Document no: P1995R0
Authors: Joshua Berne, Timur Doumler, Andrzej Krzemieński, Ryan McDougall, Herb Sutter
SG21 has gathered a large number of use cases for contracts between teh WG21 Cologne and Belfast meetings. This paper presents those use cases, along with some initial results from polling done of SG21 members to identify some level of important to the community for each individual use case.
Each use case has been assigned an identifier that can be used to reference these use cases in other papers, which will hopefully be stable. We expect this content to evolve in a number of ways:
Polling occurred through SurveyMonkey between the July 2019 WG21 Cologne meeting and the November 2019 WG21 Belfast meeting. Participants were asked to choose from three answers for each of the 195 different use cases:
Participants were also able to not answer at all for a given question (N/A), and this happened for a small number of cases. The table of use cases below lists 4 columns with the total number of each response chosen. The score column is the average value of the responses amongst those who provided a response.
|Code||As A||In Order To||I Want To||N/A||Not||Nice||Must||Score|
|dev.reason.knowl||Developer||Reason explicitly||Annotate my program anywhere in the code with my current understanding of its structure or execution||0||0||10||20||1.666667|
|dev.reason.confidence||Developer||Reason explicitly||Express a spectrum of confidence in my annotations, from "unsure" and asking for validation, to "sure" and asking for some effect to be applied (eg. "maybe", "definitely", "assume" 'something')||0||10||10||10||1.000000|
|dev.reason.importance||Developer||Reason explicitly||Express a spectrum of importance of my annotations, from "critical" (eg. bring the system down) to "minor" (eg. lead to a slower fallback)||0||15||10||5||0.666667|
|dev.reason.cost||Developer||Reason explicitly||Express a spectrum of expected cost at compile or runtime of my annotations, from "unrunnable" to "expensive" to "cheap"||0||7||12||11||1.133333|
|dev.reason.behavior||Developer||Reason about executions||Have annotations affect the execution of my program in accordance with my expectations||0||7||9||14||1.233333|
|dev.reason.sideeffects||Developer||Reason about executions||Ensure annotations do not substantially change the meaning of my program whether enabled or disabled||0||3||8||19||1.533333|
|dev.reason.behaviorcontrol||Developer||Reason about executions||Have the effect of annotations on executions be user controllable (ie. decide whether "cheap" checks or "critical" terminates)||0||4||10||16||1.400000|
|dev.adapt||Developer||Adapt and progress with my project||Be able to easily change my confidence, importance, or other properties of my annotations over time||0||2||17||11||1.300000|
|dev.readable.syntax||Developer||Have readable annotations||Have annotations with a succinct and elegant syntax||0||0||13||17||1.566667|
|dev.readable.keywords||Developer||Have readable annotations||Have annotation keywords or names with intuitive, clear, and unambiguous meanings||0||0||11||19||1.633333|
|dev.readable.priority||Developer||Have readable annotations||Have my contract specification to be visually primary, and secondary information (syntax, hints, roles, levels, etc.) to not be distracting||0||7||17||6||0.966667|
|dev.parsable||Developer||Interoperate with tools or persons||A syntax that can both be parsed and can be reasoned about semantically||0||1||6||23||1.733333|
|dev.tooling||Developer||Interoperate with tools or persons||Expose annotations to tools that might leverage them (eg. code linter, static analyzer, semantic prover, compiler sanitizer, binary analyzer, code reviewer, etc.)||0||1||9||20||1.633333|
|cppdev.syntax.familiar||C++ Developer||Get up to speed||Have annotations use familiar syntax||0||12||15||3||0.700000|
|cppdev.syntax.cpp||C++ Developer||Get up to speed||Have annotations use C++ syntax||1||8||11||10||1.068966|
|cppdev.syntax.reuse||C++ Developer||Reuse code||Have annotations use my custom types or functions||0||1||9||20||1.633333|
|cppdev.location||C++ Developer||Have a single source of truth||Use same source file for both code and annotations||0||1||4||25||1.800000|
|cppdev.syntax.macros||C++ Developer||Support modern features||Minimize use of macros||0||5||18||7||1.066667|
|cppdev.modules||C++ Developer||Support modern features||Be interoperable with modules||0||0||6||24||1.800000|
|cppdev.coroutines||C++ Developer||Support modern features||Be interoperable with coroutines||0||3||11||16||1.433333|
|cppdev.concepts||C++ Developer||Support modern features||Be interoperable with concepts||0||3||9||18||1.500000|
|cppdev.existing.std||C++ Developer||Use the standard library in-contract||Codify existing exposition-only standard library requirements||0||5||18||7||1.066667|
|cppdev.debugger||C++ Developer||Use Debugger||Have runtime able to launch a debugger from an annotation if necessary||0||10||12||8||0.933333|
|cppdev.build.legacy||C++ Developer||Use existing build modes||Have annotations affect executions depending on my existing build modes (eg. Debug or Release modes in VS)||0||9||13||8||0.966667|
|cdev.contracts||C Developer||Write contracts on my functions||Specify contracts in a way standardizable as part of the C language||1||21||6||2||0.344828|
|cdev.identifiers||C Developer||Write contracts on my functions||Use contracts with macro-safe keywords that are reserved C names (i.e., _Pre, _Post, _Assert, etc.)||1||22||6||1||0.275862|
|cdev.violationhandler||C Developer||Write contracts on my functions||Have a common violation handler for both violated C and C++ contracts||1||16||11||2||0.517241|
|cdev.ignorable||C Developer||Write contracts on my functions||Make all contract semantics optional (so as not to change WG14-N2385 6.7.11 p2)||1||23||3||3||0.310345|
|ccppdev.interop||Mixed C/C++ Developer||Maintain mixed code base||Not lose contracts when crossing languages||1||12||14||3||0.689655|
|cdev.cppinterop||Mixed C/C++ Developer||Write contracts on my functions||Expose my contracts to C++ developers through 'extern "C"' declarations of my functions||1||16||12||1||0.482759|
|api.communicate.inputsoutputs||API Developer||Communicate my interface to users||Document the expected inputs and expected outputs on my interface||0||3||5||22||1.633333|
|api.establish.check||API Developer||Establish a contract||Have validation inform me which output values are unexpected or invalid||0||2||14||14||1.400000|
