Modernize and fix the Emscripten build by RobinTF · Pull Request #2849 · ad-freiburg/qlever

RobinTF · 2026-04-29T21:13:24Z

So far, the Emscripten (WASM) build of QLever used an outdated toolchain and had several problems at runtime (for example, exceptions did not work correctly, and ICU could not find its data). This PR modernizes the build and fixes these problems. In particular:

The emsdk is bumped from version 3.1.73 to 6.0.2 and is now installed directly instead of via the Conan Center recipe, which lags behind. The CMake toolchain and the build environment are wired up so that dependencies that build via autotools (ICU and OpenSSL) also use the Emscripten compiler.
Exception handling now consistently uses -fwasm-exceptions for both compiling and linking; previously, exceptions did not work correctly.
The ICU data is now embedded statically, which fixes a U_FILE_ACCESS_ERROR at runtime.
The new class GenericCharTraits provides char_traits for non-standard single-byte character types, which newer versions of libc++ no longer provide implicitly. This completes the char_traits part of Change code to be more compatible with libc++ #3001.
Tests that trigger bugs in the Emscripten toolchain (for example, HttpTest) are excluded from the WASM build; each exclusion has a comment that names the failing compiler component.
Several redundant compilation flags are removed, and the CI workflow is adjusted accordingly.

codecov · 2026-04-29T22:54:49Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.07%. Comparing base (b3a3e4a) to head (b9e4468).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2849      +/-   ##
==========================================
+ Coverage   92.05%   92.07%   +0.01%     
==========================================
  Files         529      530       +1     
  Lines       45362    45395      +33     
  Branches     6208     6208              
==========================================
+ Hits        41757    41796      +39     
+ Misses       1759     1756       -3     
+ Partials     1846     1843       -3

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

sonarqubecloud · 2026-04-30T12:31:07Z

Quality Gate failed

Failed conditions
14.5% Duplication on New Code (required ≤ 3%)

See analysis details on SonarQube Cloud

This is a subset of #2849 that increases the compatibility with `libc++` (the standard library of LLVM). In particular: * In `libc++`, `std::string_view::begin()` and `end()` do not return a `const char*` but a wrapped iterator type, so `.data()` has to be used wherever a raw `const char*` is required (for example, for `std::from_chars`). * `libc++` provides no `char_traits` for non-`char` types, so `std::basic_string_view<std::byte>` is replaced by a `span<std::byte>`. NOTE: This PR does not yet address all instances of this second issue; for the remaining ones there are several possible solutions (some uglier than others), which will be handled in a follow-up. NOTE: This change does not alter any behavior; it only replaces non-portable uses of the standard library by portable ones.

joka921

In general nice, but i have some questions for clarification etc.

joka921

Thank you very much.

sparql-conformance · 2026-07-09T03:54:46Z

Overview

Number of Tests	Passed ✅	Intended ✅	Failed ❌	Not tested
498	414	73	11	0

Conformance check passed ✅

No test result changes.

Details: https://qlever.dev/sparql-conformance-ui?cur=40a0e7df96fb9d0e0c47baba8cf078909c30d28f&prev=f5a34398bdb016c24ba3c01fb93507cd3a670aac

sonarqubecloud · 2026-07-09T05:36:59Z

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

So far, the Emscripten (WASM) build of QLever used an outdated toolchain and had several problems at runtime (for example, exceptions did not work correctly, and `ICU` could not find its data). This PR modernizes the build and fixes these problems. In particular: 1. The `emsdk` is bumped from version 3.1.73 to 6.0.2 and is now installed directly instead of via the Conan Center recipe, which lags behind. The `CMake` toolchain and the build environment are wired up so that dependencies that build via autotools (`ICU` and `OpenSSL`) also use the Emscripten compiler. 2. Exception handling now consistently uses `-fwasm-exceptions` for both compiling and linking; previously, exceptions did not work correctly. 3. The `ICU` data is now embedded statically, which fixes a `U_FILE_ACCESS_ERROR` at runtime. 4. The new class `GenericCharTraits` provides `char_traits` for non-standard single-byte character types, which newer versions of `libc++` no longer provide implicitly. This completes the `char_traits` part of ad-freiburg#3001. 5. Tests that trigger bugs in the Emscripten toolchain (for example, `HttpTest`) are excluded from the WASM build; each exclusion has a comment that names the failing compiler component. 6. Several redundant compilation flags are removed, and the CI workflow is adjusted accordingly.

