Source file src/cmd/compile/doc.go

     1  // Copyright 2009 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  /*
     6  Compile, typically invoked as ``go tool compile,'' compiles a single Go package
     7  comprising the files named on the command line. It then writes a single
     8  object file named for the basename of the first source file with a .o suffix.
     9  The object file can then be combined with other objects into a package archive
    10  or passed directly to the linker (``go tool link''). If invoked with -pack, the compiler
    11  writes an archive directly, bypassing the intermediate object file.
    12  
    13  The generated files contain type information about the symbols exported by
    14  the package and about types used by symbols imported by the package from
    15  other packages. It is therefore not necessary when compiling client C of
    16  package P to read the files of P's dependencies, only the compiled output of P.
    17  
    18  Command Line
    19  
    20  Usage:
    21  
    22  	go tool compile [flags] file...
    23  
    24  The specified files must be Go source files and all part of the same package.
    25  The same compiler is used for all target operating systems and architectures.
    26  The GOOS and GOARCH environment variables set the desired target.
    27  
    28  Flags:
    29  
    30  	-D path
    31  		Set relative path for local imports.
    32  	-I dir1 -I dir2
    33  		Search for imported packages in dir1, dir2, etc,
    34  		after consulting $GOROOT/pkg/$GOOS_$GOARCH.
    35  	-L
    36  		Show complete file path in error messages.
    37  	-N
    38  		Disable optimizations.
    39  	-S
    40  		Print assembly listing to standard output (code only).
    41  	-S -S
    42  		Print assembly listing to standard output (code and data).
    43  	-V
    44  		Print compiler version and exit.
    45  	-asmhdr file
    46  		Write assembly header to file.
    47  	-asan
    48  		Insert calls to C/C++ address sanitizer.
    49  	-buildid id
    50  		Record id as the build id in the export metadata.
    51  	-blockprofile file
    52  		Write block profile for the compilation to file.
    53  	-c int
    54  		Concurrency during compilation. Set 1 for no concurrency (default is 1).
    55  	-complete
    56  		Assume package has no non-Go components.
    57  	-cpuprofile file
    58  		Write a CPU profile for the compilation to file.
    59  	-dynlink
    60  		Allow references to Go symbols in shared libraries (experimental).
    61  	-e
    62  		Remove the limit on the number of errors reported (default limit is 10).
    63  	-goversion string
    64  		Specify required go tool version of the runtime.
    65  		Exits when the runtime go version does not match goversion.
    66  	-h
    67  		Halt with a stack trace at the first error detected.
    68  	-importcfg file
    69  		Read import configuration from file.
    70  		In the file, set importmap, packagefile to specify import resolution.
    71  	-installsuffix suffix
    72  		Look for packages in $GOROOT/pkg/$GOOS_$GOARCH_suffix
    73  		instead of $GOROOT/pkg/$GOOS_$GOARCH.
    74  	-l
    75  		Disable inlining.
    76  	-lang version
    77  		Set language version to compile, as in -lang=go1.12.
    78  		Default is current version.
    79  	-linkobj file
    80  		Write linker-specific object to file and compiler-specific
    81  		object to usual output file (as specified by -o).
    82  		Without this flag, the -o output is a combination of both
    83  		linker and compiler input.
    84  	-m
    85  		Print optimization decisions. Higher values or repetition
    86  		produce more detail.
    87  	-memprofile file
    88  		Write memory profile for the compilation to file.
    89  	-memprofilerate rate
    90  		Set runtime.MemProfileRate for the compilation to rate.
    91  	-msan
    92  		Insert calls to C/C++ memory sanitizer.
    93  	-mutexprofile file
    94  		Write mutex profile for the compilation to file.
    95  	-nolocalimports
    96  		Disallow local (relative) imports.
    97  	-o file
    98  		Write object to file (default file.o or, with -pack, file.a).
    99  	-p path
   100  		Set expected package import path for the code being compiled,
   101  		and diagnose imports that would cause a circular dependency.
