Rf refactors Go programs. Usage: Rf applies a script of refactoring commands to the package in the current directory. For example, to unexport a field in a struct by renaming it: By default, rf writes changes back to the disk. The -diff flag causes rf to print a diff of the intended changes instead. A script is a sequence of commands, one per line. Comments are introduced by # and extend to the end of the line. Commands may be broken across lines by ending all but the last with a trailing backslash (before any comment), as in: Commands that take { } blocks need not backslash-escape line breaks inside the braces, as in: Most commands take “code addresses” as arguments. Each code address identifies some code in a program. For illustration, consider this program, prog.go: The simplest code address is the name of a top-level declaration. In this program, those addresses are C, D, F, T, V, and VT. Adding .Name to an address selects a name within the earlier address, whether that's a function variable (F.who, F.msg), a struct field (T.Field), a method (T.M, T.P), or a variable's field (V.Value, V.thing, V.thing.Field, VT.Field). Another kind of code address is a Go source file, identified by a name ending in “.go”, as in “file.go”. If the file name contains a slash, as in “../dir/file.go”, the address identifies a file in the package in “../dir”. Another kind of code address is a Go package, identified by a path containing a slash but not ending in “.go”, as in “../dir” or “example.com/pkg”. A final kind of code address is a textual range in a function body or source file, identified by the syntax Ident:Range, where Ident identifies a file, function, or method and Range identifies a section of text within. The Range syntax is as used in the Acme and Sam text editors. The most common forms are the line range “N,M”, the byte range “#N,M”, and the regular expression range “/re1/,/re2/”. For example: See http://9p.io/sys/doc/sam/sam.html Table II for details on the syntax. The add command adds text to the source code. It takes as an argument the address after which the text should be added, followed by the text itself. The address may be a declaration, text range, file, or package. In all cases, the text is added immediately after the addressed location: after the declaration, after the text range, at the end of the file, or at the end of the first file in the package (considering the file names in lexical order). Examples: The cp command is like mv (see below) but doesn't delete the source and doesn't update any references. (UNIMPLEMENTED) The ex command applies rewrites based on example snippets. The arguments to ex are interpreted as Go code consisting of a sequence of imports and then a list of special “old -> new” rules. Each rule specifies that where ex finds a pattern matching old, it should replace the code with new. For example, to replace all log.Error calls with log.Panic: Declarations introduce typed pattern variables that can be used in rules. For example, to simplify certain needlessly complex uses of fmt.Sprintf: The inline command inlines uses of declared constants, functions, and types. Each use of the named declarations is replaced by the declaration's definition. If the -rm flag is given, inline removes the declarations as well. Examples: Given the declarations in the “Code addresses” section above, the first command replaces all uses of E with 2.718281828. The second replaces all uses of TAlias with T and then removes TAlias. UNIMPLEMENTED: Inlining of functions. The key command converts all struct literals for a list of types to keyed literals. Each address must identify a struct type. All literals of those struct types are updated to use the keyed form. Example: The mv command moves and renames code. When mv moves or renames old code, it also updates any references to use the new names or locations for the code. This includes updating other packages in the current module. In general, mv aims to act appropriately for any sensible combination of old and new address form. The rest of this section enumerates the specific cases that mv handles. item → renamed item Any named item can be renamed by specifying a destination that is the same code address with the final element changed. For example: In this form, the destination address must repeat the dot-separated elements leading up to the new name, changing only the final element. The repetition here distinguishes this form from other forms. var → var field A top-level variable can be moved to a new or existing field in a global variable of struct type. For example: method → func A method can be moved to a top-level function, removing the association with the receiver type. The receiver remains the first argument of the new function. For example: func → method A function can be moved to a method on the type of its first argument, assuming that type is defined in the same package where the function appears. For example: UNIMPLEMENTED. code text → new function A text range can be moved to a new function, leaving behind an appropriate call to that function. For example: code text → new method A text range can be moved to a new method, leaving behind an approriate call to that method. For example: TODO. UNIMPLEMENTED. declaration → file If the source is a top-level declaration (const, func, method, type, var) and the destination is a file, mv moves that declaration, along