| andybons | e6a8f2bd | 2015-08-31 22:46:01 | [diff] [blame] | 1 | # Tips for debugging on Linux |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 2 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 3 | This pageisforChromium-specific debugging tips; learning how to run gdbis |
| 4 | out of scope. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 5 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 6 | [TOC] |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 7 | |
| 8 | ## Symbolized stack trace |
| 9 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 10 | The sandbox can interferewith theinternal symbolizer.Use`--no-sandbox`(but |
| 11 | keepthis temporary)or an external symbolizer(see |
| 12 | `tools/valgrind/asan/asan_symbolize.py`). |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 13 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 14 | Generally,donotuse`--no-sandbox` on waterfall bots, sandbox testingis |
| 15 | needed.Talk to security@chromium.org. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 16 | |
| 17 | ## GDB |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 18 | |
| nodir | a6074d4c | 2015-09-01 04:26:45 | [diff] [blame] | 19 | *** promo |
| 20 | GDB-7.7is requiredin order to debugChrome onLinux. |
| 21 | *** |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 22 | |
| 23 | Any prior version will fail to resolve symbolsor segfault. |
| 24 | |
| Brett Wilson | 41a7de0 | 2023-04-27 18:56:54 | [diff] [blame] | 25 | ### Setup |
| 26 | |
| 27 | #### Build setup |
| 28 | |
| 29 | In your buildset the GN build variable`symbol_level = 2`for interactive |
| 30 | debugging.(`symbol_level = 1` only provides backtrace information).Andwhile |
| 31 | release-mode debuggingis possible, things will be much easierin a debug build. |
| 32 | Set your build argswith`gn args out/<your_dir>`(substituting your build |
| 33 | directory),andset: |
| 34 | |
| 35 | ``` |
| 36 | is_debug = true |
| 37 | symbol_level = 2 |
| 38 | ``` |
| 39 | |
| 40 | #### GDB setup |
| 41 | |
| 42 | TheChrome build requires some GDB configurationfor it to be able to find |
| Taiyo Mizuhashi | d0f90c9d | 2023-05-31 00:55:23 | [diff] [blame] | 43 | source files.See[gdbinit](../gdbinit.md) to configure GDB.Thereis a similar |
| 44 | processfor[LLDB](../lldbinit.md). |
| Brett Wilson | 41a7de0 | 2023-04-27 18:56:54 | [diff] [blame] | 45 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 46 | ### Basic browser process debugging |
| 47 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 48 | gdb-tui-ex=r--argsout/Debug/chrome--disable-seccomp-sandbox \ |
| 49 | http://google.com |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 50 | |
| 51 | ### Allowing attaching to foreign processes |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 52 | |
| 53 | On distributions thatuse the |
| 54 | [Yama LSM](https://www.kernel.org/doc/Documentation/security/Yama.txt) (that |
| 55 | includesUbuntuandChrome OS), process A can attach to process B onlyif Ais |
| 56 | an ancestor of B. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 57 | |
| 58 | You will probably want to disablethis featurebyusing |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 59 | |
| 60 | echo0| sudo tee/proc/sys/kernel/yama/ptrace_scope |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 61 | |
| 62 | If you don't you'llget an error message suchas"Could not attach to process". |
| 63 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 64 | Note that you'll also probably want to use `--no-sandbox`, as explained below. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 65 | |
| 66 | ### Multiprocess Tricks |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 67 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 68 | #### Getting renderer subprocesses into gdb |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 69 | |
| 70 | Since Chromium itself spawns the renderers, it can be tricky to grab a |
| 71 | particular with gdb. This command does the trick: |
| 72 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 73 | ``` |
| 74 | chrome --no-sandbox --renderer-cmd-prefix='xterm-title renderer-e gdb--args' |
| 75 | ``` |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 76 | |
| 77 | The `--no-sandbox` flag is needed because otherwise the seccomp sandbox will |
| 78 | kill the renderer process on startup, or the setuid sandbox will prevent xterm's |
| 79 | execution.The"xterm"is necessaryor gdb will runin the current terminal, |
| 80 | which canget particularly confusing since it's running in the background, and |
| 81 | if you're also running the main processin gdb, won't work at all (the two |
| 82 | instances will fight over the terminal). To auto-start the renderers in the |
| 83 | debugger, send the "run" command to the debugger: |
| 84 | |
| nodir | a6074d4c | 2015-09-01 04:26:45 | [diff] [blame] | 85 | chrome --no-sandbox --renderer-cmd-prefix='xterm-title renderer-e gdb \ |
| Zhang Hao | a9522488 | 2021-05-12 11:26:36 | [diff] [blame] | 86 | -ex run--args' |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 87 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 88 | If you'reusingEmacsand`M-x gdb`, you cando |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 89 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 90 | chrome"--renderer-cmd-prefix=gdb --args" |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 91 | |
