1<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> 2<html xmlns="http://www.w3.org/1999/xhtml"> 3<head> 4<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" /> 5<link href="style.css" rel="stylesheet" type="text/css" /> 6<title>LLDB Tutorial</title> 7</head> 8 9<body> 10 <div class="www_title"> 11 The <strong>LLDB</strong> Debugger 12 </div> 13 14<div id="container"> 15 <div id="content"> 16 <!--#include virtual="sidebar.incl"--> 17 <div id="middle"> 18 <div class="post"> 19 <h1 class ="postheader">Getting Started</h1> 20 <div class="postcontent"> 21 22 <p>Here's a short precis of how to run lldb if you are familiar with the gdb command set. 23 We will start with some details on lldb command structure and syntax to help orient you.</p> 24 25 </div> 26 <div class="postfooter"></div> 27 28 <div class="post"> 29 <h1 class ="postheader">Command Structure</h1> 30 <div class="postcontent"> 31 32 <p>Unlike gdb's command set, which is rather free-form, we tried to make 33 the lldb command syntax fairly structured. The commands are all of the 34 form:</p> 35 36 <code color=#ff0000> 37 <noun> <verb> [-options [option-value]] [argument [argument...]] 38 </code> 39 40 <p>The command line parsing is done before command execution, so it is 41 uniform across all the commands. The command syntax for basic commands is very simple, 42 arguments, options and option values are all white-space 43 separated, and double-quotes are used to protect white-spaces in an argument. 44 If you need to put a backslash or double-quote character 45 in an argument you back-slash it in the argument. That makes the 46 command syntax more regular, but it also means you may have to 47 quote some arguments in lldb that you wouldn't in gdb.</p> 48 49 <p>Options can be placed anywhere on the command line, but if the arguments 50 begin with a "<code>-</code>" then you have to tell lldb that you're done with options 51 for the current command by adding an option termination: "<code>--</code>" 52 So for instance if you want to launch a process and give the "process launch" command 53 the "<code>--stop-at-entry</code>" option, yet you want the 54 process you are about to launch to be launched with the arguments 55 "<code>-program_arg value</code>", you would type:</p> 56 57 <code> 58 (lldb) process launch --stop-at-entry -- -program_arg value 59 </code> 60 61 <p>We also tried to reduce the number of special purpose argument 62 parsers, which sometimes forces the user to be a little more explicit 63 about stating their intentions. The first instance you'll note of 64 this is the breakpoint command. In gdb, to set a breakpoint, you 65 might enter</p> 66 67 <code> 68 (gdb) break foo.c:12 69 </code> 70 <p>to break at line 12 of foo.c, and:</p> 71 <code> 72 (gdb) break foo 73 </code> 74 75 <p>to break at the function <code>foo</code>. As time went on, the parser that tells <code>foo.c:12</code> 76 from <code>foo</code> from <code>foo.c::foo</code> (which means the function foo in the file 77 foo.c) got more and more complex and bizarre, and especially in C++ 78 there are times where there's really no way to specify the function 79 you want to break on. The lldb commands are more verbose but also more precise 80 and allow for intellegent auto completion. 81 82 <p>To set the same file and line breakpoint in LLDB you can enter either of:</p> 83 84 <code> 85 (lldb) breakpoint set --file foo.c --line 12 86 <br>(lldb) breakpoint set -f foo.c -l 12 87 </code> 88 89 <p>To set a breakpoint on a function named <code>foo</code> in LLDB you can enter either of:</p> 90 91 <code> 92 (lldb) breakpoint set --name foo 93 <br>(lldb) breakpoint set -n foo 94 </code> 95 96 <p>You can use the --name option multiple times to make a breakpoint on a set of functions as well. This is convenient 97 since it allows you to set commmon conditions or commands without having to specify them multiple times:</p> 