   102  	-pack
   103  		Write a package (archive) file rather than an object file
   104  	-race
   105  		Compile with race detector enabled.
   106  	-s
   107  		Warn about composite literals that can be simplified.
   108  	-shared
   109  		Generate code that can be linked into a shared library.
   110  	-spectre list
   111  		Enable spectre mitigations in list (all, index, ret).
   112  	-traceprofile file
   113  		Write an execution trace to file.
   114  	-trimpath prefix
   115  		Remove prefix from recorded source file paths.
   116  
   117  Flags related to debugging information:
   118  
   119  	-dwarf
   120  		Generate DWARF symbols.
   121  	-dwarflocationlists
   122  		Add location lists to DWARF in optimized mode.
   123  	-gendwarfinl int
   124  		Generate DWARF inline info records (default 2).
   125  
   126  Flags to debug the compiler itself:
   127  
   128  	-E
   129  		Debug symbol export.
   130  	-K
   131  		Debug missing line numbers.
   132  	-d list
   133  		Print debug information about items in list. Try -d help for further information.
   134  	-live
   135  		Debug liveness analysis.
   136  	-v
   137  		Increase debug verbosity.
   138  	-%
   139  		Debug non-static initializers.
   140  	-W
   141  		Debug parse tree after type checking.
   142  	-f
   143  		Debug stack frames.
   144  	-i
   145  		Debug line number stack.
   146  	-j
   147  		Debug runtime-initialized variables.
   148  	-r
   149  		Debug generated wrappers.
   150  	-w
   151  		Debug type checking.
   152  
   153  Compiler Directives
   154  
   155  The compiler accepts directives in the form of comments.
   156  To distinguish them from non-directive comments, directives
   157  require no space between the comment opening and the name of the directive. However, since
   158  they are comments, tools unaware of the directive convention or of a particular
   159  directive can skip over a directive like any other comment.
   160  */
   161  // Line directives come in several forms:
   162  //
   163  // 	//line :line
   164  // 	//line :line:col
   165  // 	//line filename:line
   166  // 	//line filename:line:col
   167  // 	/*line :line*/
   168  // 	/*line :line:col*/
   169  // 	/*line filename:line*/
   170  // 	/*line filename:line:col*/
   171  //
   172  // In order to be recognized as a line directive, the comment must start with
   173  // //line or /*line followed by a space, and must contain at least one colon.
   174  // The //line form must start at the beginning of a line.
   175  // A line directive specifies the source position for the character immediately following
   176  // the comment as having come from the specified file, line and column:
   177  // For a //line comment, this is the first character of the next line, and
   178  // for a /*line comment this is the character position immediately following the closing */.
   179  // If no filename is given, the recorded filename is empty if there is also no column number;
   180  // otherwise it is the most recently recorded filename (actual filename or filename specified
   181  // by previous line directive).
   182  // If a line directive doesn't specify a column number, the column is "unknown" until
   183  // the next directive and the compiler does not report column numbers for that range.
   184  // The line directive text is interpreted from the back: First the trailing :ddd is peeled
   185  // off from the directive text if ddd is a valid number > 0. Then the second :ddd
   186  // is peeled off the same way if it is valid. Anything before that is considered the filename
   187  // (possibly including blanks and colons). Invalid line or column values are reported as errors.
   188  //
   189  // Examples:
   190  //
   191  //	//line foo.go:10      the filename is foo.go, and the line number is 10 for the next line
   192  //	//line C:foo.go:10    colons are permitted in filenames, here the filename is C:foo.go, and the line is 10
   193  //	//line  a:100 :10     blanks are permitted in filenames, here the filename is " a:100 " (excluding quotes)
   194  //	/*line :10:20*/x      the position of x is in the current file with line number 10 and column number 20
   195  //	/*line foo: 10 */     this comment is recognized as invalid line directive (extra blanks around line number)
   196  //
   197  // Line directives typically appear in machine-generated code, so that compilers and debuggers
   198  // will report positions in the original input to the generator.
   199  /*
   200  The line directive is a historical special case; all other directives are of the form
   201  //go:name, indicating that they are defined by the Go toolchain.
   202  Each directive must be placed its own line, with only leading spaces and tabs
   203  allowed before the comment.
   204  Each directive applies to the Go code that immediately follows it,
   205  which typically must be a declaration.
   206  
   207  	//go:noescape
   208  
   209  The //go:noescape directive must be followed by a function declaration without
   210  a body (meaning that the function has an implementation not written in Go).
