Debugging

One of the main advantages of debugging cross-platform Emscripten code is that the same cross-platform source code can be debugged on either the native platform or using the web browser’s increasingly powerful toolset — including a debugger, profiler, and other tools.

This article describes the main tools and settings provided by Emscripten for debugging, organized by common developer use cases.

Overview: Emitting and Controlling Debug Information

Debugging-related information comes in several forms: in Wasm object and binary files (as DWARF sections or Wasm name section), side output files (as source maps, symbol maps, or DWARF sidecar or package files), and even in the code itself (as assertions or instrumentation, or JS whitespace and comments). For information on DWARF, see below. In addition to DWARF, wasm files may contain a name section which includes names for each function; these function names are displayed by browsers when they generate stack traces and in developer tools. Source maps are also supported by Emscripten and by browser DevTools (see below).

This document contains an overview of the flags used to emit and control debugging behavior, and use-case-based examples.

Unlike traditional Unix-style C toolchains, flags must be passed at link time to preserve or generate debug information (these defaults aim to avoid unintended bloat in production builds). The most common of these are the -g flags; see the flag documentation or the use cases below for more detail.

Flags that cause DWARF generation (e.g. -g3, -gline-tables-only) also generate a name section in the binary and suppress minification of the JS glue file (since most DWARF use cases are for interactive debugging or where the binary will be stripped). Other flags (e.g. -g2, -gsource-map) should affect only a specific behavior or type of debug info, and are generally composable.

Interactive, Source-Level Debugging

For stepping through C/C++ source code in a browser’s debugger, you can use debug information in either DWARF or source map format.

DWARF offers the best debugging experience and is supported in Chrome with an extension. See here for a detailed usage guide. Source maps are more widely supported, but they provide only location mapping and cannot be used easily to inspect variables.

DWARF

In a traditional Unix-style C toolchain, flags such as -g are passed to the compiler, placing DWARF sections in the object files. This DWARF info is combined by the linker and appears in the output, independently of any optimization settings. In contrast, although emcc supports many of the common clang flags to generate DWARF into the object files, final debug output is also controlled by link-time flags, and is more affected by optimization. For example emcc strips out most of the debug information after linking if a debugging-related flag is not provided at link time, even if the input object files contain DWARF.

DWARF can be produced at compile time with the emcc -g flag. Optimization levels above -O1 or -Og increasingly degrade LLVM debug information (as with other architectures), and optimization flags at link time also disable Emscripten’s runtime ASSERTIONS checks. Passing a -g flag at link time also affects the generated JavaScript code (preserving white-space, function names, and variable names, which makes the JavaScript debuggable).

The -g flag can also be specified with integer levels: -g0, -g1, -g2, and -g3 (equivalent to -g). At compile time these flags control the amount of DWARF in the object files. At link time, each adds sucessively more kinds of information in the wasm and JS files (DWARF is only retained after linking when using -g or -g3).

Example:

emcc source.c -c -o source.o -g # source.o has DWARF sections emcc source.o -o
program.js -g # program.wasm has DWARF and a name section

Tip

Even for medium-sized projects, DWARF debug information can be large. Debug information can be emitted in a separate file with the -gseparate-dwarf option. To speed up linking, the -gsplit-dwarf option can be used at compile time. See this article for more details on debugging large files, and see the next section for more ways to reduce debug info size.

Note

Because Binaryen optimization degrades the quality of DWARF info further, higher link-time optimization settings are not recommended. The -O1 setting will skip running the Binaryen optimizer (wasm-opt) entirely unless required by other options. You can also add the -sERROR_ON_WASM_CHANGES_AFTER_LINK option if you want to ensure the debug info is preserved. See Skipping Binaryen for more details.

Symbolizing Production Crash Logs

Even when not using an interactive debugger, it’s valuable to have source information for compiled code locations, particularly for stack traces or crash logs. This is also true for fully-optimized production builds.

Source maps are commonly used for languages that compile to JavaScript (mapping locations in the compiled JS output to locations in the original source code), but WebAssembly is also supported. Emscripten can emit source maps with the -gsource-map link-time flag. Source maps are preserved even with full post-link optimizations, so they work well for this use case. Source maps are generated by Emscripten from DWARF information. Therefore the linked object files must have DWARF. The final linked output will not have DWARF unless -g is also passed at link time.

DWARF can also be used for this purpose. Typically a binary containing DWARF would be generated at build time, and then stripped. The stripped copy would be served to users, and the original would be saved for symbolication purposes. For this use case, full information about about types and variables from the sources isn’t needed; the -gline-tables-only compile-time flag causes clang to generate only the line table information, saving DWARF size and compile/linking time.

Source maps are easier to parse and more widely supported by ecosystem tooling. And as noted above, preserving DWARF inhibits some Binaryen optimizations. However DWARF has the advantage that it includes information about inlining, which can result in more accurate stack traces.