98 99 <code> 100 (lldb) breakpoint set --name foo --name bar 101 </code> 102 103 <p>Setting breakpoints by name is even more specialized in LLDB as you can specify 104 that you want to set a breakpoint at a function by method name. To set a breakpoint 105 on all C++ methods named <code>foo</code> you can enter either of:</p> 106 107 <code> 108 (lldb) breakpoint set --method foo 109 <br>(lldb) breakpoint set -M foo 110 </code> 111 112 <p>To set a breakpoint Objective C selectors named <code>alignLeftEdges:</code> you can enter either of:</p> 113 114 <code> 115 (lldb) breakpoint set --selector alignLeftEdges: 116 <br>(lldb) breakpoint set -S alignLeftEdges: 117 </code> 118 119 <p>You can limit any breakpoints to a specific executable image by using 120 the "<code>--shlib <path></code>" ("<code>-s <path></code>" for short):</p> 121 122 <code> 123 (lldb) breakpoint set --shlib foo.dylib --name foo 124 <br>(lldb) breakpoint set -s foo.dylib -n foo 125 </code> 126 127 <p>The <code>--shlib</code> option can also be repeated to specify several shared libraries.</p> 128 129 <p>Suggestions on more interesting primitives of this sort are also very welcome.</p> 130 131 <p>Just like gdb, the lldb command interpreter does a shortest unique 132 string match on command names, so the following two commands will 133 both execute the same command:</p> 134 135 <code> 136 (lldb) breakpoint set -n "-[SKTGraphicView alignLeftEdges:]" 137 <br>(lldb) br s -n "-[SKTGraphicView alignLeftEdges:]" 138 </code> 139 140 <p>lldb also supports command completion for source file names, symbol 141 names, file names, etc. Completion is initiated by a hitting a <b>TAB</b>. 142 Individual options in a command can have different completers, so for 143 instance the "<code>--file <path></code>" option in "breakpoint" completes to source files, the 144 "<code>--shlib <path></code>" option to currently loaded shared libraries, etc. We can even do 145 things like if you specify "<code>--shlib <path></code>", and are completing on "<code>--file <path></code>", we will only 146 list source files in the shared library specified by "<code>--shlib <path></code>".</p> 147 148 <p>The individual commands are pretty extensively documented. You can 149 use the <code>help</code> command to get an overview of which commands are 150 available or to obtain details about specific commands. There is also an 151 <code>apropos</code> command that will search the help text for all commands 152 for a particular word and dump a summary help string for each matching 153 command.</p> 154 155 <p>Finally, there is a mechanism to construct aliases for commonly used 156 commands. So for instance if you get annoyed typing:</p> 157 158 <code> 159 (lldb) breakpoint set --file foo.c --line 12 160 </code> 161 162 <p>you can do:</p> 163 164 <code> 165 (lldb) command alias bfl breakpoint set -f %1 -l %2 166 <br>(lldb) bfl foo.c 12 167 </code> 168 169 <p>We have added a few aliases for commonly used commands (e.g. "step", 170 "next" and "continue") but we haven't tried to be exhaustive because 171 in our experience it is more convenient to make the basic commands 172 unique down to a letter or two, and then learn these sequences than 173 to fill the namespace with lots of aliases, and then have to type them 174 all the way out.</p> 175 176 <p>However, users are free to customize lldb's command set however they 177 like, and since lldb reads the file ~/.lldbinit at startup, you can 178 store all your aliases there and they will be generally available to 179 you. Your aliases are also documented in the help command so you can 180 remind yourself of what you've set up.