| nodir | a6074d4c | 2015-09-01 04:26:45 | [diff] [blame] | 92 | *** note |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 93 | Note:using the`--renderer-cmd-prefix` option bypasses the zygote launcher, so |
| 94 | the renderers won't be sandboxed. It is generally not an issue, except when you |
| 95 | are trying to debug interactions with the sandbox. If that's what you are doing, |
| 96 | you will need to attach yourdebugger to a running renderer process(see below). |
| nodir | a6074d4c | 2015-09-01 04:26:45 | [diff] [blame] | 97 | *** |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 98 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 99 | You may also want topass`--disable-hang-monitor` to suppress the hang monitor, |
| 100 | whichis rather annoying. |
| 101 | |
| 102 | You can alsouse`--renderer-startup-dialog`and attach to the processin order |
| 103 | to debug the renderer code.Go to |
| xiaoyin.l | 1003c0b | 2016-12-06 02:51:17 | [diff] [blame] | 104 | https://www.chromium.org/blink/getting-started-with-blink-debugging for more |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 105 | information on howthis can bedone. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 106 | |
| Alex Gough | 2092674 | 2021-05-13 20:11:30 | [diff] [blame] | 107 | For utilities you canuse`--utility-startup-dialog` to have all utilities |
| 108 | prompt,or`--utility-startup-dialog=data_decoder.mojom.DataDecoderService` |
| 109 | to debug only a particular service type. |
| 110 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 111 | #### Choosing which renderers to debug |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 112 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 113 | If you are starting multiple renderersthen the above means that multiple gdb's |
| 114 | start and fight over the console. Instead, you can set the prefix to point to |
| 115 | this shell script: |
| 116 | |
| 117 | ```sh |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 118 | #!/bin/sh |
| 119 | |
| 120 | echo "**** Child $$ starting: y to debug" |
| 121 | read input |
| 122 | if [ "$input" = "y" ] ; then |
| 123 | gdb --args $* |
| 124 | else |
| 125 | $* |
| 126 | fi |
| 127 | ``` |
| 128 | |
| Robert Flack | cdbf8c4a | 2022-11-18 18:12:34 | [diff] [blame] | 129 | #### Choosing renderer to debug by URL |
| 130 | |
| 131 | In most cases you'll want to debug the renderer whichis loading a particular |
| 132 | site.If you want a script which will automatically debug the renderer which has |
| 133 | visited a given target URLandcontinue all other renderers, you canuse the |
| 134 | following: |
| 135 | |
| 136 | ```sh |
| 137 | ./third_party/blink/tools/debug_renderer out/Default/content_shell https://example.domain/path |
| 138 | ``` |
| 139 | |
| 140 | The script also supports specifying a different URL than the navigation URL. |
| 141 | Thisis usefulwhen the renderer you want to debugisnot the top frame but one |
| 142 | of the subframes on the page.For example, you could debug a particular subframe |
| 143 | on a pagewith: |
| 144 | |
| 145 | ```sh |
| 146 | ./third_party/blink/tools/debug_renderer -d https://subframe.url/path out/Default/content_shell https://example.domain/path |
| 147 | ``` |
| 148 | |
| 149 | However,if you need more fine-grained control over which renderers to debug |
| 150 | you can run chromeor content_shell directlywith the |
| 151 | `--wait-for-debugger-on-navigation` flag which will pause each renderer at the |
| 152 | point of navigation(when the URLis known). |
| 153 | |
| 154 | This will resultin a series of lines suchas the followingin the output: |
| 155 | ``` |
| 156 | ...:content_switches_internal.cc(119)] Renderer url="https://example.domain/path" (PID) paused waiting for debugger to attach. Send SIGUSR1 to unpause. |
| 157 | ``` |
| 158 | |
| 159 | You can signal the renderers you aren't interested in to continue running with: |
| 160 | ```sh |
| 161 | kill -s SIGUSR1 <pid> |
| 162 | ``` |
| 163 | |
| 164 | And debug the renderer you are interested in debugging with: |
| 165 | ```sh |
| 166 | gdb -p <pid> |
| 167 | ``` |
| 168 | |
| Robert Flack | 6bbefe7 | 2023-02-24 18:49:51 | [diff] [blame] | 169 | #### Debugging run_web_tests.py renderers |
| 170 | |
| 171 | The `debug_renderer` script can also be used to debug the renderer running |
| Jonathan Lee | e95877f | 2023-09-26 21:16:30 | [diff] [blame] | 172 | a web test. To do so, simply call `run_{web,wpt}_tests.py` from `debug_renderer` |
| 173 | with all of the standard arguments for `run_{web,wpt}_tests.py`. For example: |
| Robert Flack | 6bbefe7 | 2023-02-24 18:49:51 | [diff] [blame] | 174 | |
| 175 | ```sh |
| 176 | ./third_party/blink/tools/debug_renderer ./third_party/blink/tools/run_web_tests.py [run_web_test args] |
| 177 | ``` |
| 178 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 179 | #### Selective breakpoints |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 180 | |
| 181 | When debugging both the browser and renderer process, you might want to have |
| 182 | separate set of breakpoints to hit. You can use gdb's command files to |