|api.establish.validate_invariants||API Developer||Establish a contract||Have validation inform me which class invariants are violated||0||2||14||14||1.400000|
|api.establish.values||API Developer||Establish a contract||Have validation inform user which input values are unexpected or invalid||0||1||11||18||1.566667|
|api.establish.preconditions||API Developer||Establish a contract||Have contracts specify their pre-conditions as logical predicates||0||2||3||25||1.766667|
|api.establish.invariants||API Developer||Establish a contract||Have contracts specify their class invariants as logical predicates||0||5||13||12||1.233333|
|api.establish.postconditions||API Developer||Establish a contract||Have contracts specify their post-conditions as logical predicates||0||5||5||20||1.500000|
|api.express.values||API Developer||Express predicates||Make reference to either the values of my inputs, or other in-scope identifiers||0||2||2||26||1.800000|
|api.establish.changedvalues||API Developer||Establish a contract||Make reference to the before and after values of in-out variables (ie. passed by pointer or reference) in post-conditions||0||7||11||12||1.166667|
|api.establish.changedmembers||API Developer||Establish a contract||Make reference to the before and after values of mutable class members (eg. new_size = old_size+1 after push_back) in post-conditions||0||7||14||9||1.066667|
|api.establish.changedstate||API Developer||Establish a contract||Make reference to the before and after values of global state (eg., global >= old(global) + 1) in post-conditions||0||13||14||3||0.666667|
|api.extend.exceptionsafety||API Developer||Extend contractual aspects||Annotate operations as being exception safe||0||13||17||0||0.566667|
|api.extend.threadsafety||API Developer||Extend contractual aspects||Annotate operations as being thread safe||0||13||17||0||0.566667|
|api.extend.atomicity||API Developer||Extend contractual aspects||Annotate operations as being atomic (ie. all or no changes become visible)||0||15||15||0||0.500000|
|api.extend.realtime||API Developer||Extend contractual aspects||Annotate operations as real-time (ie. guaranteed to complete within a time frame)||0||20||9||1||0.366667|
|api.extend.determinism||API Developer||Extend contractual aspects||Annotate operations as being deterministic (ie. same outputs for same inputs)||0||11||13||6||0.833333|
|api.extend.purity||API Developer||Extend contractual aspects||Annotate operations as functionally pure (ie. no side effects)||0||10||15||5||0.833333|
|api.extend.sideeffects||API Developer||Extend contractual aspects||Annotate operations as having global side effects (ie. write to singleton, file, network, or database)||0||13||17||0||0.566667|
|api.extend.complexity||API Developer||Extend contractual aspects||Annotate algorithmic complexity||0||19||11||0||0.366667|
|api.express.runnability||API Developer||Express unrunnable contracts||Be able to use a predicate that is not evaluated at runtime, because it might be unsafe to run or have stateful side effects||0||6||14||10||1.133333|
|api.express.undefined||API Developer||Express unrunnable contracts||Be able to use a predicate that doesn't have a definition, because it hasn't been written yet, or is infeasible to run||0||8||12||10||1.066667|
|api.express.uncheckable||API Developer||Express uncheckable contracts||Be able to use a predicate that is not evaluated, because it is simply a semantic placeholder for a tool||0||8||12||10||1.066667|
|api.express.unimplementable||API Developer||Express uncheckable contracts||Be able to use a predicate that cannot have a complete definition, because it is inexpressible in the language||0||11||9||10||0.966667|
|api.establish.responsibility||API Developer||Establish responsibility boundaries||Inform users which errors are the responsibility of the caller, and which are the callee||0||4||11||15||1.366667|
|api.resp.preassert||API Developer||Establish responsibility boundaries||Annotate assertions inside function bodies that indirectly test preconditions (such as malformed data discovered while performing the algorithm) should be reported to the caller as precondition failures||0||12||13||5||0.766667|
|api.contract.interface||API Developer||Have contract as part of my interface||Declare contract when I declare the function||0||3||6||21||1.600000|
|api.contract.private||API Developer||Keep my user interfaces clean and narrow||Be able to access private implementation details of the class so I don't have to widen public interface to declare predicates||0||5||12||13||1.266667|
|api.contract.redeclaration||API Developer||Keep my public interfaces clean and concise||Place function contract conditions on any declaration (e.g., on redeclarations at the bottom of the header, or on the definition in an implementation file, where they are less distracting).||0||12||11||7||0.833333|
|api.contract.errorhandling||API Developer||Move contract violation out of error handling||Replace uses of error handling to express contract violation (eg. operator(size_t n) noexcept [[pre: n < size()]] instead of throwing)||0||7||11||12||1.166667|
|cppapi.invariants||C++ API Developer||Write classes||Declare class invariants that all of my public functions need to maintain||1||4||13||12||1.275862|
|cppapi.class.preconditions||C++ API Developer||Maintain a class hierarchy||Ensure overriding methods have same or wider preconditions (see: Liskov substitution principle)||0||4||20||6||1.066667|
|cppapi.class.postconditions||C++ API Developer||Maintain a class hierarchy||Ensure overriding functions meet their base class postconditions when their base class preconditions are met (see: Liskov substitution principle)||0||2||20||8||1.200000|
|cppapi.class.variability||C++ API Developer||Maintain a class hierarchy.||Allow overriding functions to have narrower preconditions/wider postconditions if I want to||0||13||14||3||0.666667|
|api.class.publicinterface||C++ API Developer||Express public class invariants||Express a restriction on the public interface of a type that all callers of the type can depend upon: can mention only public members, and is checked on entry and exit from this type's code||0||9||19||2||0.766667|
|api.class.publicinvariants||C++ API Developer||Express public class invariants||Check invariants before and after every public method (when called from outside the type, not when one member function calls another)||0||5||21||4||0.966667|
