1FileCheck - Flexible pattern matching file verifier 2=================================================== 3 4SYNOPSIS 5-------- 6 7:program:`FileCheck` *match-filename* [*--check-prefix=XXX*] [*--strict-whitespace*] 8 9DESCRIPTION 10----------- 11 12:program:`FileCheck` reads two files (one from standard input, and one 13specified on the command line) and uses one to verify the other. This 14behavior is particularly useful for the testsuite, which wants to verify that 15the output of some tool (e.g. :program:`llc`) contains the expected information 16(for example, a movsd from esp or whatever is interesting). This is similar to 17using :program:`grep`, but it is optimized for matching multiple different 18inputs in one file in a specific order. 19 20The ``match-filename`` file specifies the file that contains the patterns to 21match. The file to verify is read from standard input unless the 22:option:`--input-file` option is used. 23 24OPTIONS 25------- 26 27.. option:: -help 28 29 Print a summary of command line options. 30 31.. option:: --check-prefix prefix 32 33 FileCheck searches the contents of ``match-filename`` for patterns to 34 match. By default, these patterns are prefixed with "``CHECK:``". 35 If you'd like to use a different prefix (e.g. because the same input 36 file is checking multiple different tool or options), the 37 :option:`--check-prefix` argument allows you to specify one or more 38 prefixes to match. Multiple prefixes are useful for tests which might 39 change for different run options, but most lines remain the same. 40 41.. option:: --check-prefixes prefix1,prefix2,... 42 43 An alias of :option:`--check-prefix` that allows multiple prefixes to be 44 specified as a comma separated list. 45 46.. option:: --input-file filename 47 48 File to check (defaults to stdin). 49 50.. option:: --match-full-lines 51 52 By default, FileCheck allows matches of anywhere on a line. This 53 option will require all positive matches to cover an entire 54 line. Leading and trailing whitespace is ignored, unless 55 :option:`--strict-whitespace` is also specified. (Note: negative 56 matches from ``CHECK-NOT`` are not affected by this option!) 57 58 Passing this option is equivalent to inserting ``{{^ *}}`` or 59 ``{{^}}`` before, and ``{{ *$}}`` or ``{{$}}`` after every positive 60 check pattern. 61 62.. option:: --strict-whitespace 63 64 By default, FileCheck canonicalizes input horizontal whitespace (spaces and 65 tabs) which causes it to ignore these differences (a space will match a tab). 66 The :option:`--strict-whitespace` argument disables this behavior. End-of-line 67 sequences are canonicalized to UNIX-style ``\n`` in all modes. 68 69.. option:: --implicit-check-not check-pattern 70 71 Adds implicit negative checks for the specified patterns between positive 72 checks. The option allows writing stricter tests without stuffing them with 73 ``CHECK-NOT``\ s. 74 75 For example, "``--implicit-check-not warning:``" can be useful when testing 76 diagnostic messages from tools that don't have an option similar to ``clang 77 -verify``. With this option FileCheck will verify that input does not contain 78 warnings not covered by any ``CHECK:`` patterns. 79 80.. option:: -version 81 82 Show the version number of this program. 83 84EXIT STATUS 85----------- 86 87If :program:`FileCheck` verifies that the file matches the expected contents, 88it exits with 0. Otherwise, if not, or if an error occurs, it will exit with a 89non-zero value. 