| 183 | accomplishthisby putting breakpointsin separate filesand instructing gdb to |
| 184 | load them. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 185 | |
| 186 | ``` |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 187 | gdb -x ~/debug/browser --args chrome --no-sandbox --disable-hang-monitor \ |
| 188 | --renderer-cmd-prefix='xterm -title renderer -e gdb -x ~/debug/renderer \ |
| 189 | --args ' |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 190 | ``` |
| 191 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 192 | Also, instead of running gdb, you canuse the script above, whichlet's you |
| 193 | select which renderer process to debug. Note: you might need to use the full |
| 194 | path to the script and avoid `$HOME` or `~/.` |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 195 | |
| 196 | #### Connecting to a running renderer |
| 197 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 198 | Usually `ps aux | grep chrome` will not give very helpful output. Try |
| 199 | `pstree -p | grep chrome` to get something like |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 200 | |
| 201 | ``` |
| 202 | | |-bash(21969)---chrome(672)-+-chrome(694) |
| 203 | | | |-chrome(695)---chrome(696)-+-{chrome}(697) |
| 204 | | | | \-{chrome}(709) |
| 205 | | | |-{chrome}(675) |
| 206 | | | |-{chrome}(678) |
| 207 | | | |-{chrome}(679) |
| 208 | | | |-{chrome}(680) |
| 209 | | | |-{chrome}(681) |
| 210 | | | |-{chrome}(682) |
| 211 | | | |-{chrome}(684) |
| 212 | | | |-{chrome}(685) |
| 213 | | | |-{chrome}(705) |
| 214 | | | \-{chrome}(717) |
| 215 | ``` |
| 216 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 217 | Most of those are threads. In this case the browser process would be 672 and the |
| 218 | (sole) renderer process is 696. You can use `gdb -p 696` to attach. |
| 219 | Alternatively, you might find out the process ID from Chrome's built-inTask |
| 220 | Manager(under theTools menu).Right-click on theTaskManager,and enable |
| 221 | "Process ID"in the list of columns. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 222 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 223 | Note:bydefault, sandboxed processes can't be attached by a debugger. To be |
| 224 | able to do so, you will need to pass the `--allow-sandbox-debugging` option. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 225 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 226 | If the problem only occurs with the seccomp sandbox enabled (and the previous |
| 227 | tricks don't help), you couldtry enabling core-dumps(see the**Core files** |
| 228 | section).That would allow you toget a backtraceand see somelocal variables, |
| 229 | though you won't be able to step through the running program. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 230 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 231 | Note: If you're interestedin debuggingLinuxSandboxIPC process, you can attach |
| 232 | to694in the above diagram.TheLinuxSandboxIPC process has the same command |
| 233 | line flagas the browser process so that it's easy to identify it if you run |
| 234 | `pstree -pa`. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 235 | |
| 236 | #### Getting GPU subprocesses into gdb |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 237 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 238 | Use `--gpu-launcher` flag instead of `--renderer-cmd-prefix` in the instructions |
| 239 | for renderer above. |
| 240 | |
| 241 | #### Getting `browser_tests` launched browsers into gdb |
| 242 | |
| 243 | Use environment variable `BROWSER_WRAPPER` instead of `--renderer-cmd-prefix` |
| 244 | switch in the instructions above. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 245 | |
| 246 | Example: |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 247 | |
| 248 | ```shell |
| 249 | BROWSER_WRAPPER='xterm-title renderer-e gdb--eval-command=run \ |
| 250 | --eval-command=quit--args' out/Debug/browser_tests --gtest_filter=Print |
| 251 | ``` |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 252 | |
| 253 | #### Plugin Processes |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 254 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 255 | Same strategies as renderers above, but the flag is called `--plugin-launcher`: |
| 256 | |
| 257 | chrome --plugin-launcher='xterm-e gdb--args' |
| 258 | |
| nodir | a6074d4c | 2015-09-01 04:26:45 | [diff] [blame] | 259 | *** note |
| 260 | Note: For now, this does not currently apply to PPAPI plugins because they |
| 261 | currently run in the renderer process. |
| 262 | *** |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 263 | |
| 264 | #### Single-Process mode |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 265 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 266 | Depending on whether it's relevant to the problem, it's often easier to just run |
| 267 | in "single process" mode where the renderer threads are in-process. Then you can |
| 268 | just run gdb on the main process. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 269 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 270 | gdb --args chrome --single-process |
| 271 | |