</p> 181 182 <p> One alias of note that we do include by popular demand is a weak emulator 183 of gdb's "break" command. It doesn't try to do everything that gdb's 184 break command does (for instance, it doesn't handle <code>foo.c::bar</code>. But 185 it mostly works, and makes the transition easier. Also by popular demand, it 186 is aliased to <code>b</code>. If you actually want to learn the lldb command 187 set natively, that means it will get in the way of the rest of the breakpoint 188 commands. Fortunately, if you don't like one of our aliases, you an easily 189 get rid of it by running (for example):</p> 190 191 <code> 192 (lldb) command unalias b 193 </code> 194 195 <p>I actually also do:</p> 196 197 <code> 198 (lldb) command alias b breakpoint 199 </code> 200 201 <p>so I can run the native lldb breakpoint command with just <code>b</code></p> 202 203 <p>The lldb command parser also supports "raw" commands, where, after command options 204 are stripped off, the rest of the command string is passed uninterpreted to the command. 205 This is convenient for commands whose arguments might be some complex expression that would 206 be painful to backslash protect. 207 For instance the "expression" command is a "raw" command for obvious reasons. The 208 "help" output for a command will tell you if it is "raw" or not, so you know what to expect. 209 The one thing you have to watch out for is that since raw commands still can have options, 210 if your command string has dashes in it, you'll have to indicate these are not option 211 markers by putting "--" after the command name, but before your command string. 212 213 <p>lldb also has a built-in Python interpreter, which is accessible by 214 the "script" command. All the functionality of the debugger is 215 available as classes in the Python interpreter, so the more complex 216 commands that in gdb you would introduce with the "define" command can 217 be done by writing Python functions using the lldb-Python library, 218 then loading the scripts into your running session and accessing them 219 with the "script" command.</p> 220 221 <p>Having given an overview of lldb's command syntax, we proceed to lay out the stages 222 of a standard debug session.</p> 223 224 </div> 225 <div class="postfooter"></div> 226 227 228 <div class="post"> 229 <h1 class ="postheader">Loading a program into lldb</h1> 230 <div class="postcontent"> 231 232 <p>First we need to set the program to debug. As with gdb, you 233 can start lldb and specify the file you wish to debug on the command line:</p> 234 235 <code> 236 $ lldb /Projects/Sketch/build/Debug/Sketch.app 237 <br>Current executable set to '/Projects/Sketch/build/Debug/Sketch.app' (x86_64). 238 </code> 239 240 <p>or you can specify it after the fact with the "file" command:</p> 241 242 <code> 243 $ lldb 244 <br>(lldb) file /Projects/Sketch/build/Debug/Sketch.app 245 <br>Current executable set to '/Projects/Sketch/build/Debug/Sketch.app' (x86_64). 246 </code> 247 <p> 248 </div> 249 <div class="postfooter"></div> 250 251 <div class="post"> 252 <h1 class ="postheader">Setting breakpoints</h1> 253 <div class="postcontent"> 254 255 <p>We've discussed how to set breakpoints above. You can use <code>help breakpoint set</code> 256 to see all the options for breakpoint setting. For instance, we might do:</p> 257 258 <code> 259 (lldb) breakpoint set --selector alignLeftEdges: 260 <br>Breakpoint created: 1: name = 'alignLeftEdges:', locations = 1, resolved = 1 261 </code> 262 263 <p>You can find out about the breakpoints you've set with:</p> 264 265 <pre><tt>(lldb) breakpoint list 266Current breakpoints: 2671: name = 'alignLeftEdges:', locations = 1, resolved = 1 268 1.1: where = Sketch`-[SKTGraphicView alignLeftEdges:] + 33 at /Projects/Sketch/SKTGraphicView.m:1405, address = 0x0000000100010d5b, resolved, hit count = 0 269</tt></pre> 270 271 <p>Note that setting a breakpoint creates a <i>logical</i> breakpoint, which could 272 resolve to one or more <i>locations</i>. For instance, break by selector would 273 set a breakpoint on all the methods that implement that selector in the classes in 274 your program. Similarly, a file and line breakpoint might result in multiple 275 locations if that file and line were inlined in different places in your code.