|api.class.publiccalls||C++ API Developer||Express public class invariants||Check invariants before and after calling functions that are not part of this type (including virtual calls)||0||8||20||2||0.800000|
|api.class.baseinterface||C++ API Developer||Express base class invariants||Express a restriction on the protected interface of a type that derived types can depend upon: can mention only protected and public members, and is checked on entry and exit from this type's code||0||14||15||1||0.566667|
|api.class.baseinvariants||C++ API Developer||Express base class invariants||Check invariants on entry and exit of every protected method (when called from the derived type, not when one base member function calls another)||1||14||13||2||0.586207|
|api.class.basecalls||C++ API Developer||Express base class invariants||Check invariants before and after every call to a virtual function (when calling to the derived type)||0||10||18||2||0.733333|
|api.class.privateinterface||C++ API Developer||Express private class invariants||Express an internal restriction on the private implementation of a type, can mention any member, and is checked on entry and exit from this type's code||0||12||14||4||0.733333|
|api.class.privateinvariants||C++ API Developer||Express private class invariants||Check invariants on entry and exit of every public method (when called from outside the type, not when one member function calls another)||0||9||17||4||0.833333|
|api.class.privatecalls||C++ API Developer||Express private class invariants||Check invariants before and after calling functions that are not part of this type (including virtual calls)||0||13||15||2||0.633333|
|api.class.testing||C++ API Developer||Test my classes||For every member or friend function in my class, run my unit test framework with checking enabled for every assertion at the point where it is written, and check every postcondition at every non-exceptional exit, and test my class invariants on entry and exit from this type's code||0||10||9||11||1.033333|
|cppapi.contracts.async||C++ API Developer||Enforce contracts in async code||Express contracts on callbacks such as std::function, function pointers, or references to functions, lambdas, or function objects||0||4||21||5||1.033333|
|cppapi.contracts.exception||C++ API Developer||Enforce contracts in exception safe code||Express contracts on exceptional exit||0||11||13||6||0.833333|
|cppapi.variadic||C++ API Developer||Use contracts with variadic templates||Allow predicate (fold) expansion||0||4||15||11||1.233333|
|api.coroutines||C++ API Developer||Use coroutines||Define and check pre and post conditions as I would a regular function||0||4||15||11||1.233333|
|api.coroutines.invariants||C++ API Developer||Use coroutines||Define and check invariants over all entry and exit points from a coroutine (to its awaiter or promise)||0||7||19||4||0.900000|
|int.conform.violation||Integration Developer||Conform to a contract||Be informed any time an interface's contract is violated||0||2||12||16||1.466667|
|int.conform.postconditions||Integration Developer||Conform to a contract||Verify results from a call are expected output values||0||4||9||17||1.433333|
|int.build.headeronly||Integration Developer||Build multiple libraries||Use contract-enabled header-only libraries||0||0||6||24||1.800000|
|int.build.binaries||Integration Developer||Build multiple libraries||Use contract-enabled binary libraries||1||1||7||21||1.689655|
|int.build.binarycounts||Integration Developer||Build multiple libraries||Only be required to manage a small, common set of build/link configurations||0||6||15||9||1.100000|
|int.build.control||Integration Developer||Debug multiple libraries||Enable checks only within a selected library||0||1||14||15||1.466667|
|int.build.control2||Integration Developer||Debug multiple libraries||Enable checks on multiple libraries simultaneously||0||2||14||14||1.400000|
|int.debug.callsites||Integration Developer||Debug multiple call sites||Enable checks only on selected call sites||0||5||18||7||1.066667|
|int.violations.information||Integration Developer||Correct failed checks||Be informed what check failed, when, where, and how||0||1||9||20||1.633333|
|int.violations.transmit||Integration Developer||Correct failed checks||Transmit check failure information in environment-specific ways (logs, email, special hardware traps, popup windows, blazing sirens, etc).||0||5||14||11||1.200000|
|int.violations.custom||Integration Developer||Correct failed checks||Install custom violation handler where I can inject custom logic to trap errors||0||7||9||14||1.233333|
|int.violations.common||Integration Developer||Unify violation handling||Be able to override how library violations are handled in the combined software to point into my handling code||0||6||9||15||1.300000|
|int.violations.override||Integration Developer||Be independent of build environment||Be able to define and override violation handler via source code||1||14||11||4||0.655172|
|int.build.minimize||Integration Developer||Minimize checking overhead||Disable library postconditions, asserts, and invariants, without disabling library preconditions (assuming the library is tested and stable and my code is not)||0||8||14||8||1.000000|
|int.control.build||Integrated Software Provider||Ensure the combined software is correct||Turn checks on at build time||0||3||6||21||1.600000|
|int.control.runtime||Integrated Software Provider||Ensure the combined software is correct||Turn checks on at run time||0||14||11||5||0.700000|
|int.conrol.subsets.build||Integrated Software Provider||Ensure the combined software is correct||Turn on any subset of individual (call site) checks on at build time||1||8||12||9||1.034483|
|int.control.subsets.runtime||Integrated Software Provider||Ensure the combined software is correct||Turn on any subset of individual (call site) checks on at run time||0||15||12||3||0.600000|
|int.consistency||Integrated Software Provider||Ensure the combined software is correct||Verify all annotations are globally consistent when integrated||1||6||20||3||0.896552|
|int.control.subsets||Integrated Software Provider||Ensure individual features are correct||Have a way to audit (named or semantic) subsets of checks for various deployments||1||9||16||4||0.827586|
|int.build.common||Integrated Software Provider||Manage binary delivery||Be able to use the same executable regardless of contract enforcement mode||0||16||10||4||0.600000|