| 272 | Currently, the `--disable-gpu` flag is also required, as there are known crashes |
| 273 | that occur under TextureImageTransportSurface without it. The crash described in |
| xiaoyin.l | 1003c0b | 2016-12-06 02:51:17 | [diff] [blame] | 274 | https://crbug.com/361689 can also sometimes occur, but that crash can be |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 275 | continued from without harm. |
| 276 | |
| 277 | Note that for technical reasons plugins cannot be in-process, so |
| 278 | `--single-process` only puts the renderers in the browser process. The flag is |
| 279 | still useful for debugging plugins (since it's only two processes instead of |
| 280 | three) but you'll still need to use `--plugin-launcher` or another approach. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 281 | |
| 282 | ### Printing Chromium types |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 283 | |
| Tom Anderson | f06ac38 | 2019-04-10 03:49:38 | [diff] [blame] | 284 | gdb 7 lets us use Python to write pretty-printers for Chromium types. See |
| Taiyo Mizuhashi | d0f90c9d | 2023-05-31 00:55:23 | [diff] [blame] | 285 | [gdbinit](../gdbinit.md) |
| Tom Anderson | f06ac38 | 2019-04-10 03:49:38 | [diff] [blame] | 286 | to enable pretty-printing of Chromium types. This will import Blink |
| 287 | pretty-printers as well. |
| Kenichi Ishibashi | e17b8d9f | 2018-04-26 03:32:46 | [diff] [blame] | 288 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 289 | Pretty printers for std types shouldn't be necessaryin gdb7, but they're |
| 290 | provided here in case you'reusing an older gdb.Put the followinginto |
| 291 | `~/.gdbinit`: |
| 292 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 293 | ``` |
| 294 | # Print a C++ string. |
| 295 | define ps |
| 296 | print $arg0.c_str() |
| 297 | end |
| 298 | |
| 299 | # Print a C++ wstring or wchar_t*. |
| 300 | define pws |
| 301 | printf "\"" |
| 302 | set $c = (wchar_t*)$arg0 |
| 303 | while ( *$c ) |
| 304 | if ( *$c > 0x7f ) |
| 305 | printf "[%x]", *$c |
| 306 | else |
| 307 | printf "%c", *$c |
| 308 | end |
| 309 | set $c++ |
| 310 | end |
| 311 | printf "\"\n" |
| 312 | end |
| 313 | ``` |
| 314 | |
| 315 | [More STL GDB macros](http://www.yolinux.com/TUTORIALS/src/dbinit_stl_views-1.01.txt) |
| 316 | |
| Christian Biesinger | 3332bb3a | 2019-08-13 05:45:23 | [diff] [blame] | 317 | ### JsDbg -- visualize data structures in the browser |
| 318 | |
| 319 | JsDbgis adebugger plugin to display variousChrome data structuresin a |
| 320 | browser window, suchas the accessibility tree, layoutobject tree, DOM tree, |
| 321 | and others. |
| 322 | [Installation instructions are here](https://github.com/MicrosoftEdge/JsDbg), |
| 323 | and see[here](https://github.com/MicrosoftEdge/JsDbg/blob/master/docs/FEATURES.md) |
| 324 | for screenshotsand an introduction. |
| 325 | |
| 326 | ForGooglers, please see[go/jsdbg](https://goto.google.com/jsdbg) for |
| 327 | installation instructions. |
| 328 | |
| 329 | ### Time travel debugging with rr |
| 330 | |
| 331 | You canuse[rr](https://rr-project.org) for time travel debugging, so you |
| 332 | can also stepor execute backwards.This worksby first recording a trace |
| L. David Baron | e99d91eb | 2021-03-30 20:18:00 | [diff] [blame] | 333 | andthen debugging based on that. |
| Christian Biesinger | 3332bb3a | 2019-08-13 05:45:23 | [diff] [blame] | 334 | |
| Jie Sheng | 01180e6 | 2025-03-29 03:48:01 | [diff] [blame] | 335 | ForGooglers,if you have a remote cloud machine, please followthis |
| 336 | [instruction](https://engdoc.corp.google.com/eng/doc/devguide/debugging/rr.md#setting-up-rr) |
| 337 | toset up the machinein order touse the rr tool. |
| 338 | |
| Steve Kobes | 8ce3e44d | 2022-01-28 22:36:59 | [diff] [blame] | 339 | You need an up-to-date version of rr, since rris frequently updated to support |
| 340 | new parts of theLinux system call API surface thatChromium uses.If you have |
| 341 | any issueswith the latest release version,try compiling rr |
| L. David Baron | e99d91eb | 2021-03-30 20:18:00 | [diff] [blame] | 342 | [from source](https://github.com/rr-debugger/rr/wiki/Building-And-Installing). |
| Robert Flack | c9e6995 | 2020-05-13 19:52:31 | [diff] [blame] | 343 | |
| Christian Biesinger | 3332bb3a | 2019-08-13 05:45:23 | [diff] [blame] | 344 | Once installed, you canuse it likethis: |
| 345 | ``` |
| Steve Kobes | 8ce3e44d | 2022-01-28 22:36:59 | [diff] [blame] | 346 | rr record out/Debug/content_shell --single-process |
| Christian Biesinger | 3332bb3a | 2019-08-13 05:45:23 | [diff] [blame] | 347 | rr replay |
| Steve Kobes | 8ce3e44d | 2022-01-28 22:36:59 | [diff] [blame] | 348 | (rr) c |
| 349 | (rr) break blink::NGBlockNode::Layout |
| 350 | (rr) rc # reverse-continue to the last Layout call |
| 351 | (rr) jsdbg # run JsDbg as described above to find the interesting object |
| 352 | (rr) watch -l box_->frame_rect_.size_.width_.value_ |
| 353 | (rr) rc # reverse-continue to the last time the width was changed |
| 354 | (rr) rn # reverse-next to the previous line |
| 355 | (rr) reverse-fin # run to where this function was called from |
| Christian Biesinger | 3332bb3a | 2019-08-13 05:45:23 | [diff] [blame] | 356 | ``` |
| 357 | |
| L. David Baron | e99d91eb | 2021-03-30 20:18:00 | [diff] [blame] | 358 | You can debug multi-process chromeusing`rr -f [PID]` |