with any comments immediately preceding it, to the end of the destination file. For example: Naming a single item in a declaration block is taken to indicate wanting to move all items in the block. In the example from the “Code addresses” section, “mv C x.go” moves D as well. Any time a destination file must be created, mv initializes it with the header comments (those above the package declaration and any package doc) from the file the source code is being moved from. This heuristic is meant to copy header text like copyright notices. The file may be in a different package. As usual, mv updates references to the moved declaration to refer to its new location, inserting imports as needed. If the result is an import cycle, mv reports the cycle rather than attempt some kind of automatic (and likely wrong) fix. declaration → package If the source is a top-level declaration and the destination is a package, mv moves the declaration to the a file in the destination package with the same name as the one holding the source item. If the destination package does not exist, it will be created. file → file If the source and destination are both files, mv moves all code from the source file to the end of the destination file. For example: file → package If the source is a file and the destination is a package, mv moves all code from the source file to a file in the destination with the same base name. For example: package → file If the source is a package and the destination is a file, mv moves all code from the source package to the named file. For example: UNIMPLEMENTED. package → package If the source is a package and the destination is a file, mv moves all code from the source package to the destination package. from the source package to the named file. For example: UNIMPLEMENTED. many → file, many → package If the destination is a file or package, multiple sources can be listed. The mv command moves each source item in turn to the destination. The rm command removes code. Rm deletes the old code. All address forms are valid. When deleting a declaration, rm also deletes line comments immediately preceding it, up to a blank line. UNIMPLEMENTED: removal of struct fields, interface methods, text ranges. Rf is very very rough. Everything is subject to change, and it may break your programs.
Readme
rf is an experimental refactoring tool. It is very much a work in progress.
rf is incredibly rough and likely to be buggy and change incompatibly.
FAQs
Rf refactors Go programs. Usage: Rf applies a script of refactoring commands to the package in the current directory. For example, to unexport a field in a struct by renaming it: By default, rf writes changes back to the disk. The -diff flag causes rf to print a diff of the intended changes instead. A script is a sequence of commands, one per line. Comments are introduced by # and extend to the end of the line. Commands may be broken across lines by ending all but the last with a trailing backslash (before any comment), as in: Commands that take { } blocks need not backslash-escape line breaks inside the braces, as in: Most commands take “code addresses” as arguments. Each code address identifies some code in a program. For illustration, consider this program, prog.go: The simplest code address is the name of a top-level declaration. In this program, those addresses are C, D, F, T, V, and VT. Adding .Name to an address selects a name within the earlier address, whether that's a function variable (F.who, F.msg), a struct field (T.Field), a method (T.M, T.P), or a variable's field (V.Value, V.thing, V.thing.Field, VT.Field). Another kind of code address is a Go source file, identified by a name ending in “.go”, as in “file.go”. If the file name contains a slash, as in “../dir/file.go”, the address identifies a file in the package in “../dir”. Another kind of code address is a Go package, identified by a path containing a slash but not ending in “.go”, as in “../dir” or “example.com/pkg”. A final kind of code address is a textual range in a function body or source file, identified by the syntax Ident:Range, where Ident identifies a file, function, or method and Range identifies a section of text within. The Range syntax is as used in the Acme and Sam text editors. The most common forms are the line range “N,M”, the byte range “#N,M”, and the regular expression range “/re1/,/re2/”. For example: See http://9p.io/sys/doc/sam/sam.html Table II for details on the syntax. The add command adds text to the source code. It takes as an argument the address after which the text should be added, followed by the text itself. The address may be a declaration, text range, file, or package. In all cases, the text is added immediately after the addressed location: after the declaration, after the text range, at the end of the file, or at the end of the first file in the package (considering the file names in lexical order). Examples: The cp command is like mv (see below) but doesn't delete the source and doesn't update any references. (UNIMPLEMENTED) The ex command applies rewrites based on example snippets. The arguments to ex are interpreted as Go code consisting of a sequence of imports and then a list of special “old -> new” rules. Each rule specifies that where ex finds a pattern matching old, it should replace the code with new. For example, to replace all log.Error calls with log.Panic: Declarations introduce