1#!./perl 2 3BEGIN { 4 unshift @INC, 't'; 5 require Config; 6 if (($Config::Config{'extensions'} !~ /\bB\b/) ){ 7 print "1..0 # Skip -- Perl configured without B module\n"; 8 exit 0; 9 } 10 require 'test.pl'; # we use runperl from 'test.pl', so can't use Test::More 11} 12 13plan tests => 167; 14 15require_ok("B::Concise"); 16 17$out = runperl(switches => ["-MO=Concise"], prog => '$a', stderr => 1); 18 19# If either of the next two tests fail, it probably means you need to 20# fix the section labeled 'fragile kludge' in Concise.pm 21 22($op_base) = ($out =~ /^(\d+)\s*<0>\s*enter/m); 23 24is($op_base, 1, "Smallest OP sequence number"); 25 26($op_base_p1, $cop_base) 27 = ($out =~ /^(\d+)\s*<;>\s*nextstate\(main (-?\d+) /m); 28 29is($op_base_p1, 2, "Second-smallest OP sequence number"); 30 31is($cop_base, 1, "Smallest COP sequence number"); 32 33# test that with -exec B::Concise navigates past logops (bug #18175) 34 35$out = runperl( 36 switches => ["-MO=Concise,-exec"], 37 prog => q{$a=$b && print q/foo/}, 38 stderr => 1, 39); 40#diag($out); 41like($out, qr/print/, "'-exec' option output has print opcode"); 42 43######## API tests v.60 44 45use Config; # used for perlio check 46B::Concise->import(qw( set_style set_style_standard add_callback 47 add_style walk_output reset_sequence )); 48 49## walk_output argument checking 50 51# test that walk_output rejects non-HANDLE args 52foreach my $foo ("string", [], {}) { 53 eval { walk_output($foo) }; 54 isnt ($@, '', "walk_output() rejects arg '$foo'"); 55 $@=''; # clear the fail for next test 56} 57# test accessor mode when arg undefd or 0 58foreach my $foo (undef, 0) { 59 my $handle = walk_output($foo); 60 is ($handle, \*STDOUT, "walk_output set to STDOUT (default)"); 61} 62 63{ # any object that can print should be ok for walk_output 64 package Hugo; 65 sub new { my $foo = bless {} }; 66 sub print { CORE::print @_ } 67} 68my $foo = new Hugo; # suggested this API fix 69eval { walk_output($foo) }; 70is ($@, '', "walk_output() accepts obj that can print"); 71 72# test that walk_output accepts a HANDLE arg 73SKIP: { 74 skip("no perlio in this build", 4) 75 unless $Config::Config{useperlio}; 76 77 foreach my $foo (\*STDOUT, \*STDERR) { 78 eval { walk_output($foo) }; 79 is ($@, '', "walk_output() accepts STD* " . ref $foo); 80 } 81 82 # now test a ref to scalar 83 eval { walk_output(\my $junk) }; 84 is ($@, '', "walk_output() accepts ref-to-sprintf target"); 85 86 $junk = "non-empty"; 87 eval { walk_output(\$junk) }; 88 is ($@, '', "walk_output() accepts ref-to-non-empty-scalar"); 89} 90 91## add_style 92my @stylespec; 93$@=''; 94eval { add_style ('junk_B' => @stylespec) }; 95like ($@, qr/expecting 3 style-format args/, 96 "add_style rejects insufficient args"); 97 98@stylespec = (0,0,0); # right length, invalid values 99$@=''; 100eval { add_style ('junk' => @stylespec) }; 101is ($@, '', "add_style accepts: stylename => 3-arg-array"); 102 103$@=''; 104eval { add_style (junk => @stylespec) }; 105like ($@, qr/style 'junk' already exists, choose a new name/, 106 "add_style correctly