</p> 276 277 <p>The logical breakpoint has an integer id, and it's locations have an 278 id within their parent breakpoint (the two are joined by a ".", 279 e.g. 1.1 in the example above.) </p> 280 281 <p>Also the logical breakpoints remain <i>live</i> so that if another shared library 282 were to be loaded that had another implementation of the 283 "<code>alignLeftEdges:</code>" selector, the new location would be added to 284 breakpoint 1 (e.g. a "1.2" breakpoint would be set on the newly loaded 285 selector).</p> 286 287 <p>The other piece of information in the breakpoint listing is whether the 288 breakpoint location was <i>resolved</i> or not. A location gets resolved when 289 the file address it corresponds to gets loaded into the program you are 290 debugging. For instance if you set a breakpoint in a shared library that 291 then gets unloaded, that breakpoint location will remain, but it will no 292 longer be <i>resolved</i>.</p> 293 294 <p>One other thing to note for gdb users is that lldb acts like gdb with:</p> 295 296 <code> 297 (gdb) set breakpoint pending on 298 </code> 299 300 <p>That is, lldb will always make a breakpoint from your specification, even 301 if it couldn't find any locations that match the specification. You can tell 302 whether the expression was resolved or not by checking the locations field 303 in "breakpoint list", and we report the breakpoint as "pending" when you 304 set it so you can tell you've made a typo more easily, if that was indeed 305 the reason no locations were found:</p> 306 307 <code> 308 (lldb) breakpoint set --file foo.c --line 12 309 <br>Breakpoint created: 2: file ='foo.c', line = 12, locations = 0 (pending) 310 <br>WARNING: Unable to resolve breakpoint to any actual locations. 311 </code> 312 313 <p>You can delete, disable, set conditions and ignore counts either on all the 314 locations generated by your logical breakpoint, or on any one of the particular locations 315 your specification resolved to. For instance if we wanted to add a command 316 to print a backtrace when we hit this breakpoint we could do:</p> 317 318 <code> 319 (lldb) breakpoint command add 1.1 320 <br>Enter your debugger command(s). Type 'DONE' to end. 321 <br>> bt 322 <br>> DONE 323 </code> 324 325 <p>By default, the <code> breakpoint command add</code> command takes lldb command line commands. 326 You can also specify this explicitly by passing the "<code>--command</code>" option. 327 Use "<code>--script</code>" if you want to implement your breakpoint command using the Python script instead.</p> 328 329 <p>This is an convenient point to bring up another feature of the lldb command help. Do:</p> 330 331 <code> 332 (lldb) help break command add 333 <br>Add a set of commands to a breakpoint, to be executed whenever the breakpoint is hit. 334 <br> 335 <br>Syntax: breakpoint command add <cmd-options> <breakpt-id> 336 <br> etc... 337 </code> 338 339 <p>When you see arguments to commands specified in the Syntax in angle 340 brackets like <code><breakpt-id></code>, that indicates that 341 that is some common argument type that you can get further help on from the command system. 342 So in this case you could do:</p> 343 344 <code> 345 (lldb) help <breakpt-id> 346 <br><breakpt-id> -- Breakpoint ID's consist major and minor numbers; the major 347 <br> etc... 348 </code> 349 350 </div> 351 <div class="postfooter"></div> 352 353 <div class="post"> 354 <h1 class ="postheader">Setting watchpoints</h1> 355 <div class="postcontent"> 356 357 <p>In addition to breakpoints, you can use <code>help watchpoint</code> 358 to see all the commands for watchpoint manipulations. For instance, we might do the following to watch 359 a variable called 'global' for write operation, but only stop if the condition '(global==5)' is true:</p> 360 361 <pre><tt>(lldb) watch set var global 362Watchpoint created: Watchpoint 1: addr = 0x100001018 size = 4 state = enabled type = w 363 declare @ '/Volumes/data/lldb/svn/ToT/test/functionalities/watchpoint/watchpoint_commands/condition/main.cpp:12' 364(lldb) watch modify -c '(global==5)' 365(lldb) watch list 366Current watchpoints: 367Watchpoint 1: addr = 0x100001018 size = 4 state = enabled type = w 368 declare @ '/Volumes/data/lldb/svn/ToT/test/functionalities/watchpoint/watchpoint_commands/condition/main.cpp:12' 369 condition = '(global==5)' 370(lldb) c 371Process 15562 resuming 372(lldb) about to write to 'global'... 373Process 15562 stopped and was programmatically restarted. 374Process 15562 stopped and was programmatically restarted. 375Process 15562 stopped and was programmatically restarted. 376Process 15562 stopped and was programmatically restarted. 377Process 15562 stopped 378* thread #1: tid = 0x1c03, 0x0000000100000ef5 a.out`modify + 21 at main.cpp:16, stop reason = watchpoint 1 379 frame #0: 0x0000000100000ef5 a.out`modify + 21 at main.cpp:16 380 13 381 14 static void modify(int32_t &var) { 382 15 ++var; 383-> 16 } 384 17 385 18 int main(int argc, char** argv) { 386 19 int local = 0; 387(lldb) bt 388* thread #1: tid = 0x1c03, 0x0000000100000ef5 a.out`modify + 21 at main.cpp:16, stop reason = watchpoint 1 389 frame #0: 0x0000000100000ef5 a.out`modify + 21 at main.cpp:16 390 frame #1: 0x0000000100000eac a.out`main + 108 at main.cpp:25 391 frame #2: 0x00007fff8ac9c7e1 libdyld.dylib`start + 1 392(lldb) frame var global 393(int32_t) global = 5 394(lldb) watch list -v 395Current watchpoints: 396Watchpoint 1: addr = 0x100001018 size = 4 state = enabled type = w 397 declare @ '/Volumes/data/lldb/svn/ToT/test/functionalities/watchpoint/watchpoint_commands/condition/main.cpp:12' 398 condition = '(global==5)' 399 hw_index = 0 hit_count = 5 ignore_count = 0 400(lldb) </tt></pre> 401 </div> 402 <div class="postfooter"></div> 403 404 <div class="post"> 405 <h1 class ="postheader">Starting or attaching to your Program</h1> 406 <div class="postcontent"> 407 408 <p>To launch a program in lldb we use the "<code>process launch</code>" command or 409 one of its built in aliases:</p> 410 411 <code> 412 (lldb) process launch 413 <br>(lldb) run 414 <br>(lldb) r 415 </code> 416 417 <p>You can also attach to a process by process ID or process name. 418 When attaching to a process by name, lldb also supports the "<code>--waitfor</code>" option which waits for the 419 next process that has that name to show up, and attaches to it</p> 420 421 <code> 422 (lldb) process attach --pid 123 423 <br>(lldb) process attach --name Sketch 424 <br>(lldb) process attach --name Sketch --waitfor 425 </code> 426 427 <p>After you launch or attach to a process, your process might stop 428 somewhere:</p> 429 <code> 430 (lldb) process attach -p 12345 431 <br>Process 46915 Attaching 432 <br>Process 46915 Stopped 433 <br>1 of 3 threads stopped with reasons: 434 <br>* thread #1: tid = 0x2c03, 0x00007fff85cac76a, where = libSystem.B.dylib`__getdirentries64 + 10, stop reason = signal = SIGSTOP, queue = com.apple.main-thread 435 </code> 436 437 438 <p>Note the line that says "<code>1 of 3 threads stopped with reasons:</code>" and the 439 lines that follow it. In a multi-threaded environment it is very 440 common for more than one thread to hit your breakpoint(s) before the 441 kernel actually returns control to the debugger. In that case, you 442 will see all the threads that stopped for some interesting reason 443 listed in the stop message.