|int.testing.control||Integrated Software Provider||Define "Code Under Test"||Selectively enable checking for a set of functions which could name either an individual function or an overload set||0||11||16||3||0.733333|
|int.testing.controltypes||Integrated Software Provider||Define "Code Under Test"||Selectively enable checking for a set of types and all their members||0||12||15||3||0.700000|
|int.testing.transitivity||Integrated Software Provider||Define "Code Under Test"||Selectively enable checking for a set of types and all their transitively nested types and members||0||13||14||3||0.666667|
|int.testing.modules||Integrated Software Provider||Define "Code Under Test"||Selectively enable checking for a translation unit or module and all (non transitive) types and functions within||0||6||19||5||0.966667|
|int.build.unchecked||Integrated Software Provider||Test final deliverable||Turn off build time checking to remove checking overhead||0||9||6||15||1.200000|
|int.runtime.unchecked||Integrated Software Provider||Test final deliverable||Turn off run time checking to remove checking overhead||0||3||5||22||1.633333|
|int.build.optimize||Integrated Software Provider||Test final deliverable||Turn on run time optimization to leverage annotation assumptions||0||6||12||12||1.200000|
|cpplib.headeronly||C++ Library Developer||Use templates||Be able to ship header only library||0||1||4||25||1.800000|
|cpplib.insulation||C++ Library Developer||Control the tradeoff between need for client recompilation and contract condition visibility||Insulate contract conditions with the function definition, or insulate only the definition while putting contract conditions on a redeclaration - visible to static analysis tools in all TUs.||0||11||13||6||0.833333|
|lib.maintenance.noconfig||Library Provider||Simplify maintenance||Not require extra build steps to be documented||0||14||10||6||0.733333|
|lib.maintenance.nowhining||Library Provider||Simplify maintenance||Not have users complain about my product due to modifications of annotations resulting from their build configuration||0||17||9||4||0.566667|
|lib.integration.noconfig||Library Provider||Support successful integration||Not require extra build steps to be learned or performed||0||13||13||4||0.700000|
|lib.integration.nowhining||Library Provider||Support successful integration||Not have my users accidentally modify my careful annotations||0||12||11||7||0.833333|
|arch.nomacros||Technical Architect||Maintain quality of code base||Express assertions in a way that does not rely on C macros (i.e., there is no valid technical reason for a programmer not to use the new way, including space, time, tooling, and usability/complexity reasons, compared to C's assert macro)||0||7||13||10||1.100000|
|arch.complete||Technical Architect||Have a consistent and holistic contracts facility||Specify preconditions/postconditions/assertions/invariants that express my expectations about the expected valid state of my program in the form of compilable boolean expressions, that can be checked statically or dynamically (as opposed to disjointed state where these features are factored into bits)||1||3||12||14||1.379310|
|hardware.performance||Hardware Architect||Improve system-level performance||Be able to design new hardware + optimizations, carefully dovetailed into one another, that depend on statically-unprovable facts being annotated in the code||2||8||10||10||1.071429|
|sdev.bestpractices||Senior Developer||Set an example||Demonstrate best practice in defensive programming||0||3||9||18||1.500000|
|sdev.quality||Senior Developer||Enforce code quality||Discourage reliance on observable out-of-contract behavior by causing check failure to hard stop program or build||0||8||6||16||1.266667|
|sdev.maturity||Senior Developer||Enforce mature, finalized contracts||Disable continuation on violation of stable and correct individual contracts||0||4||7||19||1.500000|
|sdev.control||Senior Developer||Enforce mature, finalized contracts||Disable remapping of semantics on stable and correct individual contracts||0||9||12||9||1.000000|
|jdev.understand.contracts||Junior Developer||Understand the API||A uniform, fluent description of expected input values, expected output values, side effects, and all logical pre and post conditions||0||1||15||14||1.433333|
|jdev.understand.violations||Junior Developer||Understand the API||Be informed when my usage is out of contract||0||0||9||21||1.700000|
|jdev.understand.buildfailures||Junior Developer||Understand the program||Know why my software is not building||0||3||6||21||1.600000|
|jdev.understand.aborting||Junior Developer||Understand the program||Know why my software is aborting||0||1||9||20||1.633333|
|jdev.understand.omniscience||Junior Developer||Understand the program||Know why my software is out of contract||0||1||11||18||1.566667|
|jdev.understand.buildviolation||Junior Developer||Understand the program||Know that my program or build was halted due to contract violation||0||0||7||23||1.766667|
|jdev.understand.all||Junior Developer||Understand the facility||Be able to build a program with contracts after reasonably short tutorial||0||4||12||14||1.333333|
|jdev.understand.keywords||Junior Developer||Understand the facility||Have keywords with precise and unambiguous meanings||0||2||13||15||1.433333|
|jdev.bestpractices||Junior Developer||Improve my code||Learn about software best practices by example||0||4||12||14||1.333333|
|adev.fast||Agile Developer||Iterate quickly||Be able to write and modify contracts quickly without heavy boiler plate or up front cost||0||1||14||15||1.466667|
|adev.evolve||Agile Developer||Safeguard evolving code||Assert against conditions I am aware of but not finished handling fully||0||8||12||10||1.066667|
|bdev.confidentiality||Business Developer||Maintain confidentiality||Not expose diagnostic information (source location, expressions, etc.) in the software I deliver to clients, even when I choose to have contracts enforced in the software I deliver||0||14||11||5||0.700000|
|pdev.speed||Performance Sensitive Developer||Enable better performance||Annotate my code with assumptions, likelihoods, or reachability information that a tool might not be able to deduce, but that I would be confident of||0||7||10||13||1.200000|
|pdev.morespeed||Performance Sensitive Developer||Enable better performance||Be able to give statically-unprovable facts to current and novel optimizers in terms of semantics my program does not depend-on but optimizers can't figure out||0||6||11||13||1.233333|