disallows re-adding same style-name" ); 107 108# test new arg-checks on set_style 109$@=''; 110eval { set_style (@stylespec) }; 111is ($@, '', "set_style accepts 3 style-format args"); 112 113@stylespec = (); # bad style 114 115eval { set_style (@stylespec) }; 116like ($@, qr/expecting 3 style-format args/, 117 "set_style rejects bad style-format args"); 118 119#### for content with doc'd options 120 121our($a, $b); 122my $func = sub{ $a = $b+42 }; # canonical example asub 123 124sub render { 125 walk_output(\my $out); 126 eval { B::Concise::compile(@_)->() }; 127 # diag "rendering $@\n"; 128 return ($out, $@) if wantarray; 129 return $out; 130} 131 132SKIP: { 133 # tests output to GLOB, using perlio feature directly 134 skip "no perlio on this build", 127 135 unless $Config::Config{useperlio}; 136 137 set_style_standard('concise'); # MUST CALL before output needed 138 139 @options = qw( 140 -basic -exec -tree -compact -loose -vt -ascii 141 -base10 -bigendian -littleendian 142 ); 143 foreach $opt (@options) { 144 ($out) = render($opt, $func); 145 isnt($out, '', "got output with option $opt"); 146 } 147 148 ## test output control via walk_output 149 150 my $treegen = B::Concise::compile('-basic', $func); # reused 151 152 { # test output into a package global string (sprintf-ish) 153 our $thing; 154 walk_output(\$thing); 155 $treegen->(); 156 ok($thing, "walk_output to our SCALAR, output seen"); 157 } 158 159 # test walkoutput acceptance of a scalar-bound IO handle 160 open (my $fh, '>', \my $buf); 161 walk_output($fh); 162 $treegen->(); 163 ok($buf, "walk_output to GLOB, output seen"); 164 165 ## test B::Concise::compile error checking 166 167 # call compile on non-CODE ref items 168 if (0) { 169 # pending STASH splaying 170 171 foreach my $ref ([], {}) { 172 my $typ = ref $ref; 173 walk_output(\my $out); 174 eval { B::Concise::compile('-basic', $ref)->() }; 175 like ($@, qr/^err: not a coderef: $typ/, 176 "compile detects $typ-ref where expecting subref"); 177 is($out,'', "no output when errd"); # announcement prints 178 } 179 } 180 181 # test against a bogus autovivified subref. 182 # in debugger, it should look like: 183 # 1 CODE(0x84840cc) 184 # -> &CODE(0x84840cc) in ??? 185 186 my ($res,$err); 187 TODO: { 188 #local $TODO = "\tdoes this handling make sense ?"; 189 190 sub declared_only; 191 ($res,$err) = render('-basic', \&declared_only); 192 like ($res, qr/coderef CODE\(0x[0-9a-fA-F]+\) has no START/, 193 "'sub decl_only' seen as having no START"); 194 195 sub defd_empty {}; 196 ($res,$err) = render('-basic', \&defd_empty); 197 my @lines = split(/\n/, $res); 198 is(scalar @lines, 3, 199 "'sub defd_empty {}' seen as 3 liner"); 200 201 is(1, $res =~ /leavesub/ && $res =~ /(next|db)state/, 202 "'sub defd_empty {}' seen as 2 ops: leavesub,nextstate"); 203 204 ($res,$err) = render('-basic', \¬_even_declared); 205 like ($res, qr/coderef CODE\(0x[0-9a-fA-F]+\) has no START/, 206 "'\¬_even_declared' seen as having no START"); 207 208 { 209 package Bar; 210 our $AUTOLOAD = 'garbage'; 211 sub AUTOLOAD { print "# in AUTOLOAD