</p> 444 445 </div> 446 <div class="postfooter"></div> 447 448 <div class="post"> 449 <h1 class ="postheader">Controlling your Program</h1> 450 <div class="postcontent"> 451 452 453 <p>After launching, we can continue until we hit our breakpoint. The primitive 454 commands for process control all exist under the "thread" command:</p> 455 456 <code> 457 (lldb) thread continue 458 <br>Resuming thread 0x2c03 in process 46915 459 <br>Resuming process 46915 460 <br>(lldb) 461 </code> 462 463 <p>At present you can only operate on one thread at a time, but the 464 design will ultimately support saying "step over the function in 465 Thread 1, and step into the function in Thread 2, and continue Thread 466 3" etc. When we eventually support keeping some threads running while 467 others are stopped this will be particularly important. For 468 convenience, however, all the stepping commands have easy aliases. 469 So "thread continue" is just "c", etc.</p> 470 471 <p>The other program stepping commands are pretty much the same as in gdb. 472 You've got:</p> 473 474 <pre><tt>(lldb) thread step-in // The same as gdb's "step" or "s" 475(lldb) thread step-over // The same as gdb's "next" or "n" 476(lldb) thread step-out // The same as gdb's "finish" or "f" 477</tt></pre> 478 479 <p>By default, lldb does defined aliases to all common gdb process control 480 commands ("<code>s</code>", "<code>step</code>", "<code>n</code>", "<code>next</code>", "<code>finish</code>"). 481 If we have missed any, please add them to your <code>~/.lldbinit</code> file 482 using the "<code>command alias</code>" command. 483 484 <p>lldb also supported the <i>step by instruction</i> versions:</p> 485 <pre><tt>(lldb) thread step-inst // The same as gdb's "stepi" / "si" 486(lldb) thread step-over-inst // The same as gdb's "nexti" / "ni" 487</tt></pre> 488 489 <p>Finally, lldb has a <i>run until line or frame exit</i> stepping mode:</p> 490 491 <code> 492 (lldb) thread until 100 493 </code> 494 495 <p>This command will run the thread in the current frame till it reaches line 100 in 496 this frame or stops if it leaves the current frame. This is a pretty 497 close equivalent to gdb's "<code>until</code>" command.</p> 498 499 <p>A process, by default, will shared the lldb terminal with the inferior 500 process. When in this mode, much like when debugging with gdb, when 501 the process is running anything you type will go to the STDIN of the 502 inferior process. To interrupt your inferior program, type CTRL+C.</p> 503 504 <p>If you attach to a process, or launch a process with the "<code>--no-stdin</code>" 505 option, the command interpreter is always available to enter commands. This 506 might be a little disconcerting to gdb users when always have an <code>(lldb)</code> 507 prompt. This allows you to set a breakpoint, etc without having to explicitly interrupt 508 the program you are debugging:</p> 509 510 <code> 511 (lldb) process continue 512 <br>(lldb) breakpoint set --name stop_here 513 </code> 514 515 <p>There are many commands that won't work while running, and the command 516 interpreter should do a good job of letting you know when this is the 517 case. If you find any instances where the command interpreter isn't 518 doing its job, please file a bug. This way of operation will set us 519 up for a future debugging mode called <i>thread centric debugging</i>. 520 This mode will allow us to run all threads and only stop the threads 521 that are at breakpoints or have exceptions or signals.</p> 522 523 <p>The commands that currently work while running include 524 interrupting the process to halt execution ("<code>process interrupt</code>"), 525 getting the process status ("<code>process status</code>"), 526 breakpoint setting and clearing ("<code> breakpoint [set|clear|enable|disable|list] ...</code>"), 527 and memory reading and writing ("<code> memory [read|write] ...</code>"). 528 </p> 529 530 <p>The question of disabling stdio when running brings up a good opportunity to 531 show how to set debugger properties in general. 