|pdev.footgun||Performance Sensitive Developer||Enable better performance||Accept responsibility for a malformed program that might result from eventually false information given by my annotations||0||7||8||15||1.266667|
|pdev.safety.isolation||Performance Sensitive Developer||Have safety critical paths||Isolate safety checks from performance annotations||0||7||11||12||1.166667|
|pdev.safety.critical||Performance Sensitive Developer||Have safety critical paths||Retain checking even when optimizing with performance annotations||0||8||8||14||1.200000|
|qdev.checkall||Quality Sensitive Developer||Enable full checking||Ensure all checks (pre, post, assert, invariant) are enabled||0||2||10||18||1.533333|
|qdev.correctness||Quality Sensitive Developer||Validate correctness||Signify the predicates that should be verified by an analysis tool||0||6||19||5||0.966667|
|qdev.tooling||Quality Sensitive Developer||Manage multiple tools||Signify subset of individual annotations to be consumed by a specific kind of verification tool||0||14||14||2||0.600000|
|qdev.tooling.control||Quality Sensitive Developer||Manage multiple tools||Signify subset of individual annotations to be consumed by a specific instance of verification tool||0||14||15||1||0.566667|
|qdev.tooling.undefined||Quality Sensitive Developer||Manage multiple tools||Use predicates that may not be understood by all instances of verification||0||9||12||9||1.000000|
|qdev.tooling.undefinedkinds||Quality Sensitive Developer||Manage multiple tools||Use predicates that may not be understood by all kinds of verification||0||8||12||10||1.066667|
|qdev.tooling.behavior||Quality Sensitive Developer||Manage multiple tools||Integrate the results of that static checker into how my program behaves in different ways: assume proven predicates, make unprovable predicates ill- formed, etc.||0||15||11||4||0.633333|
|qdev.testing||Quality Sensitive Developer||Unit test predicates||Override failure handler to trigger test failure instead of termination||0||8||9||13||1.166667|
|qdev.handler.testing||Quality Sensitive Developer||Unit test violation handlers||Have a way to run handler on all combinations of available build modes||0||11||13||6||0.833333|
|qdev.fuzz.testing||Quality Sensitive Developer||Catch unexpected failure modes||Log all predicate failure during fuzz testing||0||5||14||11||1.200000|
|crit.control||Critical Software Developer||Have a verifiable release system||Be able to control the configuration of contracts from a central point||0||4||15||11||1.233333|
|crit.noundef||Critical Software Developer||Avoid undefined behavior||Have contract violation at run-time always have well-defined behavior||0||17||4||9||0.733333|
|crit.recovery||Critical Software Developer||Not have a faulty program lead to catastrophic failure||Have access to a recovery path after contract violation||0||15||8||7||0.733333|
|crit.redundancy||Critical Software Developer||Not have a faulty program lead to catastrophic failure||Be able to express error handling that may be redundant with contract checking||0||11||8||11||1.000000|
|crit.interaction||Critical Software Developer||Not have a faulty program lead to catastrophic failure||Not have contract build or run modes possibly be able to change or disable related error handling in any way||0||13||5||12||0.966667|
|crit.locality||Critical Software Developer||Be assured a critical violation uses a critical recovery path||Couple recovery path to a specific contract within the source||0||19||6||5||0.533333|
|crit.testing||Critical Software Developer||Meet code coverage requirements||Be able to run both success and failure branches in my test environment||0||8||11||11||1.100000|
|crit.production.checking||Critical Software Developer||Have redundant layering||Be able to continue to run checks in a production environment (even after formal testing is complete)||1||6||8||15||1.310345|
|crit.more.coverage||Critical Software Developer||Maximize coverage||Be able to run checks in a production environment that are considered "cheap" compared to the expected cost of entering an invalid state||0||6||13||11||1.166667|
|crit.noassume||Critical Software Developer||Avoid unexpected or undefined behavior||Ensure checks will never be __assume'd/__builtin_assume'd by the compiler as if they were facts injected into the program (otherwise, if such an assumption ever failed, I would be running a different program that is not equivalent to the one I wrote; assumptions can expand the set of possible executions by injecting facts not otherwise knowable to the compiler)||0||13||5||12||0.966667|
|sec.noattacks||Security Sensitive Developer||Limit attack vectors||Be unable to insert code paths (eg. violation handlers) at run time (eg. build time only)||1||8||9||12||1.137931|
|sec.certify||Security Sensitive Developer||Deliver a certified product||Have build tool only link to a preapproved violation handler||0||12||15||3||0.700000|
|analysis.runtime||User of Analysis Tools||Improve runtime correctness||Have runtime checks generated by the tool||1||8||14||7||0.965517|
|analysis.optimization||User of Analysis Tools||Improve runtime performance||Have runtime optimizations generated by the tool||1||7||17||5||0.931034|
|analysis.symbolic||User of Analysis Tools||Allow symbolic analysis||Have symbolic proofs for soundness and consistency performed before compile time||1||5||20||4||0.965517|
|analysis.compiletime||User of Analysis Tools||Allow code analysis||Have code source, AST, or instruction inspection during compile time||1||8||17||4||0.862069|
|analysis.binaries||User of Analysis Tools||Allow binary analysis||Have binary inspection after compile time||1||17||10||2||0.482759|
|analysis.information||User of Analysis Tools||Improve the quality of analysis||Be able to hint to the analyzer information it may be unable to deduce from source code alone (eg. 5 / opaque(); [[ opaque() != 0]])||1||6||17||6||1.000000|
|analysis.legacy||Provider of Analysis Tools||Extend my existing engine||Be able to map pre-existing contract features in tools to a standardized language syntax||1||14||10||5||0.689655|
|teach.bestpractices||Teacher||Demonstrate best practice||Be able to express defensive programming, programming by contract, and test driven development to introductory students||0||2||13||15||1.433333|
|test.standardized||Teacher||Demonstrate best practice||Not rely on custom libraries or proprietary extensions||0||3||12||15||1.400000|
