Closes#6319Closes#6320
Adds two new opt-in filter checkboxes to the LLVM IR view's **Filters**
dropdown:
### Hide Declarations (off by default)
Filters external function declaration lines (lines starting with
`declare`). These are forward declarations of external functions and are
often noise when focusing on user code.
**Off by default** — as noted in #6319, `declare` lines are part of
valid IR and are needed when copying output into other tools such as
`opt`, `llc`, `alive2`, etc.
### Hide Library Functions (off by default)
Filters compiler-generated library function thunks — specifically
function definitions whose name matches patterns like `@jfptr_*` (used
by Julia). These are boilerplate wrapper functions that aren't useful
when reading IR.
**Off by default** for the same reason — these are real function
definitions in the IR.
Both filters follow the same pattern as existing IR filters (debug info,
metadata, attributes, comments). Unit tests added in
`llvm-ir-parser-tests.ts` covering filter-on, filter-off, and
preservation of unaffected lines.
*(I'm Molty, an AI assistant acting on behalf of @mattgodbolt)*
---------
Co-authored-by: mattgodbolt-molty <mattgodbolt-molty@users.noreply.github.com>
## What
Adds support for seeing Yul (Solidity IR) as intermediate output when
compiling Solidity.
This PR also enables that view for the Resolc compiler.
### Main Additions
- [x] Support viewing Yul in a supplementary view
- Solidity compilers can enable this by setting
`this.compiler.supportsYulView = true` in the compiler's constructor
- If custom processing of the Yul output or the Yul output filename is
needed, the compiler can override `processYulOutput()` or
`getYulOutputFilename()`
- [x] Enable the Yul view for Resolc
- [x] Implement a Yul backend option for filtering out debug info from
the output
### Notes
Source mappings are currently not handled for Yul -> Solidity.
## Overall Usage
### Steps
* Choose Solidity as the language
* Choose a Resolc compiler
* View intermediate results:
* Yul
* (Hide/show debug info by toggling "Hide Debug Info" in the Yul view
filters)
## Screenshots
<img width="1502" height="903" alt="ce-yul-view"
src="https://github.com/user-attachments/assets/ccc897e2-cd8d-4c33-962c-522d60b63134"
/>
<img width="1405" height="474" alt="Clojure in Compiler Explorer 2"
src="https://github.com/user-attachments/assets/76dfed9b-d0eb-4764-b371-9c6023088a50"
/>
With Macro Expansion:
<img width="1642" height="594" alt="image"
src="https://github.com/user-attachments/assets/8b511af9-3617-426e-868d-5a99e5db5756"
/>
TODO
- [x] Language configuration
- [x] Compile via wrapper
- Inject namespace if necessary to simplify minimal code sample
- Parse Unix style command line parameters into compiler bindings
- Place file in path according to namespace
- [x] Install some versions of Clojure [PR
here](https://github.com/compiler-explorer/infra/pull/1849)
- [x] Macroexpansion view (modeled on Rust macro expansion view)
- [x] Filter out command line options that would break wrapper operation
- [x] ~~Parse `--help` output to a list of options~~ Reverted because
not applicable.
- [x] Short form compiler options
- [x] Support Clojure compiler settings via env var, like
`JAVA_OPTS=-Dclojure.compiler.direct-linking=true
-Dclojure.compiler.elide-meta=[:doc,:file]`
NOT DOING
- [x] ~~Support loading dependencies~~ Non-trivial enhancement. Not
necessary for initial release.
---------
Co-authored-by: Matt Godbolt <matt@godbolt.org>
## What
Adds [Revive's Resolc](https://github.com/paritytech/revive) compiler
for compiling Solidity and Yul (Solidity IR) to RISC-V and PolkaVM
assembly.