   211  It specifies that the function does not allow any of the pointers passed as
   212  arguments to escape into the heap or into the values returned from the function.
   213  This information can be used during the compiler's escape analysis of Go code
   214  calling the function.
   215  
   216  	//go:uintptrescapes
   217  
   218  The //go:uintptrescapes directive must be followed by a function declaration.
   219  It specifies that the function's uintptr arguments may be pointer values that
   220  have been converted to uintptr and must be on the heap and kept alive for the
   221  duration of the call, even though from the types alone it would appear that the
   222  object is no longer needed during the call. The conversion from pointer to
   223  uintptr must appear in the argument list of any call to this function. This
   224  directive is necessary for some low-level system call implementations and
   225  should be avoided otherwise.
   226  
   227  	//go:noinline
   228  
   229  The //go:noinline directive must be followed by a function declaration.
   230  It specifies that calls to the function should not be inlined, overriding
   231  the compiler's usual optimization rules. This is typically only needed
   232  for special runtime functions or when debugging the compiler.
   233  
   234  	//go:norace
   235  
   236  The //go:norace directive must be followed by a function declaration.
   237  It specifies that the function's memory accesses must be ignored by the
   238  race detector. This is most commonly used in low-level code invoked
   239  at times when it is unsafe to call into the race detector runtime.
   240  
   241  	//go:nosplit
   242  
   243  The //go:nosplit directive must be followed by a function declaration.
   244  It specifies that the function must omit its usual stack overflow check.
   245  This is most commonly used by low-level runtime code invoked
   246  at times when it is unsafe for the calling goroutine to be preempted.
   247  
   248  	//go:linkname localname [importpath.name]
   249  
   250  The //go:linkname directive conventionally precedes the var or func
   251  declaration named by ``localname``, though its position does not
   252  change its effect.
   253  This directive determines the object-file symbol used for a Go var or
   254  func declaration, allowing two Go symbols to alias the same
   255  object-file symbol, thereby enabling one package to access a symbol in
   256  another package even when this would violate the usual encapsulation
   257  of unexported declarations, or even type safety.
   258  For that reason, it is only enabled in files that have imported "unsafe".
   259  
   260  It may be used in two scenarios. Let's assume that package upper
   261  imports package lower, perhaps indirectly. In the first scenario,
   262  package lower defines a symbol whose object file name belongs to
   263  package upper. Both packages contain a linkname directive: package
   264  lower uses the two-argument form and package upper uses the
   265  one-argument form. In the example below, lower.f is an alias for the
   266  function upper.g:
   267  
   268      package upper
   269      import _ "unsafe"
   270      //go:linkname g
   271      func g()
   272  
   273      package lower
   274      import _ "unsafe"
   275      //go:linkname f upper.g
   276      func f() { ... }
   277  
   278  The linkname directive in package upper suppresses the usual error for
   279  a function that lacks a body. (That check may alternatively be
   280  suppressed by including a .s file, even an empty one, in the package.)
   281  
   282  In the second scenario, package upper unilaterally creates an alias
   283  for a symbol in package lower. In the example below, upper.g is an alias
   284  for the function lower.f.
   285  
   286      package upper
   287      import _ "unsafe"
   288      //go:linkname g lower.f
   289      func g()
   290  
   291      package lower
   292      func f() { ... }
   293  
   294  The declaration of lower.f may also have a linkname directive with a
   295  single argument, f. This is optional, but helps alert the reader that
   296  the function is accessed from outside the package.
   297  
   298  	//go:wasmimport importmodule importname
   299  
   300  The //go:wasmimport directive is wasm-only and must be followed by a
   301  function declaration.
   302  It specifies that the function is provided by a wasm module identified
   303  by ``importmodule`` and ``importname``.
   304  
   305  	//go:wasmimport a_module f
   306  	func g()
   307  
   308  The types of parameters and return values to the Go function are translated to
   309  Wasm according to the following table:
   310  
   311      Go types        Wasm types
   312      int32, uint32   i32
   313      int64, uint64   i64
   314      float32         f32
   315      float64         f64
   316      unsafe.Pointer  i32
   317  
   318  Any other parameter types are disallowed by the compiler.
   319  
   320  */
   321  package main
   322  

View as plain text