|teach.lifecycle||Teacher||Demonstrate best practice||Demonstrate mock lifecycle by switching simple compiler flags to control which checks are enabled||0||5||19||6||1.033333|
|teach.portable||Teacher||Manage many students||Have examples compilable by a standard compiler on any system||0||4||3||23||1.633333|
|teach.dumbstudents||Teacher||Manage many students||Have examples that are easy to build without digression into build systems||0||6||9||15||1.300000|
|teach.teachable||Teacher||Build layers of understanding||Have simple explanation of assertions and their use to support simple programming tasks, including debugging erroneous programs.||0||3||9||18||1.500000|
|teach.layering||Teacher||Build layers of understanding||Support the ability for advanced uses of contracts to be distributed across many different courses in a C++-focused computer science curriculum.||0||6||15||9||1.100000|
|compiler.benice||Compiler Developer||Deliver best experience to my customers||Maximize implementation freedom by limiting what is strictly required by the standard||0||12||12||6||0.800000|
|compiler.best||Compiler Developer||Deliver the best implementation||Have a clear and simple specification that meets clear need||0||1||10||19||1.600000|
|large.complex||Large Codebase Developer||Debug complex issues||Have composable and fine grained control over which checks are run, without requiring source code changes. Specifically the checks for only one function or some grouping of functions||0||5||14||11||1.200000|
|large.critical||Large Codebase Developer||Enable/Disable checking on critical/hot paths||Control whether checks are run based on where they are being called from||0||9||13||8||0.966667|
|large.modernize||Large Codebase Owner||Modernize my code base||Introduce standardized contracts to replace my macro-based contracts||0||1||12||17||1.533333|
|large.stillmacros||Large Codebase Owner||Modernize my code base||Have my existing macro-based facilities interoperate smoothly with standardized contracts so I can do the migration gradually||0||5||16||9||1.133333|
|large.observation||Large Codebase Owner||Introduce new contracts into an existing system||Have failed individual checks from existing code optionally warn instead of hard stop||0||7||9||14||1.233333|
|large.introduction||Large Codebase Owner||Introduce new contracts into an existing system||Have failed checks from a new library optionally warn instead of hard stop||0||7||9||14||1.233333|
|large.separability||Large Codebase Owner||Introduce new parameters or invariants into a contracts based system||Be able to include distinct clauses for each parameter or invariant with their own individual failure or build controls||0||17||8||5||0.600000|
|large.newenvironment||Large Codebase Owner||Introduce new elements into a contracts based system||Have failed checks caused by a change in environment optionally warn instead of hard stop||0||10||13||7||0.900000|
|large.newcompiler||Large Codebase Owner||Introduce new elements into a contracts based system||Have failed checks caused by a change in compiler optionally warn instead of hard stop||0||10||15||5||0.833333|
|large.nogoingback||Large Codebase Owner||Prevent regressions||Have trusted contracts fail fast and hard stop||0||2||11||17||1.500000|
|large.scalability||Large Codebase Owner||Scale violation handling||Be able to log violations in my organization specific format||0||5||11||14||1.300000|
|large.simulation.disable||Large Codebase Owner||Allow simulation or post-mortem testing of known failure modes||Optionally disable checking on a subset of individual annotations||0||10||10||10||1.000000|
|large.simulation.enable||Large Codebase Owner||Allow simulation or post-mortem testing of known failure modes||Optionally allow checking of a subset of individual annotations to fail and access its recovery path||0||13||9||8||0.833333|
|large.simulation.ignore||Large Codebase Owner||Allow simulation or post-mortem testing of known failure modes||Optionally allow checking of a subset of individual annotations to fail and continue failing||0||14||9||7||0.766667|
|large.perfcontrol.build||Large Codebase Owner||Manage performance cost||Constrain the set of built time checks according to their performance overhead||0||13||12||5||0.733333|
|large.perfcontrol.runtime||Large Codebase Owner||Manage performance cost||Constrain the set of runtime checks according to their performance overhead||0||6||11||13||1.233333|
|large.narrowing||Large Codebase Owner||Tune contract width in complex system||Be able to narrow individual contract so it fails in testing not in production||0||11||13||6||0.833333|
|embedded.nochecking||Small Machine Developer||Minimize executable footprint||Remove all checking and diagnostic (eg. source location) overhead entirely from the final binary||0||4||7||19||1.500000|
|embedded.nologging||Small Machine Developer||Minimize executable footprint||Remove all logging and diagnostic (but not checking) overhead from the final binary||1||4||14||11||1.241379|
|embedded.minimize||Small Machine Developer||Minimize executable footprint||Remove all but the most important diagnostic overhead from the final binary||0||7||14||9||1.066667|
|wg21.everythingelse||Language Developer||Interoperate with Contracts||Have a clear way to understand how contracts will interact with the standard library||1||4||11||14||1.344828|
|wg21.otherfeatures||Language Developer||Extend contracts beyond pre/post conditions on functions||Be able to use contract-like syntax on past or present runtime checkable language features such as switches, pattern matching, etc. or what might happen on signed integer overflow, etc. This might allow configuration of trapping, logging, or assuming in other areas of language UB.||0||12||14||4||0.733333|
This general use case expresses the desire to place information about a program's expected execution state in many different places throughout the program - possibly including "upon function entry", "whenever this line of code is executed", "at all times when a class of this type is not actively executing a member function", or others.
This high level use case expresses the need to attach information to contract conditions that is beyond the statement of the condition itself, and instead has user-provided metadata about both the confidence in the condition and the desired behavior of the program in relation to that condition.