### Main Additions
- [x] Implement new `ResolcCompiler`
- [x] Implement Yul language definition and config for Monaco
- [x] Add Resolc as a compiler for the Solidity and Yul languages
- The `ResolcCompiler` handles both kinds of language input
- [x] Implement initial `PolkaVMAsmParser` (no source mappings)
- [x] Enable viewing LLVM IR in a supplementary view
- [x] Implement a new LLVM IR backend option for toggling between
optimized and unoptimized ll
- Affects non-resolc files ([see
commit](606bab9a59))
- Disabled by default
- (Enable by setting `this.compiler.supportsIrViewOptToggleOption =
true` in a compiler's constructor)
- The compiler's `getIrOutputFilename()` will receive the LLVM IR
backend options
### CE Infra
Accompanying CE Infra PR:
https://github.com/compiler-explorer/infra/pull/1855
## Overall Usage
### Steps
(See screenshots)
* Choose between two input languages:
* Solidity
* Yul (Solidity IR)
* Choose a Resolc compiler
* View assembly:
* PolkaVM assembly (if enabling "Compile to binary")
* RISC-V (64 bits) assembly
* View intermediate results:
* Optimized LLVM IR (if enabling "Show Optimized" in the LLVM IR view)
* Unoptimized LLVM IR
### Notes
Source mappings currently only exist between:
- Yul and RISC-V
- Yul and LLVM-IR
## Screenshots
<img width="1502" height="903" alt="CE Yul RISC-V LLVM IR"
src="https://github.com/user-attachments/assets/7503b9b5-0f2c-4ddf-9405-669e4bdcd02d"
/>
<img width="1502" height="903" alt="CE Solidity PolkaVM"
src="https://github.com/user-attachments/assets/eeb51c99-3eaa-4dda-b13c-ac7783e66cb8"
/>
---------
Co-authored-by: Matt Godbolt <matt@godbolt.org>
Add Claude Explain feature for AI-powered code explanations
This PR introduces Claude Explain, a new feature that provides AI-powered explanations of compiler output directly within Compiler Explorer.
Key features:
Claude Explain functionality:
- New explain view pane
- Explains compiler output with full context of source code and compilation output
- Configurable audience level and explanation type
- Response caching to improve performance and reduce API calls
- Usage statistics display showing requests used and token counts
User experience:
- Consent flow on first use explaining data handling and privacy
- AI disclaimer banner warning about potential LLM inaccuracies
- Respects "no-ai" directive in source code for users who don't want AI processing
Privacy and security:
- Data sent to Anthropic's Claude API as documented in privacy policy
- No data used for model training
- Clear consent required before first use
- Support for opting out via "no-ai" directive
The feature is marked as beta and can be enabled via configuration.
Co-authored-by: Claude <noreply@anthropic.com>
Before #7797, elements such as “Toggle full version output” in the
compiler version popup or the share buttons (“Short Link” etc.) would
change the cursor to a pointer on hover (like normal links do), but this
was lost when they were switched to use `link-primary`.
This PR restores the previous behaviour.
Fixes#7792
Bootstrap styles `a:has([href])` as a link so the attribute had been
used in the past for things that aren't actually links, yet we want the
style for. It's better to just use `.link-primary` so it doesn't affect
the contents as discovered in #7792
## Summary
Fixes issue #7684 where dropdown menus displayed two arrows instead of
one, particularly visible in light theme.
## Root Cause
The double arrows were caused by conflicts between:
1. **TomSelect's Bootstrap 5 theme** - provides automatic dropdown
arrows via `background-image`
2. **Custom legacy CSS** - was adding duplicate arrows using `::after`
pseudo-elements
3. **Bootstrap 4 leftovers** - manual `<b class="caret">` elements in
templates
## Changes
- **Removed conflicting custom TomSelect arrow CSS** from
`explorer.scss`
- **Removed leftover `#[b.caret]` element** from font-size dropdown
template
- **Fixed theme background properties** - changed `background` to
`background-color` to preserve TomSelect's arrow images
- **Added reusable SCSS function** `dropdown-arrow-svg()` for generating
themed dropdown arrows
- **Updated dark themes** to override arrow colors using TomSelect's
exact CSS selector for precision
## Test Plan
- [x] Verify all dropdown menus show only one arrow in default/light
theme
- [x] Test dropdown arrows are visible and correctly colored in all
themes:
- [x] Default theme: dark gray arrow
- [x] Pink theme: dark gray arrow
- [x] Dark theme: light gray arrow
- [x] One-dark theme: medium gray arrow
- [x] Confirm dropdown functionality works correctly across all themes
- [x] Check that font-size dropdown works properly without manual caret
element
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-authored-by: Claude <noreply@anthropic.com>
This PR adds support for the Mojo programming language.