532 If you always want to run in the <code>--no-stdin</code> mode, you can set this 533 as a generic process property using the lldb "<code>settings</code>&qout; command, 534 which is equivalent to gdb's "<code>set</code>" command. For instance, 535 in this case you would say:</p> 536 537 <code> 538 (lldb) settings set target.process.disable-stdio true 539 </code> 540 541 <p>Over time, gdb's "<code>set</code> command became a wilderness of disordered options, 542 so that there were useful options that even experienced gdb users didn't know about 543 because they were too hard to find. We tried to organize the settings hierarchically 544 using the structure of the basic entities in the debugger. For the most part anywhere 545 you can specify a setting on a generic entity (threads, for example) you can also apply 546 the option to a particular instance, which can also be convenient at times. 547 You can view the available settings with "<code>settings list</code>" and 548 there is help on the settings command explaining how it works more generally.</p> 549 550 </div> 551 <div class="postfooter"></div> 552 553 <div class="post"> 554 <h1 class ="postheader">Examining Thread State</h1> 555 <div class="postcontent"> 556 557 <p>Once you've stopped, lldb will choose a current thread, usually the 558 one that stopped "for a reason", and a current frame in that thread (on stop this is always the bottom-most frame). 559 Many the commands for inspecting state work on this current 560 thread/frame.</p> 561 562 <p>To inspect the current state of your process, you can start with the 563 threads:</p> 564 565 <pre><tt>(lldb) thread list 566Process 46915 state is Stopped 567* thread #1: tid = 0x2c03, 0x00007fff85cac76a, where = libSystem.B.dylib`__getdirentries64 + 10, stop reason = signal = SIGSTOP, queue = com.apple.main-thread 568 thread #2: tid = 0x2e03, 0x00007fff85cbb08a, where = libSystem.B.dylib`kevent + 10, queue = com.apple.libdispatch-manager 569 thread #3: tid = 0x2f03, 0x00007fff85cbbeaa, where = libSystem.B.dylib`__workq_kernreturn + 10 570</tt></pre> 571 572 <p>The * indicates that Thread 1 is the current thread. To get a 573 backtrace for that thread, do:</p> 574 575 <pre><tt>(lldb) thread backtrace 576thread #1: tid = 0x2c03, stop reason = breakpoint 1.1, queue = com.apple.main-thread 577 frame #0: 0x0000000100010d5b, where = Sketch`-[SKTGraphicView alignLeftEdges:] + 33 at /Projects/Sketch/SKTGraphicView.m:1405 578 frame #1: 0x00007fff8602d152, where = AppKit`-[NSApplication sendAction:to:from:] + 95 579 frame #2: 0x00007fff860516be, where = AppKit`-[NSMenuItem _corePerformAction] + 365 580 frame #3: 0x00007fff86051428, where = AppKit`-[NSCarbonMenuImpl performActionWithHighlightingForItemAtIndex:] + 121 581 frame #4: 0x00007fff860370c1, where = AppKit`-[NSMenu performKeyEquivalent:] + 272 582 frame #5: 0x00007fff86035e69, where = AppKit`-[NSApplication _handleKeyEquivalent:] + 559 583 frame #6: 0x00007fff85f06aa1, where = AppKit`-[NSApplication sendEvent:] + 3630 584 frame #7: 0x00007fff85e9d922, where = AppKit`-[NSApplication run] + 474 585 frame #8: 0x00007fff85e965f8, where = AppKit`NSApplicationMain + 364 586 frame #9: 0x0000000100015ae3, where = Sketch`main + 33 at /Projects/Sketch/SKTMain.m:11 587 frame #10: 0x0000000100000f20, where = Sketch`start + 52 588</tt></pre> 589 590 <p>You can also provide a list of threads to backtrace, or the keyword 591 "all" to see all threads:</p> 592 593 <code> 594 (lldb) thread backtrace all 595 </code> 596 597 <p>You can select the current thread, which will be used by default in all the commands in 598 the next section, with the "thread select" command:</p> 599 600 <code> 601 (lldb) thread select 2 602 </code> 603 604 <p>where the thread index is just the one shown in the "<code>thread list</code>" listing. 