 22nd, 2016 10 / 32
Fundamentals Concepts briefly explained (2) similar to the way a virtual function sets requirements on derived types but at compile-time Germán Diago Gómez The STL design: an overview May 22nd, 2016 10 / 32
Fundamentals Concepts briefly explained (2) similar to the way a virtual function sets requirements on derived types but at compile-time when a type fullfills the requirements of a concept we say it models that concept Germán Diago Gómez The STL design: an overview May 22nd, 2016 10 / 32
Fundamentals Concepts briefly explained (2) similar to the way a virtual function sets requirements on derived types but at compile-time when a type fullfills the requirements of a concept we say it models that concept similar to when we say a derived class implements requirements from a base class Germán Diago Gómez The STL design: an overview May 22nd, 2016 10 / 32
Fundamentals Concepts briefly explained (2) similar to the way a virtual function sets requirements on derived types but at compile-time when a type fullfills the requirements of a concept we say it models that concept similar to when we say a derived class implements requirements from a base class types do not need to inherit from any class to model a Concept → more loosely coupled Germán Diago Gómez The STL design: an overview May 22nd, 2016 10 / 32
Fundamentals Concepts example: LessThanComparable syntactic requirements Germán Diago Gómez The STL design: an overview May 22nd, 2016 11 / 32
Fundamentals Concepts example: LessThanComparable syntactic requirements your type must work with operator< and return a type convertible to bool Germán Diago Gómez The STL design: an overview May 22nd, 2016 11 / 32
Fundamentals Concepts example: LessThanComparable syntactic requirements your type must work with operator< and return a type convertible to bool semantic requirements for LessThanComparable operator< Germán Diago Gómez The STL design: an overview May 22nd, 2016 11 / 32
Fundamentals Concepts example: LessThanComparable syntactic requirements your type must work with operator< and return a type convertible to bool semantic requirements for LessThanComparable operator< ∀a, !(a < a) Germán Diago Gómez The STL design: an overview May 22nd, 2016 11 / 32
Fundamentals Concepts example: LessThanComparable syntactic requirements your type must work with operator< and return a type convertible to bool semantic requirements for LessThanComparable operator< ∀a, !(a < a) a < b ⇐⇒ !(b < a) Germán Diago Gómez The STL design: an overview May 22nd, 2016 11 / 32
Fundamentals Concepts example: LessThanComparable syntactic requirements your type must work with operator< and return a type convertible to bool semantic requirements for LessThanComparable operator< ∀a, !(a < a) a < b ⇐⇒ !(b < a) a < b ∧ b < c =⇒ a < c Germán Diago Gómez The STL design: an overview May 22nd, 2016 11 / 32
Fundamentals Concepts example: LessThanComparable syntactic requirements your type must work with operator< and return a type convertible to bool semantic requirements for LessThanComparable operator< ∀a, !(a < a) a < b ⇐⇒ !(b < a) a < b ∧ b < c =⇒ a < c associated types (optionally) Germán Diago Gómez The STL design: an overview May 22nd, 2016 11 / 32
Fundamentals More Concepts examples BinaryPredicate DefaultConstructible Many others Germán Diago Gómez The STL design: an overview May 22nd, 2016 12 / 32
Containers Introduction Containers allow you to store collections of data in different data structures that best suite your needs. Germán Diago Gómez The STL design: an overview May 22nd, 2016 13 / 32
Containers STL containers classification sequence containers Germán Diago Gómez The STL design: an overview May 22nd, 2016 14 / 32
Containers STL containers classification sequence containers std::array<T>, std::vector<T>, std::list<T>. . . Germán Diago Gómez The STL design: an overview May 22nd, 2016 14 / 32
Containers STL containers classification sequence containers std::array<T>, std::vector<T>, std::list<T>. . . associative containers Germán Diago Gómez The STL design: an overview May 22nd, 2016 14 / 32
Containers STL containers classification sequence containers std::array<T>, std::vector<T>, std::list<T>. . . associative containers ordered → std::set. . . Germán Diago Gómez The STL design: an overview May 22nd, 2016 14 / 32
Containers STL containers classification sequence containers std::array<T>, std::vector<T>, std::list<T>. . . associative containers ordered → std::set. . . unordered → std::unordered_map. . . Germán Diago Gómez The STL design: an overview May 22nd, 2016 14 / 32
Containers STL containers classification sequence containers std::array<T>, std::vector<T>, std::list<T>. . . associative containers ordered → std::set. . . unordered → std::unordered_map. . . container adaptors Germán Diago Gómez The STL design: an overview May 22nd, 2016 14 / 32