This high level use case expresses the desire to have metadata on contract conditions associated with what downsides violations might have, perhaps indicating that while a certain condition is expected to be met, the library does guarantee that the downsides will not be catastrophic.
This use case expresses the desire to have metadata about contract conditions that capture at least 3 (if not more) granularities of "cost" to be used as input in some way to other decisions about what the contracts might do or how they might be interpreted.
The desire to have behaviors (such as generating runtime checks or optimizations) and the expected behavior explicitly defined by the standard in order to help users reason about what a program will do/has done when a contract is violated. Note that having this in the standard allows for leveraging expectations of behavior across all compliant platforms.
This high-level use case expresses a desire that any execution effective transformations an annotation might apply be "reasonable", ie. not surprising or counter intuitive (such as time-travel optimizations).
There is a tension between allowing side effects in contract conditions and disallowing them completely. In general, writing code with absolutely no side effects is very hard, and there are pitfalls if the language is actively hostile to accidental side effects. On the other hand, conditions with side effects are also not elidable by the compiler, since the act of checking them is observable.
This encompasses some specific need to control, either when writing code or when building code, what behavior (if any) is associated with a contract condition.
Any keywords chosen for use within the feature should be easily distinguishable (to avoid name churn such as expects/ensures to pre/post) and be very clearly matched to what they will do within the language (unlike, for example, axiom).
This use case indicates a preference to have any metadata about a contract be visually very minimal or come after the predicate itself.
Rather than requiring special functions or types (or some completely new thing), contract should be able to leverage any program logic related to their statement that is already written/writable in C++.
This as a preference to needing to specify contracts in a completely separate metadata file of some sort.
The desire is to not require use of macros in order to satisfy the majority of use cases.
All contract control features and behaviors should be interoperable in a reasonable way with any partially or fully modularized C++ program.
Contracts on coroutines open up a number of new situations to consider because there might be requirements on what the state of the program is whenever a coroutine resumes execution, and there might be promises a coroutine makes prior to each time it suspends.
In some shape or form, anything documented in the contracts of the standard library's functions (preconditions, postconditions, other behaviors) is a candidate for something that should be codifiable as a contract condition.
When handling a contract violation, the ability to trigger a debugger when possible is currently not standardized but should be available.
Note that "debug" and "release" are not standardized things, but nothing specified in the standard should preclude those from having some impact on what behaviors contracts take on in those two modes. It should be possible for a vendor to map contract modes to their existing native modes.
Important considerations for C are related to any contracts that can go on a normal C function. C does have attributes, but it also explicitly calls out that a conforming implementation can ignore all attributes, as opposed to C++ which has made that a commonly held assumption but has not actually put wording in the standard to that effect.
Adding new identifiers with meaning in C is generally not acceptable for standardization, so compatible contracts would either need to use no contracts or support alternate reserved words for use with contracts.
Importantly, satisfying this requirement would mean making the the argument to a violation handler meaningful in both C and C++, or requiring platform vendors to shim between the two (in the case of a C violation handler receiving a violation from C++ code).
Assuming contracts continue to be rendered as attributes, C standardization would require they be semantically optional.
Note that, just as with C++, the ':' in the previous contract syntax does not match the grammar for attributes in either language, so by a strict reading of the standards there is no obligation to be ignorable. ('[[a:b]]' is not a valid attribute and should be diagnosed as invalid on any current C or C++ compiler). Many have expressed the view that this opinion is pedantic and that the spirit of the law is that anything between []s should be ignorable.
Additionally, it is currently a conforming extension to throw away all tokens between a pair of []s, and there exist numerous compilers that take advantage of that fact which would be broken by requiring behavior of any constructs that look like an attribute.
Class invariants have historically been considered of general use, but the performance impact of checking them can be surprisingly huge. Previously these have been left out of proposals for C++ to be added in at a later date.
This establishes the basic requirement on preconditions in terms of input values.
The more detail available to users about how a contract condition has been violated the more useful they become. This means each specific condition is benefited by being separate (eg. 'x != 0' and 'y != 0' as distinct conditions instead of requiring that users use 'x != 0 && y != 0').
In addition to that, anything that might be able to capture those values and expose them to violation handlers for logging would again help benefit problem diagnosis.
This establishes the basic requirement to use predicates to evaluate input values.
This establishes the basic requirement for invariants in terms of predicates on class members.
This establishes the basic requirement to use predicates to evaluate return values.
This establishes the basic requirement to reference in-scope variables in order capture values and compute predicates.
Postconditions often need to state things about how values of output parameters have changed, or how the values of other global state might have changed as a result of a function call. The ability to store copies of state from before the function call and reference that in a postcondition predicate will enable a wider variety of conditions to be formulatable.
Ada has this functionality builtin by providing ways to reference explicitly the original value of a function parameter. C++ makes this more complicated with the need to consider the handling of move-only or generally non-copyable types. More importantly, such copying should absolutely not happen if a contract is not being evaluated at runtime.
Capturing the pre-function state of member variables might also be needed to state the contracts on many member functions.
Capturing arbitrary global state for use in a postcondition might prove useful.
This extends to the state of arbitrary other expressions and how that might change
due to the invocation of a function. Consider this example of what might
be a postcondition of a typical
old(now()) <= now() + sleep_time
Exception safety guarantees are often part of the English language contract of a function, and being able to state that (in a way that tools might be able to then pick up on and verify) would be useful.
The proper use of a type in a multithreaded environment can benefit greatly from being documented in a way that can then be checked and verified.
Establishing more details about what might happen on failure, such as what external state might be changed in a remote database, is another useful condition to state.
Numerous predicates that might be useful to state also might change the state of a program in a paradoxical way if checked. The common example (originally put forward in P0380) is validating available items on an input iterator. Other important examples include checks that might require accessing special hardware, or doing extensive computations that would invalidate the results just by taking the time to do them.
