Update build by processing changes using fine-grained dependencies. Use fine-grained dependencies to update targets in other modules that may be affected by externally-visible changes in the changed modules. This forms the core of the fine-grained incremental daemon mode. This module is not used at all by the 'classic' (non-daemon) incremental mode. Here is some motivation for this mode: * By keeping program state in memory between incremental runs, we only have to process changed modules, not their dependencies. The classic incremental mode has to deserialize the symbol tables of all dependencies of changed modules, which can be slow for large programs. * Fine-grained dependencies allow processing only the relevant parts of modules indirectly affected by a change. Say, if only one function in a large module is affected by a change in another module, only this function is processed. The classic incremental mode always processes an entire file as a unit, which is typically much slower. * It's possible to independently process individual modules within an import cycle (SCC). Small incremental changes can be fast independent of the size of the related SCC. In classic incremental mode, any change within a SCC requires the entire SCC to be processed, which can slow things down considerably. Some terms: * A *target* is a function/method definition or the top level of a module. We refer to targets using their fully qualified name (e.g. 'mod.Cls.method'). Targets are the smallest units of processing during fine-grained incremental checking. * A *trigger* represents the properties of a part of a program, and it gets triggered/fired when these properties change. For example, '<mod.func>' refers to a module-level function. It gets triggered if the signature of the function changes, or if the function is removed, for example. Some program state is maintained across multiple build increments in memory: * The full ASTs of all modules are stored in memory all the time (this includes the type map). * A fine-grained dependency map is maintained, which maps triggers to affected program locations (these can be targets, triggers, or classes). The latter determine what other parts of a program need to be processed again due to a fired trigger. Here's a summary of how a fine-grained incremental program update happens: * Determine which modules have changes in their source code since the previous update. * Process changed modules one at a time. Perform a separate full update for each changed module, but only report the errors after all modules have been processed, since the intermediate states can generate bogus errors due to only seeing a partial set of changes. * Each changed module is processed in full. We parse the module, and run semantic analysis to create a new AST and symbol table for the module. Reuse the existing ASTs and symbol tables of modules that have no changes in their source code. At the end of this stage, we have two ASTs and symbol tables for the changed module (the old and the new versions). The latter AST has not yet been type checked. * Take a snapshot of the old symbol table. This is used later to determine which properties of the module have changed and which triggers to fire. * Merge the old AST with the new AST, preserving the identities of externally visible AST nodes for which we can find a corresponding node in the new AST. (Look at mypy.server.astmerge for the details.) This way all external references to AST nodes in the changed module will continue to point to the right nodes (assuming they still have a valid target). * Type check the new module. * Take another snapshot of the symbol table of the changed module. Look at the differences between the old and new snapshots to determine which parts of the changed modules have changed. The result is a set of fired triggers. * Using the dependency map and the fired triggers, decide which other targets have become stale and need to be reprocessed. * Create new fine-grained dependencies for the changed module. We don't garbage collect old dependencies, since extra dependencies are relatively harmless (they take some memory and can theoretically slow things down a bit by causing redundant work). This is implemented in mypy.server.deps. * Strip the stale AST nodes that we found above. This returns them to a state resembling the end of semantic analysis pass 1. We'll run semantic analysis again on the existing AST nodes, and since semantic analysis is not idempotent, we need to revert some changes made during semantic analysis. This is implemented in mypy.server.aststrip. * Run semantic analyzer passes 2 and 3 on the stale AST nodes, and type check them. We also need to do the symbol table snapshot comparison dance to find any changes, and we need to merge ASTs to preserve AST node identities. * If some triggers haven been fired, continue processing and repeat the previous steps until no triggers are fired. This is module is tested using end-to-end fine-grained incremental mode test cases (test-data/unit/fine-grained*.test).
| Class | |
Undocumented |
| Class | |
No class docstring; 0/12 instance variable, 6/6 methods documented |
| Class | |
Undocumented |
| Function | calculate |
Determine activated triggers by comparing old and new symbol tables. |
| Function | dedupe |
Undocumented |
| Function | delete |
Undocumented |
| Function | ensure |
Ensure that the dependencies on a module are loaded. |
| Function | ensure |
Ensure that the modules in initial and their deps have loaded trees. |
| Function | extract |
Undocumented |
| Function | extract |
Undocumented |
| Function | find |
Find a module in a list that directly imports no other module in the list. |
| Function | find |
Find all nested symbol tables. |
| Function | find |
Find names of all targets that need to reprocessed, given some triggers. |
| Function | find |
Find all the deps of the nodes in initial that haven't had their tree loaded. |
| Function | get |
Undocumented |
| Function | get |
Undocumented |
| Function | is |
Undocumented |
| Function | lookup |
Look up a target by fully-qualified name. |
| Function | propagate |
Transitively rechecks targets based on triggers and the dependency map. |
| Function | refresh |
Look for submodules that are now suppressed in target package. |
| Function | replace |
Replace modules with newly builds versions. |
| Function | reprocess |
Reprocess a set of nodes within a single module. |
| Function | sort |
Sort messages so that the order of files is preserved. |
| Function | target |
Return the target name corresponding to a deferred node. |
| Function | update |
Undocumented |
| Function | update |
Build a new version of one changed module only. |
| Constant | INIT |
Undocumented |
| Constant | MAX |
Undocumented |
| Constant | SENSITIVE |
Undocumented |
| Type Alias | |
Undocumented |
BuildManager, old_snapshots: dict[, new_modules: dict[) -> set[:
(source)
¶
Determine activated triggers by comparing old and new symbol tables. For example, if only the signature of function m.f is different in the new symbol table, return {'<m.f>'}.