90 91TUTORIAL 92-------- 93 94FileCheck is typically used from LLVM regression tests, being invoked on the RUN 95line of the test. A simple example of using FileCheck from a RUN line looks 96like this: 97 98.. code-block:: llvm 99 100 ; RUN: llvm-as < %s | llc -march=x86-64 | FileCheck %s 101 102This syntax says to pipe the current file ("``%s``") into ``llvm-as``, pipe 103that into ``llc``, then pipe the output of ``llc`` into ``FileCheck``. This 104means that FileCheck will be verifying its standard input (the llc output) 105against the filename argument specified (the original ``.ll`` file specified by 106"``%s``"). To see how this works, let's look at the rest of the ``.ll`` file 107(after the RUN line): 108 109.. code-block:: llvm 110 111 define void @sub1(i32* %p, i32 %v) { 112 entry: 113 ; CHECK: sub1: 114 ; CHECK: subl 115 %0 = tail call i32 @llvm.atomic.load.sub.i32.p0i32(i32* %p, i32 %v) 116 ret void 117 } 118 119 define void @inc4(i64* %p) { 120 entry: 121 ; CHECK: inc4: 122 ; CHECK: incq 123 %0 = tail call i64 @llvm.atomic.load.add.i64.p0i64(i64* %p, i64 1) 124 ret void 125 } 126 127Here you can see some "``CHECK:``" lines specified in comments. Now you can 128see how the file is piped into ``llvm-as``, then ``llc``, and the machine code 129output is what we are verifying. FileCheck checks the machine code output to 130verify that it matches what the "``CHECK:``" lines specify. 131 132The syntax of the "``CHECK:``" lines is very simple: they are fixed strings that 133must occur in order. FileCheck defaults to ignoring horizontal whitespace 134differences (e.g. a space is allowed to match a tab) but otherwise, the contents 135of the "``CHECK:``" line is required to match some thing in the test file exactly. 136 137One nice thing about FileCheck (compared to grep) is that it allows merging 138test cases together into logical groups. For example, because the test above 139is checking for the "``sub1:``" and "``inc4:``" labels, it will not match 140unless there is a "``subl``" in between those labels. If it existed somewhere 141else in the file, that would not count: "``grep subl``" matches if "``subl``" 142exists anywhere in the file. 143 144The FileCheck -check-prefix option 145~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 146 147The FileCheck :option:`-check-prefix` option allows multiple test 148configurations to be driven from one `.ll` file. This is useful in many 149circumstances, for example, testing different architectural variants with 150:program:`llc`. Here's a simple example: 151 152.. code-block:: llvm 153 154 ; RUN: llvm-as < %s | llc -mtriple=i686-apple-darwin9 -mattr=sse41 \ 155 ; RUN: | FileCheck %s -check-prefix=X32 156 ; RUN: llvm-as < %s | llc -mtriple=x86_64-apple-darwin9 -mattr=sse41 \ 157 ; RUN: | FileCheck %s -check-prefix=X64 158 159 define <4 x i32> @pinsrd_1(i32 %s, <4 x i32> %tmp) nounwind { 160 %tmp1 = insertelement <4 x i32>; %tmp, i32 %s, i32 1 161 ret <4 x i32> %tmp1 162 ; X32: pinsrd_1: 163 ; X32: pinsrd $1, 4(%esp), %xmm0 164 165 ; X64: pinsrd_1: 166 ; X64: pinsrd $1, %edi, %xmm0 167 } 168 169In this case, we're testing that we get the expected code generation with 170both 32-bit and 64-bit code generation. 171 172The "CHECK-NEXT:" directive 173~~~~~~~~~~~~~~~~~~~~~~~~~~~ 174 175Sometimes you want to match lines and would like to verify that matches 176happen on exactly consecutive lines with no other lines in between them. In 177this case, you can use "``CHECK:``" and "``CHECK-NEXT:``" directives to specify 178this. If you specified a custom check prefix, just use "``<PREFIX>-NEXT:``". 