Containers STL containers classification sequence containers std::array<T>, std::vector<T>, std::list<T>. . . associative containers ordered → std::set. . . unordered → std::unordered_map. . . container adaptors std::priority_queue, std::stack. . . Germán Diago Gómez The STL design: an overview May 22nd, 2016 14 / 32
Containers STL containers traversal depends on the kind iterators provided (more later) Germán Diago Gómez The STL design: an overview May 22nd, 2016 15 / 32
Containers STL containers traversal depends on the kind iterators provided (more later) random access, bidirectional, forward containers, etc. Germán Diago Gómez The STL design: an overview May 22nd, 2016 15 / 32
Iterators Introduction Iterators are used to traverse ranges Germán Diago Gómez The STL design: an overview May 22nd, 2016 16 / 32
Iterators Introduction Iterators are used to traverse ranges Iterators are an abstraction for positions Germán Diago Gómez The STL design: an overview May 22nd, 2016 16 / 32
Iterators Introduction Iterators are used to traverse ranges Iterators are an abstraction for positions they are modeled after C pointers in C++’s concrete case Germán Diago Gómez The STL design: an overview May 22nd, 2016 16 / 32
Iterators Introduction Iterators are used to traverse ranges Iterators are an abstraction for positions they are modeled after C pointers in C++’s concrete case you can do a traversal with two iterators Germán Diago Gómez The STL design: an overview May 22nd, 2016 16 / 32
Iterators Introduction Iterators are used to traverse ranges Iterators are an abstraction for positions they are modeled after C pointers in C++’s concrete case you can do a traversal with two iterators they are not the same as an iterator in other languages Germán Diago Gómez The STL design: an overview May 22nd, 2016 16 / 32
Iterators Introduction Iterators are used to traverse ranges Iterators are an abstraction for positions they are modeled after C pointers in C++’s concrete case you can do a traversal with two iterators they are not the same as an iterator in other languages Java/C# iterator similar to a pair of iterators in C++ Germán Diago Gómez The STL design: an overview May 22nd, 2016 16 / 32
Iterators Introduction Iterators are used to traverse ranges Iterators are an abstraction for positions they are modeled after C pointers in C++’s concrete case you can do a traversal with two iterators they are not the same as an iterator in other languages Java/C# iterator similar to a pair of iterators in C++ a pair of iterators is usually called a range in C++ Germán Diago Gómez The STL design: an overview May 22nd, 2016 16 / 32
Iterators Iterator categories several kinds of iterators Germán Diago Gómez The STL design: an overview May 22nd, 2016 17 / 32
Iterators Iterator categories several kinds of iterators iterators form a hierarchy of Concepts Germán Diago Gómez The STL design: an overview May 22nd, 2016 17 / 32
Iterators Iterator categories several kinds of iterators iterators form a hierarchy of Concepts not implemented as inheritance in the library Germán Diago Gómez The STL design: an overview May 22nd, 2016 17 / 32
Iterators Iterator categories several kinds of iterators iterators form a hierarchy of Concepts not implemented as inheritance in the library they are Concepts Germán Diago Gómez The STL design: an overview May 22nd, 2016 17 / 32
Iterators Iterator categories several kinds of iterators iterators form a hierarchy of Concepts not implemented as inheritance in the library they are Concepts RandomAccessIterator → BidirectionalIterator → ForwardIterator → InputIterator Germán Diago Gómez The STL design: an overview May 22nd, 2016 17 / 32
Iterators Iterator categories several kinds of iterators iterators form a hierarchy of Concepts not implemented as inheritance in the library they are Concepts RandomAccessIterator → BidirectionalIterator → ForwardIterator → InputIterator OutputIterator is any iterator that can write to the element it points to Germán Diago Gómez The STL design: an overview May 22nd, 2016 17 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access increment iterator position by one Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access increment iterator position by one single traversal Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access increment iterator position by one single traversal ForwardIterator Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access increment iterator position by one single traversal ForwardIterator InputIterator + support for multiple traversals Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access increment iterator position by one single traversal ForwardIterator InputIterator + support for multiple traversals BidirectionalIterator Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access increment iterator position by one single traversal ForwardIterator InputIterator + support for multiple traversals BidirectionalIterator ForwardIterator + supports decrementing position by one Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access increment iterator position by one single traversal ForwardIterator InputIterator + support for multiple traversals BidirectionalIterator ForwardIterator + supports decrementing position by one RandomAccessIterator Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access increment iterator position by one single traversal ForwardIterator InputIterator + support for multiple traversals BidirectionalIterator ForwardIterator + supports decrementing position by one RandomAccessIterator BidirectionalIterator + supports jumping to arbitrary position in O(1) Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Iterator categories (2) every iterator category supports a set of operations InputIterator element access increment iterator position by one single traversal ForwardIterator InputIterator + support for multiple traversals BidirectionalIterator ForwardIterator + supports decrementing position by one RandomAccessIterator BidirectionalIterator + supports jumping to arbitrary position in O(1) think of a pointer to C array position Germán Diago Gómez The STL design: an overview May 22nd, 2016 18 / 32
Iterators Why are iterators important abstract traversal from the underlying data structure Germán Diago Gómez The STL design: an overview May 22nd, 2016 19 / 32
Iterators Why are iterators important abstract traversal from the underlying data structure decouple algorithms implementation from data structure (later more) Germán Diago Gómez The STL design: an overview May 22nd, 2016 19 / 32
Iterators Why are iterators important abstract traversal from the underlying data structure decouple algorithms implementation from data structure (later more) can implement algorithms based on its iterator category (later more) Germán Diago Gómez The STL design: an overview May 22nd, 2016 19 / 32
Iterators Why are iterators important abstract traversal from the underlying data structure decouple algorithms implementation from data structure (later more) can implement algorithms based on its iterator category (later more) containers just need to provide a begin/end sequence pair Germán Diago Gómez The STL design: an overview May 22nd, 2016 19 / 32
Iterators Why are iterators important abstract traversal from the underlying data structure decouple algorithms implementation from data structure (later more) can implement algorithms based on its iterator category (later more) containers just need to provide a begin/end sequence pair let you reduce implementation effort on algorithms Germán Diago Gómez The STL design: an overview May 22nd, 2016 19 / 32
Algorithms Introduction algorithms let you perform operations on your data Germán Diago Gómez The STL design: an overview May 22nd, 2016 20 / 32
Algorithms Introduction algorithms let you perform operations on your data STL algorithms design choices are a bit suprising at first Germán Diago Gómez The STL design: an overview May 22nd, 2016 20 / 32
Algorithms Algorithm design choices it was decided to implement algorithms as free functions, not as member functions Germán Diago Gómez The STL design: an overview May 22nd, 2016 21 / 32
Algorithms Algorithm design choices it was decided to implement algorithms as free functions, not as member functions it was decided that it would take Iterator s as input, instead of Container s Germán Diago Gómez The STL design: an overview May 22nd, 2016 21 / 32
Algorithms Algorithm design choices it was decided to implement algorithms as free functions, not as member functions it was decided that it would take Iterator s as input, instead of Container s consequence: algorithms are decoupled from containers Germán Diago Gómez The STL design: an overview May 22nd, 2016 21 / 32
Algorithms STL algorithms/iterators conventions algorithm specifications use half-open ranges for algorithms Germán Diago Gómez The STL design: an overview May 22nd, 2016 22 / 32
Algorithms STL algorithms/iterators conventions algorithm specifications use half-open ranges for algorithms (a, b] Germán Diago Gómez The STL design: an overview May 22nd, 2016 22 / 32
Algorithms STL algorithms/iterators conventions algorithm specifications use half-open ranges for algorithms (a, b] a is included in the range. b is not included in the range Germán Diago Gómez The STL design: an overview May 22nd, 2016 22 / 32
Algorithms STL algorithms/iterators conventions algorithm specifications use half-open ranges for algorithms (a, b] a is included in the range. b is not included in the range containers provide begin, end and its variants Germán Diago Gómez The STL design: an overview May 22nd, 2016 22 / 32
Algorithms STL algorithms/iterators conventions algorithm specifications use half-open ranges for algorithms (a, b] a is included in the range. b is not included in the range containers provide begin, end and its variants begin points to the first element of the container Germán Diago Gómez The STL design: an overview May 22nd, 2016 22 / 32