typed pattern variables that can be used in rules. For example, to simplify certain needlessly complex uses of fmt.Sprintf: The inline command inlines uses of declared constants, functions, and types. Each use of the named declarations is replaced by the declaration's definition. If the -rm flag is given, inline removes the declarations as well. Examples: Given the declarations in the “Code addresses” section above, the first command replaces all uses of E with 2.718281828. The second replaces all uses of TAlias with T and then removes TAlias. UNIMPLEMENTED: Inlining of functions. The key command converts all struct literals for a list of types to keyed literals. Each address must identify a struct type. All literals of those struct types are updated to use the keyed form. Example: The mv command moves and renames code. When mv moves or renames old code, it also updates any references to use the new names or locations for the code. This includes updating other packages in the current module. In general, mv aims to act appropriately for any sensible combination of old and new address form. The rest of this section enumerates the specific cases that mv handles. item → renamed item Any named item can be renamed by specifying a destination that is the same code address with the final element changed. For example: In this form, the destination address must repeat the dot-separated elements leading up to the new name, changing only the final element. The repetition here distinguishes this form from other forms. var → var field A top-level variable can be moved to a new or existing field in a global variable of struct type. For example: method → func A method can be moved to a top-level function, removing the association with the receiver type. The receiver remains the first argument of the new function. For example: func → method A function can be moved to a method on the type of its first argument, assuming that type is defined in the same package where the function appears. For example: UNIMPLEMENTED. code text → new function A text range can be moved to a new function, leaving behind an appropriate call to that function. For example: code text → new method A text range can be moved to a new method, leaving behind an approriate call to that method. For example: TODO. UNIMPLEMENTED. declaration → file If the source is a top-level declaration (const, func, method, type, var) and the destination is a file, mv moves that declaration, along with any comments immediately preceding it, to the end of the destination file. For example: Naming a single item in a declaration block is taken to indicate wanting to move all items in the block. In the example from the “Code addresses” section, “mv C x.go” moves D as well. Any time a destination file must be created, mv initializes it with the header comments (those above the package declaration and any package doc) from the file the source code is being moved from. This heuristic is meant to copy header text like copyright notices. The file may be in a different package. As usual, mv updates references to the moved declaration to refer to its new location, inserting imports as needed. If the result is an import cycle, mv reports the cycle rather than attempt some kind of automatic (and likely wrong) fix. declaration → package If the source is a top-level declaration and the destination is a package, mv moves the declaration to the a file in the destination package with the same name as the one holding the source item. If the destination package does not exist, it will be created. file → file If the source and destination are both files, mv moves all code from the source file to the end of the destination file. For example: file → package If the source is a file and the destination is a package, mv moves all code from the source file to a file in the destination with the same base name. For example: package → file If the source is a package and the destination is a file, mv moves all code from the source package to the named file. For example: UNIMPLEMENTED. package → package If the source is a package and the destination is a file, mv moves all code from the source package to the destination package. from the source package to the named file. For example: UNIMPLEMENTED. many → file, many → package If the destination is a file or package, multiple sources can be listed. The mv command moves each source item in turn to the destination. The rm command removes code. Rm deletes the old code. All address forms are valid. When deleting a declaration, rm also deletes line comments immediately preceding it, up to a blank line. UNIMPLEMENTED: removal of struct fields, interface methods, text ranges. Rf is very very rough. Everything is subject to change, and it may break your programs.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
The June TC39 meeting wrapped up this week with eight proposals moving on to the next stage. Here's a quick roundup of the features that the committee approved to advance.
Security News
Cyber extortion in the US and Canada hit record levels in 2023, with ransomware attacks surging and median ransom demands skyrocketing, though fewer companies are choosing to pay ransoms.
Security News
Ecma TC39 is meeting this week and has moved key ECMAScript proposals forward, advancing Deferred Import Evaluation, Error.isError(), RegExp Escaping, and Promise.try to the next stages.