body: $AUTOLOAD\n" } 212 } 213 ($res,$err) = render('-basic', Bar::auto_func); 214 like ($res, qr/unknown function \(Bar::auto_func\)/, 215 "Bar::auto_func seen as unknown function"); 216 217 ($res,$err) = render('-basic', \&Bar::auto_func); 218 like ($res, qr/coderef CODE\(0x[0-9a-fA-F]+\) has no START/, 219 "'\&Bar::auto_func' seen as having no START"); 220 221 ($res,$err) = render('-basic', \&Bar::AUTOLOAD); 222 like ($res, qr/in AUTOLOAD body: /, "found body of Bar::AUTOLOAD"); 223 224 } 225 ($res,$err) = render('-basic', Foo::bar); 226 like ($res, qr/unknown function \(Foo::bar\)/, 227 "BC::compile detects fn-name as unknown function"); 228 229 # v.62 tests 230 231 pass ("TEST POST-COMPILE OPTION-HANDLING IN WALKER SUBROUTINE"); 232 233 my $sample; 234 235 my $walker = B::Concise::compile('-basic', $func); 236 walk_output(\$sample); 237 $walker->('-exec'); 238 like($sample, qr/goto/m, "post-compile -exec"); 239 240 walk_output(\$sample); 241 $walker->('-basic'); 242 unlike($sample, qr/goto/m, "post-compile -basic"); 243 244 245 # bang at it combinatorically 246 my %combos; 247 my @modes = qw( -basic -exec ); 248 my @styles = qw( -concise -debug -linenoise -terse ); 249 250 # prep samples 251 for $style (@styles) { 252 for $mode (@modes) { 253 walk_output(\$sample); 254 reset_sequence(); 255 $walker->($style, $mode); 256 $combos{"$style$mode"} = $sample; 257 } 258 } 259 # crosscheck that samples are all text-different 260 @list = sort keys %combos; 261 for $i (0..$#list) { 262 for $j ($i+1..$#list) { 263 isnt ($combos{$list[$i]}, $combos{$list[$j]}, 264 "combos for $list[$i] and $list[$j] are different, as expected"); 265 } 266 } 267 268 # add samples with styles in different order 269 for $mode (@modes) { 270 for $style (@styles) { 271 reset_sequence(); 272 walk_output(\$sample); 273 $walker->($mode, $style); 274 $combos{"$mode$style"} = $sample; 275 } 276 } 277 # test commutativity of flags, ie that AB == BA 278 for $mode (@modes) { 279 for $style (@styles) { 280 is ( $combos{"$style$mode"}, 281 $combos{"$mode$style"}, 282 "results for $style$mode vs $mode$style are the same" ); 283 } 284 } 285 286 my %save = %combos; 287 %combos = (); # outputs for $mode=any($order) and any($style) 288 289 # add more samples with switching modes & sticky styles 290 for $style (@styles) { 291 walk_output(\$sample); 292 reset_sequence(); 293 $walker->($style); 294 for $mode (@modes) { 295 walk_output(\$sample); 296 reset_sequence(); 297 $walker->($mode); 298 $combos{"$style/$mode"} = $sample; 299 } 300 } 301 # crosscheck that samples are all text-different 302 @nm = sort keys %combos; 303 for $i (0..$#nm) { 304 for $j ($i+1..