179For example, something like this works as you'd expect: 180 181.. code-block:: llvm 182 183 define void @t2(<2 x double>* %r, <2 x double>* %A, double %B) { 184 %tmp3 = load <2 x double>* %A, align 16 185 %tmp7 = insertelement <2 x double> undef, double %B, i32 0 186 %tmp9 = shufflevector <2 x double> %tmp3, 187 <2 x double> %tmp7, 188 <2 x i32> < i32 0, i32 2 > 189 store <2 x double> %tmp9, <2 x double>* %r, align 16 190 ret void 191 192 ; CHECK: t2: 193 ; CHECK: movl 8(%esp), %eax 194 ; CHECK-NEXT: movapd (%eax), %xmm0 195 ; CHECK-NEXT: movhpd 12(%esp), %xmm0 196 ; CHECK-NEXT: movl 4(%esp), %eax 197 ; CHECK-NEXT: movapd %xmm0, (%eax) 198 ; CHECK-NEXT: ret 199 } 200 201"``CHECK-NEXT:``" directives reject the input unless there is exactly one 202newline between it and the previous directive. A "``CHECK-NEXT:``" cannot be 203the first directive in a file. 204 205The "CHECK-SAME:" directive 206~~~~~~~~~~~~~~~~~~~~~~~~~~~ 207 208Sometimes you want to match lines and would like to verify that matches happen 209on the same line as the previous match. In this case, you can use "``CHECK:``" 210and "``CHECK-SAME:``" directives to specify this. If you specified a custom 211check prefix, just use "``<PREFIX>-SAME:``". 212 213"``CHECK-SAME:``" is particularly powerful in conjunction with "``CHECK-NOT:``" 214(described below). 215 216For example, the following works like you'd expect: 217 218.. code-block:: llvm 219 220 !0 = !DILocation(line: 5, scope: !1, inlinedAt: !2) 221 222 ; CHECK: !DILocation(line: 5, 223 ; CHECK-NOT: column: 224 ; CHECK-SAME: scope: ![[SCOPE:[0-9]+]] 225 226"``CHECK-SAME:``" directives reject the input if there are any newlines between 227it and the previous directive. A "``CHECK-SAME:``" cannot be the first 228directive in a file. 229 230The "CHECK-NOT:" directive 231~~~~~~~~~~~~~~~~~~~~~~~~~~ 232 233The "``CHECK-NOT:``" directive is used to verify that a string doesn't occur 234between two matches (or before the first match, or after the last match). For 235example, to verify that a load is removed by a transformation, a test like this 236can be used: 237 238.. code-block:: llvm 239 240 define i8 @coerce_offset0(i32 %V, i32* %P) { 241 store i32 %V, i32* %P 242 243 %P2 = bitcast i32* %P to i8* 244 %P3 = getelementptr i8* %P2, i32 2 245 246 %A = load i8* %P3 247 ret i8 %A 248 ; CHECK: @coerce_offset0 249 ; CHECK-NOT: load 250 ; CHECK: ret i8 251 } 252 253The "CHECK-DAG:" directive 254~~~~~~~~~~~~~~~~~~~~~~~~~~ 255 256If it's necessary to match strings that don't occur in a strictly sequential 257order, "``CHECK-DAG:``" could be used to verify them between two matches (or 258before the first match, or after the last match). For example, clang emits 259vtable globals in reverse order. Using ``CHECK-DAG:``, we can keep the checks 260in the natural order: 261 262.. code-block:: c++ 263 264 // RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s 265 266 struct Foo { virtual void method(); }; 267 Foo f; // emit vtable 268 // CHECK-DAG: @_ZTV3Foo = 269 270 struct Bar { virtual void method(); }; 271 Bar b; 272 // CHECK-DAG: @_ZTV3Bar = 273 274``CHECK-NOT:`` directives could be mixed with ``CHECK-DAG:`` directives to 275exclude strings between the surrounding ``CHECK-DAG:`` directives. As a result, 276the surrounding ``CHECK-DAG:`` directives cannot be reordered, i.e. all 277occurrences matching ``CHECK-DAG:`` before ``CHECK-NOT:`` must not fall behind 278occurrences matching ``CHECK-DAG:`` after ``CHECK-NOT:``. For example, 279 280.. code-block:: llvm 281 282 ; CHECK-DAG: BEFORE 283 ; CHECK-NOT: NOT 284 ; CHECK-DAG: AFTER 285 286This case will reject input strings where ``BEFORE`` occurs after ``AFTER``. 287 288With captured variables, ``CHECK-DAG:`` is able to match valid topological 289orderings of a DAG with edges from the definition of a variable to its use. 290It's useful, e.g., when your test cases need to match different output 291sequences from the instruction scheduler. For example, 292 293.. code-block:: llvm 294 295 ; CHECK-DAG: add [[REG1:r[0-9]+]], r1, r2 296 ; CHECK-DAG: add [[REG2:r[0-9]+]], r3, r4 297 ; CHECK: mul r5, [[REG1]], [[REG2]] 298 299In this case, any order of that two ``add`` instructions will be allowed. 