| Steve Kobes | 8ce3e44d | 2022-01-28 22:36:59 | [diff] [blame] | 359 | for processes`fork()`edfrom a[zygote process](zygote.md) withoutexec, |
| 360 | which includes renderer processes, |
| 361 | or`rr -p [PID]`for other processes. |
| 362 | To find the process id you can either run`rr ps` after recording,orfor |
| 363 | renderer processesuse`--vmodule=render_frame_impl=1` which will log a |
| 364 | message on navigations.Example: |
| Robert Flack | e13e0b1 | 2020-04-16 17:03:58 | [diff] [blame] | 365 | |
| 366 | ``` |
| Steve Kobes | 8ce3e44d | 2022-01-28 22:36:59 | [diff] [blame] | 367 | $ rr record out/Debug/content_shell --disable-hang-monitor --vmodule=render_frame_impl=1 https://www.google.com/ |
| Robert Flack | e13e0b1 | 2020-04-16 17:03:58 | [diff] [blame] | 368 | rr: Saving execution to trace directory `...'. |
| 369 | ... |
| 370 | [128515:128515:0320/164124.768687:VERBOSE1:render_frame_impl.cc(4244)] Committed provisional load: https://www.google.com/ |
| 371 | ``` |
| 372 | |
| 373 | From the log message we can see that the site was loaded into process 128515 |
| 374 | and can set a breakpoint for when that process is forked. |
| 375 | |
| 376 | ``` |
| 377 | rr replay -f 128515 |
| 378 | ``` |
| 379 | |
| L. David Baron | e99d91eb | 2021-03-30 20:18:00 | [diff] [blame] | 380 | If you want to call debugging functions from gdb that use `LOG()`, |
| Steve Kobes | 8ce3e44d | 2022-01-28 22:36:59 | [diff] [blame] | 381 | then those functions need to disable the printing of timestamps using |
| 382 | [`SetLogItems`](https://source.chromium.org/search?q=SetLogItems&sq=&ss=chromium%2Fchromium%2Fsrc). |
| 383 | See `LayoutObject::ShowLayoutObject()` for an example of this, and |
| 384 | [issue 2829](https://github.com/rr-debugger/rr/issues/2829) for why it is needed. |
| L. David Baron | e99d91eb | 2021-03-30 20:18:00 | [diff] [blame] | 385 | |
| Steve Kobes | 8ce3e44d | 2022-01-28 22:36:59 | [diff] [blame] | 386 | If rr doesn't work correctly, the rr developers are generally quite responsive |
| 387 | to[bug reports](https://github.com/rr-debugger/rr/issues), |
| L. David Baron | e99d91eb | 2021-03-30 20:18:00 | [diff] [blame] | 388 | especially ones that have enough information so that |
| 389 | they don't have to build Chromium. |
| 390 | |
| 391 | See Also: |
| Steve Kobes | 8ce3e44d | 2022-01-28 22:36:59 | [diff] [blame] | 392 | |
| L. David Baron | e99d91eb | 2021-03-30 20:18:00 | [diff] [blame] | 393 | * [The Chromium Chronicle #13: Time-Travel Debugging with RR](https://developer.chrome.com/blog/chromium-chronicle-13/) |
| Steve Kobes | 8ce3e44d | 2022-01-28 22:36:59 | [diff] [blame] | 394 | * [@davidbaron demo using rr](https://twitter.com/davidbaron/status/1473761042278887433) |
| 395 | * [@davidbaron demo using pernosco](https://twitter.com/davidbaron/status/1475836824409022469) |
| 396 | (Googlers: see [go/pernosco](https://goto.google.com/pernosco)) |
| L. David Baron | e99d91eb | 2021-03-30 20:18:00 | [diff] [blame] | 397 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 398 | ### Graphical Debugging Aid for Chromium Views |
| 399 | |
| 400 | The following link describes a tool that can be used on Linux, Windows and Mac under GDB. |
| 401 | |
| sisidovski | f270241c | 2021-08-04 07:07:44 | [diff] [blame] | 402 | [graphical_debugging_aid_chromium_views](../graphical_debugging_aid_chromium_views.md) |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 403 | |
| 404 | ### Faster startup |
| 405 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 406 | Use the `gdb-add-index` script (e.g. |
| 407 | `build/gdb-add-index out/Debug/browser_tests`) |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 408 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 409 | Only makes sense if you run the binary multiple times or maybe if you use the |
| nodir | a6074d4c | 2015-09-01 04:26:45 | [diff] [blame] | 410 | component build since most `.so` files won'trequire reindexing on a rebuild. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 411 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 412 | See |
| 413 | https://groups.google.com/a/chromium.org/forum/#!searchin/chromium-dev/gdb-add-index/chromium-dev/ELRuj1BDCL4/5Ki4LGx41CcJ |
| 414 | for more info. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 415 | |
| Takuto Ikuta | a093dc2 | 2020-11-11 03:17:50 | [diff] [blame] | 416 | You can improve GDB load time significantly at the cost of link timebynot |
| brettw | 20d800c | 2016-04-12 00:10:49 | [diff] [blame] | 417 | splitting symbolsfrom theobject files.In GN,set`use_debug_fission=false`in |
| 418 | your"gn args". |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 419 | |
| 420 | ## Core files |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 421 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 422 | `ulimit -c unlimited` should cause allChrome processes(runfrom that shell) to |
| 423 | dump cores,with the possible exception of some sandboxed processes. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 424 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 425 | Some sandboxed subprocesses mightnotdump coresunless youpass the |
| 426 | `--allow-sandbox-debugging` flag. |
| 427 | |