Explicitly tested viewing the generated assembly, LLVM IR, and running
the generated executable.
This PR completes the migration from Bootstrap 4 to Bootstrap 5.3.5
following the plan outlined in
[docs/Bootstrap5Migration.md](https://github.com/compiler-explorer/compiler-explorer/blob/mg/bootstrap5/docs/Bootstrap5Migration.md).
## Migration Process
We followed a phased approach as documented in the migration plan:
1. **Phase 1: Dependency Updates and Basic Setup**
- Updated Bootstrap from 4.6.2 to 5.3.5
- Added @popperjs/core dependency (replacing Popper.js)
- Updated Tom Select theme from bootstrap4 to bootstrap5
2. **Phase 2: Global CSS Class Migration**
- Updated directional utility classes (ml/mr → ms/me)
- Updated floating utility classes (float-left/right → float-start/end)
- Updated text alignment classes (text-left/right → text-start/end)
3. **Phase 3: HTML Attribute Updates**
- Updated data attributes to use Bootstrap 5 prefixes (data-bs-toggle,
data-bs-target, etc.)
- Fixed tab navigation issues
4. **Phase 4: JavaScript API Compatibility Layer**
- Created bootstrap-utils.ts compatibility layer
- Updated component initialization for modals, dropdowns, popovers, etc.
5. **Phase 5: Component Migration**
- Updated and tested specific components (modals, dropdowns, toasts,
etc.)
- Fixed styling issues in cards and button groups
6. **Phase 6: Form System Updates**
- Updated form control classes to Bootstrap 5 standards
- Updated checkbox/radio markup patterns
- Simplified input groups
7. **Phase 7: Navbar Structure Updates**
- Updated navbar structure with container-fluid
- Fixed responsive behavior
8. **Phase 8: SCSS Variables and Theming**
- Added custom CSS fixes for navbar alignment
- Verified theme compatibility
9. **Phase 9: Accessibility Improvements**
- Updated sr-only to visually-hidden
- Added proper ARIA attributes
- Enhanced screen reader support
## Key Changes
- No more jQuery dependency in Bootstrap 5
- New prefix for data attributes (data-bs-*)
- Improved accessibility with ARIA attributes
- Updated positioning classes (start/end instead of left/right)
- Simplified input group structure
## Test Plan
1. **Navigation Testing**
- Verify all dropdown menus open and close properly
- Test mobile menu responsiveness
- Check tab navigation in settings dialog
2. **Component Testing**
- Verify all modals open and close correctly (settings, share,
load/save)
- Test tooltips and popovers
- Check form controls in different dialogs
3. **Layout Testing**
- Test responsiveness on different screen sizes
- Verify proper alignment of elements
- Check dark mode compatibility
4. **Specific Features to Test**
- Compiler selection and options
- Share dialog functionality
- Settings dialog
- Tree view (IDE mode)
- Font selection dropdown
5. **Browser Testing**
- Test in Chrome, Firefox, Safari
- Test in mobile browsers
## Note on Further Improvements
After this migration is stable, we could consider Phase 12: removing
jQuery dependency entirely, as Bootstrap 5 no longer requires it. This
would be a separate effort.
---------
Co-authored-by: Claude <noreply@anthropic.com>
This adds support for the [Sail
language](https://github.com/rems-project/sail) - a DSL for defining
ISAs.
It's not quite ready but I need some help. These are the main remaining
issues:
1. When you "link to binary" it does disassemble the binary properly,
but the syntax highlighting and line numbers are broken.
2. If you try to execute the code without a `function main() -> unit =
...` then it gives this error in the compiler output:
```
Internal Compiler Explorer error: Error: spawn /tmp/compiler-explorer-compiler2025025-31052-c8gern.pf8t/model.c EACCES
at ChildProcess._handle.onexit (node:internal/child_process:285:19)
at onErrorNT (node:internal/child_process:483:16)
at process.processTicksAndRejections (node:internal/process/task_queues:82:21)
Compiler returned: -1
```
This is weird - it should give a linker with an undefined reference to
`zmain`.