These predicates are still useful for those reading the code, still might benefit from being validated against postconditions of other functions, and might provide some identifiable code improvements when analyzed by the compiler without actually executing them at runtime.
Often during development it helps to write contracts first, and that might even predate having enough of an implementation to fully define the conditions. Having them in code ensures that the placeholder API to check them is maintained. Some functions may never be implementable due to time constraints, while others will eventually be filled in as time allows.
Some predicates might even have no meaning within the language itself, but benefit other tools by being placed into the same contract framework as other predicates, with meaning to those external tools.
Many external static analysis tools, and even compilers, can check that some state is being handled properly which cannot be properly validated within the language itself. Consider functions such as "is_deletable" or "is_reachable".
One common realization with having checked contracts in the language is that they will then identify bugs that exist in programs. Identifying the source of those bugs swiftly is important, and for any contract that is dependent on a function being called properly that responsibility does not usually lie with the line of code where the contract is written, but with the place where the function has been called.
Some function preconditions can be expressed as c++ expressions, but are runtime-expensive. However, if expressed inside function body as assertions after some function logic has already been executed, the check whether the function has everything that is required is cheap. If I use an assertion like this, I want to annotate this assertion as being "precondition-like". The semantics would be that if it fails, it is reported as "caller's bug" rather than callee's bug, and if the build is configured to runtime-check only preconditions, such assertion should also be runtime-checked.
As an example, consider a binary search that wishes to check that the input list is sorted. There is an expensive check (O(n)) that can be done to verify that the whole list is sorted on each call, but this would negate the performance benefits of doing a binary search to begin with. Another option is to check just that the elements adjacent to those that are visited are in properly sorted order. This will provide some chance of identify unsorted inputs without making performance unusable, but is also most cleanly done during the execution of the search algorithm. Unlike a normal in-body assertion, a violation identified by this kind of check should be reported to the caller, as it is a failure on the calling code's part to provide a properly sorted input list.
Putting contracts on the first function declaration makes them easily visible to anyone looking at a header file for understanding an API.
Often, an interface might be very narrow - consider the interface on an iterator, which generally just has a few operators and no other functions - yet there might be ways to check parts of a narrow contract that depend on internal state. Exposing this state so that it can be checked publicly would require widening the API interface, which is counter to the desire to keep the use of a type as simple as possible.
On a similar note, sometimes the internal state only approximates the actual contract, and while better than no checking there is a desire to not expose that data publicly so that there is no assumption that the data represents the complete set of requirements on calling code.
Many development methodologies revolve around making readable, well-documented header files to provide a way to understand API usage. Some contracts might be simple and readable, but others might involve complex conditions that distract from readability, and match very simple to read and understand prose contracts in the documentation.
The benefits (checkability, analysis, etc.) of having the encoded contracts should not come at the cost of making the primary visible public interface (the header file) more unreadable.
Historically, rather than providing functions with narrow contracts, software might have been written to report misuse through exceptions. Contracts should provide a better alternative than giving all functions a fully wide contract.
P4810 did not allow any variation in contract conditions through overrides of a virtual function, though this could be roughly accomplished with assertions in function bodies instead.
In general, the pre and postconditions of the interface you call through should be checked to be sure you're meeting the requirements you want to meet. Similarly, the concrete type's conditions should be checked to be sure it is behaving properly. Doing both checks if they are different might be an acceptable solution to allow flexibility without any increase in risk, at the possible cost of some performance in checked builds.
Requirements on an interface to a function should be impossible when entering into a type's functions, but there might be a desire to distinguish that so that checking is done differently when a type calls its own functions.
General class invariants should be checked for validity before and after any call to a public member function of a class. Calls to friend functions need to be considered as well, as they might benefit from invalidating/validating invariants. Similarly, invariants might be checked for all objects of the type that are being touched whenever a public function is starting or finishing.
Calling from a member function to a function of another type (or a possible subtype) might require verifying that the object's invariants hold prior to calling the function and that they still hold when the other function returns.
Similar to the public API of a class, the protected API is exposed to subclasses and invariants of that API should be checked whenever the boundary might be crossed.
In general, any callback based solution exposes functions that will be called that might benefit from having information about their contracts also transmitted.
This might be implemented by incorporating contracts deeply into the type system, or in a more limitted way by just facilitating library types like a function with contracts.
Generally, postconditions have been assumed to hold when a function returns normally. Expressing conditions that will hold if a function returns through an exception, or both, could also be useful.
Fold expressions using && might be useful for a predicate, but they then lose any information about which particular argument might have violated the condition. Enabling more useful information to narrow that down would be helpful.
Like regular functions, pre and postconditions on coroutines should have similar semantics.
Similar to a class, the state of a coroutine, or what it expects of the state of the world as it suspends and resumes, might be stated as expected invariants that could be checked whenever entering and exiting.
Often it is helpful to limit what gets enabled to functions called from some subset of a program, and the rest of the program might not perform acceptably enough to test if hose same functions are always checked.
Contract conditions on only the first declaration (as proposed by P4810) mean that the condition clutters the readable interface and changes in a contract condition force client recompilation. Contract conditions not in the header file are not going to be visible to tools that are attempting to limit what they consume to a single translation unit. Combined, these result in a desire to sometimes put the contracts with the implementing definition (in a .cpp file) and sometimes put them in a redeclaration in a header file (where one might also put inline function definitions).
For free functions this would just require relaxing the requirements on where contracts are. For member functions, allowing member function redeclaration would be needed to put the contracts in a less obtrusive place that is still visible outside of the defining TU. See P1320 for more discussion of this.
Thanks to everyone who has participated in the SG21 reflector, telecon, and our first meeting in Belfast. Herb Sutter contributed greatly by taking the vast raw data that made up our use cases and painfully populated the online survey that everyone participated in. Thanks also to all 30 respondents who took a significant portion of their personal time to read, understand, and express their opinions on the 195 use cases presented here.