| 428 | If the problemis a freeze rather than a crash, you may be able to trigger a |
| 429 | core-dumpby sending SIGABRT to the relevant process: |
| 430 | |
| 431 | kill-6[process id] |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 432 | |
| 433 | ## Breakpad minidump files |
| 434 | |
| Tom Anderson | abdbd6a | 2020-01-09 16:59:27 | [diff] [blame] | 435 | See[minidump_to_core.md](minidump_to_core.md) |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 436 | |
| 437 | ## Running Tests |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 438 | |
| 439 | Many ofour tests bring up windows on screen.This can be annoying(they steal |
| 440 | your focus)and hard to debug(they receive extra eventsas you mouse over them). |
| 441 | Instead,use`Xvfb`or`Xephyr` to run a nested X session to debug them,as |
| sisidovski | f270241c | 2021-08-04 07:07:44 | [diff] [blame] | 442 | outlined on[testing/web_tests_linux.md](../testing/web_tests_linux.md). |
| Brett Wilson | 41a7de0 | 2023-04-27 18:56:54 | [diff] [blame] | 443 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 444 | ### Browser tests |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 445 | |
| 446 | Bydefault the`browser_tests` forks anew browserfor each test.To debug the |
| 447 | browser side of a single test,use a command like |
| 448 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 449 | ``` |
| Thomas Lukaszewicz | 2c5fb614 | 2019-10-14 19:20:05 | [diff] [blame] | 450 | gdb --args out/Debug/browser_tests --single-process-tests --gtest_filter=MyTestName |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 451 | ``` |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 452 | |
| Thomas Lukaszewicz | 2c5fb614 | 2019-10-14 19:20:05 | [diff] [blame] | 453 | **note theuse of`single-process-tests`**--this makes the test harnessand |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 454 | browser process share the outermost process. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 455 | |
| Brett Wilson | 41a7de0 | 2023-04-27 18:56:54 | [diff] [blame] | 456 | Theswitch`--gtest_break_on_failure` can also be useful to automatically stop |
| 457 | debugger upon`ASSERT`or`EXPECT` failures. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 458 | |
| 459 | To debug a renderer processinthiscase,use the tips above about renderers. |
| 460 | |
| Kent Tamura | 59ffb02 | 2018-11-27 05:30:56 | [diff] [blame] | 461 | ### Web tests |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 462 | |
| sisidovski | f270241c | 2021-08-04 07:07:44 | [diff] [blame] | 463 | See[testing/web_tests_linux.md](../testing/web_tests_linux.md)for some tips.In particular, |
| Kent Tamura | 59ffb02 | 2018-11-27 05:30:56 | [diff] [blame] | 464 | note that it's possible to debug a web test via `ssh`ing to a Linux box; you |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 465 | don't need anything on screenif youuse`Xvfb`. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 466 | |
| 467 | ### UI tests |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 468 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 469 | UI tests are runin forked browsers.Unlike browser tests, you cannotdo any |
| 470 | single process tricks here to debug the browser.See below about |
| 471 | `BROWSER_WRAPPER`. |
| 472 | |
| 473 | Topass flags to the browser,use a command line like |
| 474 | `--extra-chrome-flags="--foo --bar"`. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 475 | |
| 476 | ### Timeouts |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 477 | |
| 478 | UI tests have a confusing array of timeoutsin place.(Pawelis working on |
| 479 | reducing the number of timeouts.)To disable themwhile you debug,set the |
| 480 | timeout flags to a large value: |
| 481 | |
| 482 | *`--test-timeout=100000000` |
| 483 | *`--ui-test-action-timeout=100000000` |
| 484 | *`--ui-test-terminate-timeout=100000000` |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 485 | |
| 486 | ### To replicate Window Manager setup on the bots |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 487 | |
| 488 | Chromiumtry botsand main waterfall's bots run tests under Xvfb&openbox |
| 489 | combination. Xvfb is an X11 server that redirects the graphical output to the |
| 490 | memory, and openbox is a simple window manager that is running on top of Xvfb. |
| 491 | The behavior of openbox is markedly different when it comes to focus management |
| 492 | and other window tasks, so test that runs fine locally may fail or be flaky on |
| 493 | try bots. To run the tests on a local machine as on a bot, follow these steps: |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 494 | |
| 495 | Make sure you have openbox: |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 496 | |
| 497 | apt-get install openbox |
| 498 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 499 | Start Xvfb and openbox on a particular display: |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 500 | |
| 501 | Xvfb :6.0 -screen 0 1280x1024x24 & DISPLAY=:6.0 openbox & |
| 502 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 503 | Run your tests with graphics output redirected to that display: |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 504 | |