Examples:

emcc source.c -c -o source.o -g # source.o has DWARF sections (-gsource-map also works here)
emcc source.o -o program.js -gsource-map # program.wasm.map contains a source map

emcc source.o -o program2.js -g # program2.wasm has DWARF
llvm-strip program2.wasm -o program2_stripped.wasm # program2_stripped.wasm has no debug info

Emscripten includes a tool called emsymbolizer that can map wasm code addresses to sources using several different kinds of debug info, including DWARF (in wasm object or linked files) and source maps for line/column info, and symbol maps (see --emit-symbol-map), name sections and object file symbol tables for function names.

Fast Edit+Compile with minimal debug information

When you want the fastest builds, you generally want to avoid generating large debug information during compile, because it takes time to link into the final binary. It is still worthwhile to use the -g2 flag (at link time only) because browsers understand the name section even when devtools are not in use, resulting in more useful stack traces at minimal cost.

Example:

emcc source.c -c -o source.o # source.o has no debug info
emcc source.o -o program.js -g2 # program.wasm has a name section, program.js is unminified

Sometimes the use of the -O1 or -Og flag at compile time can also result in faster builds, because optimizations early in the pipeline can reduce the amount of IR that is processed by later phases such as instruction selection and linking. It also of course reduces test runtime.

Detecting Memory Errors and Undefined Behavior

The best tools for detecting memory safety and undefined behavior issues are Clang’s sanitizers, such as the Undefined Behavior Sanitizer (UBSan) and the Address Sanitizer (ASan). For more information, see Debugging with Sanitizers.

Emscripten has several other compiler settings that can be useful for catching errors at runtime. These are set using the emcc -s option. For example:

emcc -O1 -sASSERTIONS test/hello_world.c

Some important settings are:

• ASSERTIONS=1 is used to enable runtime checks for many types of common errors. It also defines how Emscripten should handle errors in program flow. The value can be set to ASSERTIONS=2 in order to run additional tests. ASSERTIONS=1 is enabled by default at -O0 and disabled at higher optimization levels, but can be overridden.

• SAFE_HEAP=1 adds additional memory access checks with a Binaryen pass, and will give clear errors for problems like dereferencing 0 and memory alignment issues. You can also set SAFE_HEAP_LOG to log SAFE_HEAP operations. ASan provides most of the functionality of this pass (plus some extras) and is generally preferred to try first unless alignment issues are important for your platform.

• STACK_OVERFLOW_CHECK=1 adds a runtime magic token value at the end of the stack, which is checked in certain locations to verify that the user code does not accidentally write past the end of the stack. While overrunning the Emscripten stack is not a security issue for JavaScript (which is unaffected), writing past the stack causes memory corruption in global data and dynamically allocated memory sections in the Emscripten HEAP, which makes the application fail in unexpected ways. The value STACK_OVERFLOW_CHECK=2 enables slightly more detailed stack guard checks, which can give a more precise callstack at the expense of some performance. Default value is 1 if ASSERTIONS=1 is set, and disabled otherwise.

A number of other useful debug settings are defined in src/settings.js. For more information, search that file for the keywords “check” and “debug”.

Profiling Performance

Speed

To profile your code for speed, build with profiling info using --profiling, (which is currently the same as -g2), and then run the code in the browser’s devtools profiler. You should then be able to see in which functions most of the time is spent.

Memory

The browser’s memory profiling tools generally only understand allocations at the JavaScript level. From that perspective, the entire linear memory that the emscripten-compiled application uses is a single big allocation (of a WebAssembly.Memory). To get information about usage inside that object, you need other tools:

• Emscripten supports the mallinfo(), API, which gives you information from dlmalloc about current allocations.

• Emscripten also has a --memoryprofiler option that displays memory usage in a visual manner. Note that you need to emit HTML (e.g. with a command like emcc test/hello_world.c --memoryprofiler -o page.html) as the memory profiler output is rendered onto the page. To view it, load page.html in your browser (remember to use a local webserver). The display auto-updates, so you can open the devtools console and run a command like _malloc(1024 * 1024). That will allocate 1MB of memory, which will then show up on the memory profiler display.

Manual print debugging

You can also manually instrument the source code with printf() statements, then compile and run the code to investigate issues. The output from the stdout and stderr streams is copied to the browser console by default. Note that printf() is line-buffered, so make sure to add \n to see output in the console. The functions in the console.h header can also be used to access the console more directly.

Emscripten-Specific Issues

Memory Alignment Issues

The Emscripten memory representation is compatible with C and C++. In WebAssembly, unaligned loads and stores will work; each may be annotated with its expected alignment. However if the actual alignment does not match, it may be very slow on some systems.

Tip

SAFE_HEAP can be used to reveal memory alignment issues.