$#nm) { 305 isnt ($combos{$nm[$i]}, $combos{$nm[$j]}, 306 "results for $nm[$i] and $nm[$j] are different, as expected"); 307 } 308 } 309 310 # add samples with switching styles & sticky modes 311 for $mode (@modes) { 312 walk_output(\$sample); 313 reset_sequence(); 314 $walker->($mode); 315 for $style (@styles) { 316 walk_output(\$sample); 317 reset_sequence(); 318 $walker->($style); 319 $combos{"$mode/$style"} = $sample; 320 } 321 } 322 # test commutativity of flags, ie that AB == BA 323 for $mode (@modes) { 324 for $style (@styles) { 325 is ( $combos{"$style/$mode"}, 326 $combos{"$mode/$style"}, 327 "results for $style/$mode vs $mode/$style are the same" ); 328 } 329 } 330 331 332 #now do double crosschecks: commutativity across stick / nostick 333 %combos = (%combos, %save); 334 335 # test commutativity of flags, ie that AB == BA 336 for $mode (@modes) { 337 for $style (@styles) { 338 339 is ( $combos{"$style$mode"}, 340 $combos{"$style/$mode"}, 341 "$style$mode VS $style/$mode are the same" ); 342 343 is ( $combos{"$mode$style"}, 344 $combos{"$mode/$style"}, 345 "$mode$style VS $mode/$style are the same" ); 346 347 is ( $combos{"$style$mode"}, 348 $combos{"$mode/$style"}, 349 "$style$mode VS $mode/$style are the same" ); 350 351 is ( $combos{"$mode$style"}, 352 $combos{"$style/$mode"}, 353 "$mode$style VS $style/$mode are the same" ); 354 } 355 } 356} 357 358 359# test proper NULLING of pointer, derefd by CvSTART, when a coderef is 360# undefd. W/o this, the pointer can dangle into freed and reused 361# optree mem, which no longer points to opcodes. 362 363# Using B::Concise to render Config::AUTOLOAD's optree at BEGIN-time 364# triggers this obscure bug, cuz AUTOLOAD has a bootstrap version, 365# which is used at load-time then undeffed. It is normally 366# re-vivified later, but not in time for this (BEGIN/CHECK)-time 367# rendering. 368 369$out = runperl ( switches => ["-MO=Concise,Config::AUTOLOAD"], 370 prog => 'use Config; BEGIN { $Config{awk} }', 371 stderr => 1 ); 372 373like($out, qr/Config::AUTOLOAD exists in stash, but has no START/, 374 "coderef properly undefined"); 375 376$out = runperl ( switches => ["-MO=Concise,Config::AUTOLOAD"], 377 prog => 'use Config; CHECK { $Config{awk} }', 378 stderr => 1 ); 379 380like($out, qr/Config::AUTOLOAD exists in stash, but has no START/, 381 "coderef properly undefined"); 382 383# test -stash and -src rendering 384$out = runperl ( switches => ["-MO=-qq,Concise,-stash=B::Concise,-src"], 385 prog => '-e 1', stderr => 1 ); 386 387like($out, qr/FUNC: \*B::Concise::concise_cv_obj/, 388 "stash rendering of B::Concise includes Concise::concise_cv_obj"); 389 390like($out, qr/FUNC: \*B::Concise::walk_output/, 391 "stash rendering includes Concise::walk_output"); 392 393like($out, qr/\# 4\d\d: \s+ \$l->concise\(\$level\);/, 394 "src-line rendering works"); 395 396$out = runperl ( switches => ["-MStorable", "-MO=Concise,-stash=Storable,-src"], 397 prog => '-e 1', stderr => 1 ); 398 399like($out, qr/FUNC: \*Storable::BIN_MAJOR/, 400 "stash rendering has constant sub: Storable::BIN_MAJOR"); 401 402like($out, qr/BIN_MAJOR is a constant sub, optimized to a IV/, 403 "stash rendering identifies it as constant"); 404 405$out = runperl ( switches => ["-MO=Concise,-stash=ExtUtils::Mksymlists,-src,-exec"], 406 prog => '-e 1', stderr => 1 ); 407 408like($out, qr/FUNC: \*ExtUtils::Mksymlists::_write_vms/, 409 "stash rendering loads package as needed"); 410 411$out = runperl ( switches => ["-MO=Concise,-stash=Data::Dumper,-src,-exec"], 412 prog => '-e 1', stderr => 1 ); 413 414SKIP: { 415 skip "Data::Dumper is statically linked", 1 416 if $Config{static_ext} =~ m|\bData/Dumper\b|; 417 like($out, qr/FUNC: \*Data::Dumper::format_refaddr/, 418 "stash rendering loads package as needed"); 419} 420 421my $prog = q{package FOO; sub bar { print q{bar} } package main; FOO::bar(); }; 422 423# this would fail if %INC used for -stash test 424$out = runperl ( switches => ["-MO=Concise,-src,-stash=FOO,-main"], 425 prog => $prog, stderr => 1 ); 426 427like($out, qr/FUNC: \*FOO::bar/, 428 "stash rendering works on inlined package"); 429 430# Test that consecutive nextstate ops are not nulled out when PERLDBf_NOOPT 431# is set. 432# XXX Does this test belong here? 433 434$out = runperl ( switches => ["-MO=Concise"], 435 prog => 'BEGIN{$^P = 0x04} 1 if 0; print', 436 stderr => 1 ); 437like $out, qr/nextstate.*nextstate/s, 438 'nulling of nextstate-nextstate happeneth not when $^P | PERLDBf_NOOPT'; 439 440 441# A very basic test for -tree output 442$out = 443 runperl( 444 switches => ["-MO=Concise,-tree"], prog => 'print', stderr => 1 445 ); 446ok index $out=~s/\r\n/\n/gr=~s/gvsv\(\*_\)/gvsv[*_]/r, <<'end'=~s/\r\n/\n/gr =>>= 0, '-tree output'; 447<6>leave[1 ref]-+-<1>enter 448 |-<2>nextstate(main 1 -e:1) 449 `-<5>print-+-<3>pushmark 450 `-ex-rv2sv---<4>gvsv[*_] 451end 452 453# -nobanner 454$out = 455 runperl( 456 switches => ["-MO=Concise,-nobanner,foo"], prog=>'sub foo{}', stderr => 1 457 ); 458unlike $out, qr/main::foo/, '-nobanner'; 459 460# glob 461$out = 462 runperl( 463 switches => ["-MO=Concise"], prog=>'glob(q{.})', stderr => 1 464 ); 465like $out, qr/\*<none>::/, 'glob(q{.})'; 466 467# Test op_other in -debug 468$out = runperl( 469 switches => ["-MO=Concise,-debug,xx"], 470 prog => q{sub xx { if ($a) { return $b } }}, 471 stderr => 1, 472); 473 474$out =~s/\r\n/\n/g; 475 476# Look for OP_AND 477$end = <<'EOF'; 478LOGOP \(0x\w+\) 479 op_next 0x\w+ 480 op_other (0x\w+) 481 op_sibling 0 482 op_ppaddr PL_ppaddr\[OP_AND\] 483EOF 484 485$end =~ s/\r\n/\n/g; 486 487like $out, qr/$end/, 'OP_AND has op_other'; 488 489# like(..) above doesn't fill in $1 490$out =~ $end; 491my $next = $1; 492 493# Check it points to a PUSHMARK 494$end = <<'EOF'; 495OP \(<NEXT>\) 496 op_next 0x\w+ 497 op_sibling 0x\w+ 498 op_ppaddr PL_ppaddr\[OP_PUSHMARK\] 499EOF 500 501$end =~ s/<NEXT>/$next/; 502 503like $out, qr/$end/, 'OP_AND->op_other points correctly'; 504 505# test nextstate hints display 506 507{ 508 509 $out = runperl( 510 switches => ["-MO=Concise"], 511 prog => q{my $x; use strict; use warnings; $x++; use feature q(:5.11); $x++}, 512 stderr => 1, 513 ); 514 515 my @hints = $out =~ /nextstate\([^)]+\) (.*) ->/g; 516 517 # handle test script run with PERL_UNICODE="" 518 s/>,<,// for @hints; 519 s/%,// for @hints; 520 521 is(scalar(@hints), 3, "3 hints"); 522 is($hints[0], 'v:{', "hints[0]"); 523 is($hints[1], 'v:*,&,{,x*,x&,x$,$', "hints[1]"); 524 is($hints[2], 'v:us,*,&,{,x*,x&,x$,$,fea=15', "hints[2]"); 525} 526 527__END__ 528