605 606 </div> 607 <div class="postfooter"></div> 608 609 <div class="post"> 610 <h1 class ="postheader">Examining Stack Frame State</h1> 611 <div class="postcontent"> 612 613 614 <p>The most convenient way to inspect a frame's arguments and local variables is to use the "<code>frame variable</code>" command:</p> 615 616 <code> 617 (lldb) frame variable 618 <br>self = (SKTGraphicView *) 0x0000000100208b40 619 <br>_cmd = (struct objc_selector *) 0x000000010001bae1 620 <br>sender = (id) 0x00000001001264e0 621 <br>selection = (NSArray *) 0x00000001001264e0 622 <br>i = (NSUInteger) 0x00000001001264e0 623 <br>c = (NSUInteger) 0x00000001001253b0 624 </code> 625 626 <p>As you see above, if you don't specify any variable names, all arguments 627 and locals will be shown. If you call "<code>frame variable</code>" 628 passing in the names of a particular local(s), only those variables 629 will be printed. For instance: 630 </p> 631 632 <code> 633 (lldb) frame variable self 634 <br>(SKTGraphicView *) self = 0x0000000100208b40 635 </code> 636 637 <p>You can also pass in a path to some subelement of one of the available locals, 638 and that sub-element will be printed. For instance: 639 </p> 640 641 <code> 642 <br>(lldb) frame variable self.isa 643 <br>(struct objc_class *) self.isa = 0x0000000100023730 644 </code> 645 646 <p>The "<code>frame variable</code>" command is not a full expression 647 parser but it does support a few simple operations like &, *, ->, [] (no overloaded 648 operators). The array brackets can be used on pointers to treat pointers 649 as arrays:</p> 650 651 <code> 652 (lldb) frame variable *self 653 <br>(SKTGraphicView *) self = 0x0000000100208b40 654 <br>(NSView) NSView = { 655 <br>(NSResponder) NSResponder = { 656 <br>... 657 <br> 658 <br>(lldb) frame variable &self 659 <br>(SKTGraphicView **) &self = 0x0000000100304ab 660 <br> 661 <br>(lldb) frame variable argv[0] 662 <br>(char const *) argv[0] = 0x00007fff5fbffaf8 "/Projects/Sketch/build/Debug/Sketch.app/Contents/MacOS/Sketch" 663 </code> 664 665 <p>The frame variable command will also perform "object printing" operations on 666 variables (currently we only support ObjC printing, using the object's "description" method. 667 Turn this on by passing the -o flag to frame variable:</p> 668 669 <code> 670 (lldb) frame variable -o self 671 (SKTGraphicView *) self = 0x0000000100208b40 <SKTGraphicView: 0x100208b40> 672 </code> 673 674 <p>You can select another frame to view with the "<code>frame select</code>" command</p> 675 676 <code> 677 (lldb) frame select 9 678 <br>frame #9: 0x0000000100015ae3, where = Sketch`function1 + 33 at /Projects/Sketch/SKTFunctions.m:11 679 </code> 680 681 <p>You can also move up and down the stack by passing the "<code>--relative</code>" ("<code>-r</code>") 682 option. And we have built-in aliases "<code>u</code>" and "<code>d</code>" which 683 behave like their gdb equivalents. 684 685 <p>If you need to view more complex data or change program data, you can 686 use the general "expression" command. It takes an expression and 687 evaluates it in the scope of the currently selected frame. For instance:</p> 688 689 <code> 690 (lldb) expr self 691 <br>$0 = (SKTGraphicView *) 0x0000000100135430 692 <br>(lldb) expr self = 0x00 693 <br>$1 = (SKTGraphicView *) 0x0000000000000000 694 <br>(lldb) frame var self 695 <br>(SKTGraphicView *) self = 0x0000000000000000 696 </code> 697 698 <p>You can also call functions:</p> 699 700 <code> 701 (lldb) expr (int) printf ("I have a pointer 0x%llx.\n", self) 702 <br>$2 = (int) 22 703 <br>I have a pointer 0x0. 704 </code> 705 706 <p>As I said above, "expression" is one of the "raw" commands. So 707 you don't have to quote your whole expression, nor backslash protect quotes, 708 etc...</p> 709 710 <p>Finally, the results of the expressions are stored in persistent variables 711 (of the form $[0-9]+) that you can use in further expressions, like:</p> 712 713 <code> 714 (lldb) expr self = $0 715 <br>$4 = (SKTGraphicView *) 0x0000000100135430 716 </code> 717 <p> 718 </div> 719 <div class="postfooter"></div> 720 721 </div> 722 </div> 723 </div> 724</div> 725</body> 726</html> 727