Generally it is best to avoid unaligned reads and writes. Often they occur as the result of undefined behavior. In some cases, however, they are unavoidable — for example if the code to be ported reads an int from a packed structure in some pre-existing data format. In that case, to make it as fast as possible in WebAssembly, you can make sure that the compiler knows the load or store is unaligned. To do so you can:

• Manually read individual bytes and reconstruct the full value

• Use the emscripten_align* typedefs, which define unaligned versions of the basic types (short, int, float, double). All operations on those types are not fully aligned (use the 1 variants in most cases, which mean no alignment whatsoever).

Function Pointer Issues

If you get an abort() from a function pointer call to nullFunc or b0 or b1 (possibly with an error message saying “incorrect function pointer”), the problem is that the function pointer was not found in the expected function pointer table when called.

Note

nullFunc is the function used to populate empty index entries in the function pointer tables (b0 and b1 are shorter names used for nullFunc in more optimized builds). A function pointer to an invalid index will call this function, which simply calls abort().

There are several possible causes:

• Your code is calling a function pointer that has been cast from another type (this is undefined behavior but it does happen in real-world code). In optimized Emscripten output, each function pointer type is stored in a separate table based on its original signature, so you must call a function pointer with that same signature to get the right behavior (see Function Pointer Issues in the code portability section for more information).

• Your code is calling a method on a NULL pointer or dereferencing 0. This sort of bug can be caused by any sort of coding error, but manifests as a function pointer error because the function can’t be found in the expected table at runtime.

To debug these sorts of issues:

• Compile with -Werror (or otherwise fix warnings, many of which highlight undefined behavior).

• Use -sASSERTIONS=2 to get some useful information about the function pointer being called, and its type.

• Look at the browser stack trace to see where the error occurs and which function should have been called.

• Enable clang warnings on dangerous function pointer casts using -Wcast-function-type.

• Build with SAFE_HEAP=1.

• Debugging with Sanitizers can help here, in particular UBSan.

Infinite loops

Infinite loops cause your page to hang. After a period the browser will notify the user that the page is stuck and offer to halt or close it. If your code hits an infinite loop, one easy way to find the problem code is to use a JavaScript profiler. In the Firefox profiler, if the code enters an infinite loop you will see a block of code doing the same thing repeatedly near the end of the profile.

Note

The Browser main loop may need to be re-coded if your application uses an infinite main loop.

Debugging Emscripten

Debugging the compiler driver

Compiling with the emcc -v will cause emcc to output the sub-commands that it runs as well as passes -v to Clang. The EMCC_DEBUG environment variable can be set to emit even more debug output and generate intermediate files for the compiler’s various stages.

AutoDebugger

The AutoDebugger is the ‘nuclear option’ for debugging Emscripten code. It will rewrite the output so it prints out each store to memory. This is useful for comparing the output for different compiler settings in order to detect regressions. To run the AutoDebugger, compile with the environment variable EMCC_AUTODEBUG=1 set.

Warning

This option is primarily intended for Emscripten core developers.

The AutoDebugger will rewrite the output so it prints out each store to memory. This is useful because you can compare the output for different compiler settings in order to detect regressions.

The AutoDebugger can potentially find any problem in the generated code, so it is strictly more powerful than the CHECK_* settings and SAFE_HEAP. One use of the AutoDebugger is to quickly emit lots of logging output, which can then be reviewed for odd behavior. The AutoDebugger is also particularly useful for debugging regressions.

The AutoDebugger has some limitations:

• It generates a lot of output. Using diff can be very helpful for identifying changes.

• It prints out simple numerical values rather than pointer addresses (because pointer addresses change between runs, and hence can’t be compared). This is a limitation because sometimes inspection of addresses can show errors where the pointer address is 0 or impossibly large. It is possible to modify the tool to print out addresses as integers in tools/autodebugger.py.

To run the AutoDebugger, compile with the environment variable EMCC_AUTODEBUG=1 set. For example:

# Linux or macOS
EMCC_AUTODEBUG=1 emcc test/hello_world.cpp -o hello.html
# Windows
set EMCC_AUTODEBUG=1
emcc test/hello_world.cpp -o hello.html
set EMCC_AUTODEBUG=0

AutoDebugger Regression Workflow

Use the following workflow to find regressions with the AutoDebugger:

• Compile the working code with EMCC_AUTODEBUG=1 set in the environment.

• Compile the code using EMCC_AUTODEBUG=1 in the environment again, but this time with the settings that cause the regression. Following this step we have one build before the regression and one after.

• Run both versions of the compiled code and save their output.

• Compare the output using a diff tool.

Any difference between the outputs is likely to be caused by the bug.

Note

You may want to use -sDETERMINISTIC which will ensure that timing and other issues don’t cause false positives.

Useful Links

• Links to Wasm debugging-related documents

Need help?

The Emscripten Test Suite contains good examples of almost all functionality offered by Emscripten. If you have a problem, it is a good idea to search the suite to determine whether test code with similar behavior is able to run.

If you’ve tried the ideas here and you need more help, please Get in touch.