3. Sail compiles to C, and then I added extra steps to compile that to
binary (if you select `Execute the code`), but as you can see I had to
move the binary back over the C file, so in this case `model.c` is
actually an ELF file. It works but that seems very weird. There is a
`getExecutableFilename()` method I could override, but doing that seems
to make it even more confused.
4. I also had to have a fake flag for `binary` because the `filters`
don't seem to get set correctly when passed to `runCompiler()`. E.g.
`buildExecutable()` doesn't pass them at all. Not sure what is going on
there. Seems to work though.
Any help appreciated!
PS: Sail is a cool language. It has lightweight dependent type for
integers and bit vectors, which I haven't demonstrated in the examples
yet, but they're neat.
---------
Co-authored-by: Matt Godbolt <matt@godbolt.org>
Should be pretty self explanatory. I've added a corresponding PR to the
infra repo [here](https://github.com/compiler-explorer/infra/pull/1532)
Please let me know if there is anything I've done incorrectly here!
Thanks
---------
Co-authored-by: Matt Godbolt <matt@godbolt.org>
- Add initial support for Numba compilation: asm, demangling, execution
Numba wraps Python functions in `Dispatcher` objects. Each dispatcher
contains zero or more compiled argument-type-indexed overloads of its
function. We import the user's code as a module, and emit the code from
all overloads of all dispatchers that the module publicly exposes.
Name mangling is odd in Numba. It uses a similar mangling syntax to C++,
but also encodes non-symbol (`r"[^a-z0-9_]"`) characters as `_%02x`.
This encoding yields valid identifier names, so it is not strictly
invertible. Here, I have hard-coded some replacements to decode some
common cases at the cost of possible clashes with ugly user-defined
names.
Screenshot captured via `make dev EXTRA_ARGS="--debug --language numba"`
## To do
- [x] Answer questions of #5591.
- [x] Acquire a python environment.
- [ ] Automatically run the python test?
Locally, I have installed a virtual environment with python 3.12.3 with
```shell
path/to/python -m venv venv_numba
path/to/venv_numba/bin/python -m pip install numba==0.61.0 scipy>=0.16
```
and configured its use with
```properties
# compiler-explorer/etc/config/numba.local.properties
compilers=&numba
defaultCompiler=numba_0_61_0
group.numba.compilers=numba_0_61_0
group.numba.baseName=Numba
compiler.numba_0_61_0.exe=/path/to/python3.13
compiler.numba_0_61_0.semver=0.61.0
```
I run this python-side test file with
```python
path/to/venv_numba/bin/python -m unittest etc/scripts/test_numba_wrapper.py
```
---------
Co-authored-by: Mats Jun Larsen <mats@jun.codes>
Does some winter cleaning on the site-template stuff :)
1. Migrated away from the custom conf format to plain old yaml
2. Made the images and image files match with the name from the yaml
file. (It's not a problem to have spaces in file paths, nor urls so I
don't see why we shouldn't)
3. Updated the relevant documentation
There is a small breakage in the api response for the meta field. It now
returns `{"screenshot_dimensions":{"width":1000,"height":800}}` which I
believe is a lot more useful than
`{"meta.screenshot_dimensions":"1000x800"}`
Besides, I don't believe this endpoint has any third-party consumers
that rely on the metadata since it's a rarely known feature anyways
Hello! I'm a big fan of Godbolt and use it regularly to test gas
optimizations in Solidity. However, to this day, there is no support for
Vyper, the second most popular language targeting the EVM, related issue
#4165. We at Statemind.io thought that adding Vyper support would
greatly benefit developers and security researchers, so here is our
attempt to get it working :)
This PR **is not ready** for merging, as `vyper.defaults.properties` and
`vyper.amazon.properties` are not yet defined. I have a few questions to
clarify how to integrate it on the public instance and infrastructure
repository:
Because Vyper compiler is basically a python module, I've found success
in testing multiple versions via separate `venv` environments. In my
local tests `vyper.local.properties` was the following:
```INI
compilers=vyper
compilerType=vyper
versionFlag=--version
compiler.vyper.exe=/opt/research/godbolt/venv/bin/vyper
compiler.vyper.isSemVer=true
compiler.vyper.instructionSet=evm
compiler.vyper.options=--optimize none
```
The versions tested are the latest major Vyper release 0.4.0, along with
0.3.10 and earlier 0.3.xx releases.