| 505 | DISPLAY=:6.0 out/Debug/browser_tests --gtest_filter="MyBrowserTest.MyActivateWindowTest" |
| 506 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 507 | You can look at a snapshot of the output by: |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 508 | |
| 509 | xwd -display :6.0 -root | xwud |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 510 | |
| 511 | Alternatively, you can use testing/xvfb.py to set up your environment for you: |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 512 | |
| thomasanderson | 3d07428 | 2016-12-06 18:21:12 | [diff] [blame] | 513 | testing/xvfb.py out/Debug/browser_tests \ |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 514 | --gtest_filter="MyBrowserTest.MyActivateWindowTest" |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 515 | |
| nodir | a6074d4c | 2015-09-01 04:26:45 | [diff] [blame] | 516 | ### BROWSER_WRAPPER |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 517 | |
| 518 | You can also get the browser under a debugger by setting the `BROWSER_WRAPPER` |
| 519 | environment variable. (You can use this for `browser_tests` too, but see above |
| 520 | for discussion of a simpler way.) |
| 521 | |
| 522 | BROWSER_WRAPPER='xterm-e gdb--args' out/Debug/browser_tests |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 523 | |
| qyearsley | c0dc6f4 | 2016-12-02 22:13:39 | [diff] [blame] | 524 | ### Replicating try bot Slowness |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 525 | |
| qyearsley | c0dc6f4 | 2016-12-02 22:13:39 | [diff] [blame] | 526 | Try bots are pretty stressed, and can sometimes expose timing issues you can't |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 527 | normally reproduce locally. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 528 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 529 | You can simulatethisby shutting down all but one of theCPUs |
| 530 | (http://www.cyberciti.biz/faq/debian-rhel-centos-redhat-suse-hotplug-cpu/) and |
| 531 | running a CPU loading tool(e.g., http://www.devin.com/lookbusy/). Now run your |
| qyearsley | c0dc6f4 | 2016-12-02 22:13:39 | [diff] [blame] | 532 | test.It will run slowly, but any flakiness foundby thetry bot should replicate |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 533 | locally now-and often nearly100% of the time. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 534 | |
| 535 | ## Logging |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 536 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 537 | ### Seeing all LOG(foo) messages |
| 538 | |
| 539 | Default log level hides`LOG(INFO)`.Runwith`--log-level=0`and |
| 540 | `--enable-logging=stderr` flags. |
| 541 | |
| qyearsley | c0dc6f4 | 2016-12-02 22:13:39 | [diff] [blame] | 542 | Newer versions ofChromiumwith VLOG may need--v=1 too.For more VLOG tips, see |
| xiaoyin.l | 1003c0b | 2016-12-06 02:51:17 | [diff] [blame] | 543 | [the chromium-dev thread](https://groups.google.com/a/chromium.org/group/chromium-dev/browse_thread/thread/dcd0cd7752b35de6?pli=1). |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 544 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 545 | ## Profiling |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 546 | |
| 547 | See |
| 548 | https://sites.google.com/a/chromium.org/dev/developers/profiling-chromium-and-webkit |
| Tom Anderson | 93e49e49 | 2019-12-23 19:55:37 | [diff] [blame] | 549 | and[LinuxProfiling](profiling.md). |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 550 | |
| 551 | ## i18n |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 552 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 553 | We obey your system locale.Try something like: |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 554 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 555 | LANG=ja_JP.UTF-8out/Debug/chrome |
| 556 | |
| 557 | Ifthis doesn't work, make sure that the `LANGUAGE`, `LC_ALL` and `LC_MESSAGE` |
| 558 | environment variables aren'tset-- they have higher priority than LANGin the |
| 559 | order listed.Alternatively, justdothis: |
| 560 | |
| 561 | LANGUAGE=frout/Debug/chrome |
| 562 | |
| 563 | Note that because weuse GTK, some locale data comesfrom the system--for |
| 564 | example, file save boxesand whether the current languageis considered RTL. |
| 565 | Without all the language data available,Chrome willuse a mixture of your |
| 566 | system languageand the language you runChromein. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 567 | |
| 568 | Here's how to install the Arabic (ar) and Hebrew (he) language packs: |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 569 | |
| 570 | sudo apt-get install language-pack-ar language-pack-he \ |
| 571 | language-pack-gnome-ar language-pack-gnome-he |
| 572 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 573 | Note that the `--lang` flag does **not** work properly for this. |
| 574 | |
| Tom Anderson | 287339e | 2018-08-22 21:52:02 | [diff] [blame] | 575 | On non-Debian systems, you need the `gtk30.mo` files. (Please update these docs |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 576 | with the appropriate instructions if you know what they are.) |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 577 | |