300 301If you are defining `and` using variables in the same ``CHECK-DAG:`` block, 302be aware that the definition rule can match `after` its use. 303 304So, for instance, the code below will pass: 305 306.. code-block:: llvm 307 308 ; CHECK-DAG: vmov.32 [[REG2:d[0-9]+]][0] 309 ; CHECK-DAG: vmov.32 [[REG2]][1] 310 vmov.32 d0[1] 311 vmov.32 d0[0] 312 313While this other code, will not: 314 315.. code-block:: llvm 316 317 ; CHECK-DAG: vmov.32 [[REG2:d[0-9]+]][0] 318 ; CHECK-DAG: vmov.32 [[REG2]][1] 319 vmov.32 d1[1] 320 vmov.32 d0[0] 321 322While this can be very useful, it's also dangerous, because in the case of 323register sequence, you must have a strong order (read before write, copy before 324use, etc). If the definition your test is looking for doesn't match (because 325of a bug in the compiler), it may match further away from the use, and mask 326real bugs away. 327 328In those cases, to enforce the order, use a non-DAG directive between DAG-blocks. 329 330The "CHECK-LABEL:" directive 331~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 332 333Sometimes in a file containing multiple tests divided into logical blocks, one 334or more ``CHECK:`` directives may inadvertently succeed by matching lines in a 335later block. While an error will usually eventually be generated, the check 336flagged as causing the error may not actually bear any relationship to the 337actual source of the problem. 338 339In order to produce better error messages in these cases, the "``CHECK-LABEL:``" 340directive can be used. It is treated identically to a normal ``CHECK`` 341directive except that FileCheck makes an additional assumption that a line 342matched by the directive cannot also be matched by any other check present in 343``match-filename``; this is intended to be used for lines containing labels or 344other unique identifiers. Conceptually, the presence of ``CHECK-LABEL`` divides 345the input stream into separate blocks, each of which is processed independently, 346preventing a ``CHECK:`` directive in one block matching a line in another block. 347For example, 348 349.. code-block:: llvm 350 351 define %struct.C* @C_ctor_base(%struct.C* %this, i32 %x) { 352 entry: 353 ; CHECK-LABEL: C_ctor_base: 354 ; CHECK: mov [[SAVETHIS:r[0-9]+]], r0 355 ; CHECK: bl A_ctor_base 356 ; CHECK: mov r0, [[SAVETHIS]] 357 %0 = bitcast %struct.C* %this to %struct.A* 358 %call = tail call %struct.A* @A_ctor_base(%struct.A* %0) 359 %1 = bitcast %struct.C* %this to %struct.B* 360 %call2 = tail call %struct.B* @B_ctor_base(%struct.B* %1, i32 %x) 361 ret %struct.C* %this 362 } 363 364 define %struct.D* @D_ctor_base(%struct.D* %this, i32 %x) { 365 entry: 366 ; CHECK-LABEL: D_ctor_base: 367 368The use of ``CHECK-LABEL:`` directives in this case ensures that the three 369``CHECK:`` directives only accept lines corresponding to the body of the 370``@C_ctor_base`` function, even if the patterns match lines found later in 371the file. Furthermore, if one of these three ``CHECK:`` directives fail, 372FileCheck will recover by continuing to the next block, allowing multiple test 373failures to be detected in a single invocation. 374 375There is no requirement that ``CHECK-LABEL:`` directives contain strings that 376correspond to actual syntactic labels in a source or output language: they must 377simply uniquely match a single line in the file being verified. 378 379``CHECK-LABEL:`` directives cannot contain variable definitions or uses. 380 381FileCheck Pattern Matching Syntax 382~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 383 384All FileCheck directives take a pattern to match. 385For most uses of FileCheck, fixed string matching is perfectly sufficient. For 386some things, a more flexible form of matching is desired. To support this, 387FileCheck allows you to specify regular expressions in matching strings, 388surrounded by double braces: ``{{yourregex}}``. Because we want to use fixed 389string matching for a majority of what we do, FileCheck has been designed to 390support mixing and matching fixed string matching with regular expressions. 