Ensure that the dependencies on a module are loaded. Dependencies are loaded into the 'deps' dictionary. This also requires loading dependencies from any parent modules, since dependencies will get stored with parent modules when a module doesn't exist.
BuildManager, graph: dict[, initial: Sequence[):
(source)
¶
Ensure that the modules in initial and their deps have loaded trees.
list[, graph: Graph) -> tuple[:
(source)
¶
Find a module in a list that directly imports no other module in the list. If no such module exists, return the lexicographically first module from the list. Always return one of the items in the modules list. NOTE: If both 'abc' and 'typing' have changed, an effect of the above rule is that we prefer 'abc', even if both are in the same SCC. This works around a false positive in 'typing', at least in tests. Args: modules: List of (module, path) tuples (non-empty) graph: Program import graph that contains all modules in the module list
str, symbols: SymbolTable) -> dict[:
(source)
¶
Find all nested symbol tables. Args: prefix: Full name prefix (used for return value keys and to filter result so that cross references to other modules aren't included) symbols: Root symbol table Returns a dictionary from full name to corresponding symbol table.
BuildManager, graph: Graph, triggers: set[, deps: dict[, up_to_date_modules: set[) -> tuple[:
(source)
¶
Find names of all targets that need to reprocessed, given some triggers. Returns: A tuple containing a: * Dictionary from module id to a set of stale targets. * A set of module ids for unparsed modules with stale targets.
BuildManager, graph: dict[, initial: Sequence[) -> list[:
(source)
¶
Find all the deps of the nodes in initial that haven't had their tree loaded. The key invariant here is that if a module is loaded, so are all of their dependencies. This means that when we encounter a loaded module, we don't need to explore its dependencies. (This invariant is slightly violated when dependencies are added, which can be handled by calling find_unloaded_deps directly on the new dependencies.)
FileSystemCache, modules: dict[, changed_modules: list[, followed: bool) -> list[:
(source)
¶
Undocumented
BuildManager, target: str) -> tuple[:
(source)
¶
Look up a target by fully-qualified name. The first item in the return tuple is a list of deferred nodes that needs to be reprocessed. If the target represents a TypeInfo corresponding to a protocol, return it as a second item in the return tuple, otherwise None.
BuildManager, graph: dict[, deps: dict[, triggered: set[, up_to_date_modules: set[, targets_with_errors: set[, processed_targets: list[) -> list[:
(source)
¶
Transitively rechecks targets based on triggers and the dependency map. Returns a list (module id, path) tuples representing modules that contain a target that needs to be reprocessed but that has not been parsed yet. Processed targets should be appended to processed_targets (used in tests only, to test the order of processing targets).
str, path: str|None, deps: dict[, graph: Graph, fscache: FileSystemCache, refresh_file: Callable[) -> list[:
(source)
¶
Look for submodules that are now suppressed in target package. If a submodule a.b gets added, we need to mark it as suppressed in modules that contain "from a import b". Previously we assumed that 'a.b' is not a module but a regular name. This is only relevant when following imports normally. Args: module: target package in which to look for submodules path: path of the module refresh_file: function that reads the AST of a module (returns error messages) Return a list of errors from refresh_file() if it was called. If the return value is None, we didn't call refresh_file().
BuildManager, graph: dict[, old_modules: dict[, new_modules: dict[):
(source)
¶
Replace modules with newly builds versions. Retain the identities of externally visible AST nodes in the old ASTs so that references to the affected modules from other modules will still be valid (unless something was deleted or replaced with an incompatible definition, in which case there will be dangling references that will be handled by propagate_changes_using_dependencies).
BuildManager, graph: dict[, module_id: str, nodeset: set[, deps: dict[, processed_targets: list[) -> set[:
(source)
¶
Reprocess a set of nodes within a single module. Return fired triggers.
list[, prev_messages: list[) -> list[:
(source)
¶
Sort messages so that the order of files is preserved. An update generates messages so that the files can be in a fairly arbitrary order. Preserve the order of files to avoid messages getting reshuffled continuously. If there are messages in additional files, sort them towards the end.
str, node: (FuncDef|MypyFile)|OverloadedFuncDef) -> str|None:
(source)
¶
Return the target name corresponding to a deferred node. Args: module: Must be module id of the module that defines 'node' Returns the target name, or None if the node is not a valid target in the given module (for example, if it's actually defined in another module).
str, nodes: list[, graph: dict[, deps: dict[, options: Options):
(source)
¶
Undocumented
str, path: str, manager: BuildManager, previous_modules: dict[, graph: Graph, force_removed: bool, followed: bool) -> UpdateResult:
(source)
¶
Build a new version of one changed module only. Don't propagate changes to elsewhere in the program. Raise CompileError on encountering a blocking error. Args: module: Changed module (modified, created or deleted) path: Path of the changed module manager: Build manager graph: Build graph force_removed: If True, consider the module removed from the build even it the file exists Returns a named tuple describing the result (see above for details).