| 578 | ## Breakpad |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 579 | |
| Tom Anderson | 93e49e49 | 2019-12-23 19:55:37 | [diff] [blame] | 580 | See the last section of [Linux Crash Dumping](crash_dumping.md). |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 581 | |
| 582 | ## Drag and Drop |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 583 | |
| 584 | If you break in a debugger during a drag, Chrome will have grabbed your mouse |
| 585 | and keyboard so you won't be able to interactwith thedebugger!To work around |
| 586 | this, run via`Xephyr`.Instructionsfor how touse`Xephyr` are on the |
| sisidovski | f270241c | 2021-08-04 07:07:44 | [diff] [blame] | 587 | [Running web tests onLinux](../testing/web_tests_linux.md) page. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 588 | |
| 589 | ## Tracking Down Bugs |
| 590 | |
| 591 | ### Isolating Regressions |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 592 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 593 | Old builds are archived here: |
| xiaoyin.l | 1003c0b | 2016-12-06 02:51:17 | [diff] [blame] | 594 | https://build.chromium.org/buildbot/snapshots/chromium-rel-linux/ |
| nodir | a6074d4c | 2015-09-01 04:26:45 | [diff] [blame] | 595 | (TODO: doesnot exist). |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 596 | |
| 597 | `tools/bisect-builds.py`in the tree automates bisecting through the archived |
| 598 | builds.Despite a computer science education, I am still amazed how quickly |
| 599 | binary search will find its target. |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 600 | |
| 601 | ### Screen recording for bug reports |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 602 | |
| 603 | sudo apt-get install gtk-recordmydesktop |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 604 | |
| 605 | ## Version-specific issues |
| 606 | |
| 607 | ### Google Chrome |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 608 | |
| 609 | GoogleChrome binaries don't include symbols. Googlers can read where to get |
| 610 | symbols from |
| 611 | [the Google-internal wiki](http://wiki/Main/ChromeOfficialBuildLinux#The_Build_Archive). |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 612 | |
| 613 | ### Ubuntu Chromium |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 614 | |
| 615 | Since we don't build theUbuntu packages(Ubuntu does) we can't get useful |
| 616 | backtraces from them. Direct users to https://wiki.ubuntu.com/Chromium/Debugging |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 617 | |
| 618 | ### Fedora'sChromium |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 619 | |
| 620 | LikeUbuntu, but direct users to |
| 621 | https://fedoraproject.org/wiki/TomCallaway/Chromium_Debug |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 622 | |
| 623 | ### Xlib |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 624 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 625 | If you're trying to track down X errors like: |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 626 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 627 | ``` |
| 628 | The program 'chrome' received an X Window System error. |
| 629 | This probably reflects a bug in the program. |
| 630 | The error was 'BadDrawable(invalidPixmaporWindow parameter)'. |
| 631 | ``` |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 632 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 633 | Some strategies are: |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 634 | |
| 635 | * pass `--sync` on the command line to make all X calls synchronous |
| 636 | * run chrome via [xtrace](http://xtrace.alioth.debian.org/) |
| 637 | * turn on IPC debugging (see above section) |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 638 | |
| 639 | ### Window Managers |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 640 | |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 641 | To test on various window managers, you can use a nested X server like `Xephyr`. |
| 642 | Instructions for how to use `Xephyr` are on the |
| sisidovski | f270241c | 2021-08-04 07:07:44 | [diff] [blame] | 643 | [Running web tests on Linux](../testing/web_tests_linux.md) page. |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 644 | |
| 645 | If you need to test something with hardware accelerated compositing |
| 646 | (e.g., compiz), you can use `Xgl` (`sudo apt-get install xserver-xgl`). E.g.: |
| 647 | |
| 648 | Xgl :1 -ac -accel glx:pbuffer -accel xv:pbuffer -screen 1024x768 |
| 649 | |
| andybons | 3322f76 | 2015-08-24 21:37:09 | [diff] [blame] | 650 | ## Mozilla Tips |
| andybons | ad92aa3 | 2015-08-31 02:27:44 | [diff] [blame] | 651 | |
| 652 | https://developer.mozilla.org/en/Debugging_Mozilla_on_Linux_FAQ |
| Joshua Pawlicki | c2464187d | 2022-02-23 14:27:51 | [diff] [blame] | 653 | |
| 654 | ## Google Chrome Symbol Files |
| 655 | |
| 656 | Symbols for Google Chrome's official builds are availablefrom |
| 657 | `https://edgedl.me.gvt1.com/chrome/linux/symbols/google-chrome-debug-info-linux64-${VERSION}.zip` |
| 658 | where ${VERSION}is any version ofGoogleChrome that has recently been served |
| Robert Flack | cdbf8c4a | 2022-11-18 18:12:34 | [diff] [blame] | 659 | toStable,Beta,orUnstable(Dev) channels onLinux. |