391This allows you to write things like this: 392 393.. code-block:: llvm 394 395 ; CHECK: movhpd {{[0-9]+}}(%esp), {{%xmm[0-7]}} 396 397In this case, any offset from the ESP register will be allowed, and any xmm 398register will be allowed. 399 400Because regular expressions are enclosed with double braces, they are 401visually distinct, and you don't need to use escape characters within the double 402braces like you would in C. In the rare case that you want to match double 403braces explicitly from the input, you can use something ugly like 404``{{[{][{]}}`` as your pattern. 405 406FileCheck Variables 407~~~~~~~~~~~~~~~~~~~ 408 409It is often useful to match a pattern and then verify that it occurs again 410later in the file. For codegen tests, this can be useful to allow any register, 411but verify that that register is used consistently later. To do this, 412:program:`FileCheck` allows named variables to be defined and substituted into 413patterns. Here is a simple example: 414 415.. code-block:: llvm 416 417 ; CHECK: test5: 418 ; CHECK: notw [[REGISTER:%[a-z]+]] 419 ; CHECK: andw {{.*}}[[REGISTER]] 420 421The first check line matches a regex ``%[a-z]+`` and captures it into the 422variable ``REGISTER``. The second line verifies that whatever is in 423``REGISTER`` occurs later in the file after an "``andw``". :program:`FileCheck` 424variable references are always contained in ``[[ ]]`` pairs, and their names can 425be formed with the regex ``[a-zA-Z][a-zA-Z0-9]*``. If a colon follows the name, 426then it is a definition of the variable; otherwise, it is a use. 427 428:program:`FileCheck` variables can be defined multiple times, and uses always 429get the latest value. Variables can also be used later on the same line they 430were defined on. For example: 431 432.. code-block:: llvm 433 434 ; CHECK: op [[REG:r[0-9]+]], [[REG]] 435 436Can be useful if you want the operands of ``op`` to be the same register, 437and don't care exactly which register it is. 438 439FileCheck Expressions 440~~~~~~~~~~~~~~~~~~~~~ 441 442Sometimes there's a need to verify output which refers line numbers of the 443match file, e.g. when testing compiler diagnostics. This introduces a certain 444fragility of the match file structure, as "``CHECK:``" lines contain absolute 445line numbers in the same file, which have to be updated whenever line numbers 446change due to text addition or deletion. 447 448To support this case, FileCheck allows using ``[[@LINE]]``, 449``[[@LINE+<offset>]]``, ``[[@LINE-<offset>]]`` expressions in patterns. These 450expressions expand to a number of the line where a pattern is located (with an 451optional integer offset). 452 453This way match patterns can be put near the relevant test lines and include 454relative line number references, for example: 455 456.. code-block:: c++ 457 458 // CHECK: test.cpp:[[@LINE+4]]:6: error: expected ';' after top level declarator 459 // CHECK-NEXT: {{^int a}} 460 // CHECK-NEXT: {{^ \^}} 461 // CHECK-NEXT: {{^ ;}} 462 int a 463 464Matching Newline Characters 465~~~~~~~~~~~~~~~~~~~~~~~~~~~ 466 467To match newline characters in regular expressions the character class 468``[[:space:]]`` can be used. For example, the following pattern: 469 470.. code-block:: c++ 471 472 // CHECK: DW_AT_location [DW_FORM_sec_offset] ([[DLOC:0x[0-9a-f]+]]){{[[:space:]].*}}"intd" 473 474matches output of the form (from llvm-dwarfdump): 475 476.. code-block:: llvm 477 478 DW_AT_location [DW_FORM_sec_offset] (0x00000233) 479 DW_AT_name [DW_FORM_strp] ( .debug_str[0x000000c9] = "intd") 480 481letting us set the :program:`FileCheck` variable ``DLOC`` to the desired value 482``0x00000233``, extracted from the line immediately preceding "``intd``". 483