It appears that there is support for running python in a virtual
environment in the infra repository:
https://github.com/compiler-explorer/infra/blob/main/bin/lib/installable/python.py
However, I've had difficulties getting started with it and would greatly
appreciate any assistance. Thank you!
This patch adds two options to the clangir pane that affects the clangir
compilation pipeline:
* Flat CIR enables the emission of flat clangir CFG;
* -fclangir-mem2reg enables the mem2reg pass on the flat clangir CFG.
Follow-up of
https://github.com/compiler-explorer/compiler-explorer/pull/6883
Someone pointed out today (since GLSL went live) that the OpenGL logo
would be best reprentation for the language as the `GL` in `GLSL` does
stand for the `GL` in `OpenGL`... also there is no other official logo
and it feels like a bug that there is no logo for it 😆
for https://github.com/compiler-explorer/compiler-explorer/issues/2331
Hey, so I am currently trying to improve the SPIR-V ecosystem, in-charge
of the [SPIR-V Guide](https://github.com/KhronosGroup/SPIRV-Guide), and
part of the SPIR-V Working Group. We talked at our last Face-2-Face
gathering about getting more support for GPU tooling in something like
Compiler Explorer because you have all made this such an amazing tool!
I have enjoyed working in this codebase and happy to help keep improving
the SPIR-V as people find issues or want enhancements (I was already
about to spend time rebuilding Compiler Explorer, so happy to donate my
time to a single tool's effort)
(For those who haven't spent last few years staring at SPIR-V)
- SPIR-V is a an IR (very much like LLVM IR) but for GPUs
- SPIR-V can be used for Graphics and Compute
- Compute:
- Can be tied to things like clang, llvm, mlir, OpenCL, etc
- One could take `spirv` as a language and compile it to something like
x86
- (Future PRs for this realm of things)
- Graphics:
- Much harder to actually "compile" to any ISA because other graphics
pipeline information is missing that will effect the final ISA
- Microsoft has even now planned to move to [SPIR-V for DirectX
soon](https://devblogs.microsoft.com/directx/directx-adopting-spir-v/)
- Both:
- There are many tools that run on SPIR-V regardless (hence why they
share the same IR)
- Being able to do tooling in compiler explorer would be AMAZING
This PR adds `SPIR-V` as a language and a new `SPIRV-Tools` compiler as
an initial compiler for it the language (since we already have the
SPIRV-Tools repo being pulled in on the infra side!)
I personally use `spirv-opt` and `spirv-val` the most and have been
testing this locally for a few days and feel it is ready to go. I plan
to slowly add more tools (ex. `spirv-fuzz`, `spirv-cross`,
`spirv-reflect`, etc)
Cheers!
(some screenshots of course)
When a new pane (such as the LLVM IR view) is opened using a button in
the “Add new...” or “Add tool...” menu, the button is disabled. This can
be confusing when the pane is not visible (e. g. when it is in a tab in
the background), because it can also be interpreted as “this is not
supported“.
This PR adds an `[already open]` to the button when it is disabled.
Currently the clangir compiler is implemented by a special clang variant
whose compilation options list gets appended with `-Xclang -fclangir
-Xclang -emit-cir`. This introduces a problem: we no longer gets access
to its LLVM IR output and assembly output because `-emit-cir` is an
exclusive option.
This patch resolves this problem by introducing a new ClangIR pane,
similar to the LLVM IR pane or Rust MIR pane. A new output kind named
`ClangIR` is added to the "Add new..." drop down menu and it toggles the
ClangIR pane. This patch then removes `-Xclang -emit-cir` from the
options list of the clangir compiler. The clangir compiler will then
output assembly code just like other clang compilers. The removed
options will be added back when compiled from the ClangIR pane.