This is a subset of ad-freiburg#2849 that increases the compatibility with `libc++` (the standard library of LLVM). In particular: * In `libc++`, `std::string_view::begin()` and `end()` do not return a `const char*` but a wrapped iterator type, so `.data()` has to be used wherever a raw `const char*` is required (for example, for `std::from_chars`). * `libc++` provides no `char_traits` for non-`char` types, so `std::basic_string_view<std::byte>` is replaced by a `span<std::byte>`. NOTE: This PR does not yet address all instances of this second issue; for the remaining ones there are several possible solutions (some uglier than others), which will be handled in a follow-up. NOTE: This change does not alter any behavior; it only replaces non-portable uses of the standard library by portable ones.

So far, the Emscripten (WASM) build of QLever used an outdated toolchain and had several problems at runtime (for example, exceptions did not work correctly, and `ICU` could not find its data). This PR modernizes the build and fixes these problems. In particular: 1. The `emsdk` is bumped from version 3.1.73 to 6.0.2 and is now installed directly instead of via the Conan Center recipe, which lags behind. The `CMake` toolchain and the build environment are wired up so that dependencies that build via autotools (`ICU` and `OpenSSL`) also use the Emscripten compiler. 2. Exception handling now consistently uses `-fwasm-exceptions` for both compiling and linking; previously, exceptions did not work correctly. 3. The `ICU` data is now embedded statically, which fixes a `U_FILE_ACCESS_ERROR` at runtime. 4. The new class `GenericCharTraits` provides `char_traits` for non-standard single-byte character types, which newer versions of `libc++` no longer provide implicitly. This completes the `char_traits` part of ad-freiburg#3001. 5. Tests that trigger bugs in the Emscripten toolchain (for example, `HttpTest`) are excluded from the WASM build; each exclusion has a comment that names the failing compiler component. 6. Several redundant compilation flags are removed, and the CI workflow is adjusted accordingly.

This is a subset of ad-freiburg#2849 that increases the compatibility with `libc++` (the standard library of LLVM). In particular: * In `libc++`, `std::string_view::begin()` and `end()` do not return a `const char*` but a wrapped iterator type, so `.data()` has to be used wherever a raw `const char*` is required (for example, for `std::from_chars`). * `libc++` provides no `char_traits` for non-`char` types, so `std::basic_string_view<std::byte>` is replaced by a `span<std::byte>`. NOTE: This PR does not yet address all instances of this second issue; for the remaining ones there are several possible solutions (some uglier than others), which will be handled in a follow-up. NOTE: This change does not alter any behavior; it only replaces non-portable uses of the standard library by portable ones.

So far, the Emscripten (WASM) build of QLever used an outdated toolchain and had several problems at runtime (for example, exceptions did not work correctly, and `ICU` could not find its data). This PR modernizes the build and fixes these problems. In particular: 1. The `emsdk` is bumped from version 3.1.73 to 6.0.2 and is now installed directly instead of via the Conan Center recipe, which lags behind. The `CMake` toolchain and the build environment are wired up so that dependencies that build via autotools (`ICU` and `OpenSSL`) also use the Emscripten compiler. 2. Exception handling now consistently uses `-fwasm-exceptions` for both compiling and linking; previously, exceptions did not work correctly. 3. The `ICU` data is now embedded statically, which fixes a `U_FILE_ACCESS_ERROR` at runtime. 4. The new class `GenericCharTraits` provides `char_traits` for non-standard single-byte character types, which newer versions of `libc++` no longer provide implicitly. This completes the `char_traits` part of ad-freiburg#3001. 5. Tests that trigger bugs in the Emscripten toolchain (for example, `HttpTest`) are excluded from the WASM build; each exclusion has a comment that names the failing compiler component. 6. Several redundant compilation flags are removed, and the CI workflow is adjusted accordingly.

RobinTF added 3 commits April 29, 2026 23:11

Update GH CI build

fcb19af

Next attempt

1d4448d

Fix compilation error

f654ef1

RobinTF added 9 commits April 30, 2026 01:13

Fix next compilation error

6856a90

Next compilation error fix

b233de9

Fix next compiler error

4b98246

Next error

7e4ed82

Next attempt

2076315

Next 2 issues

155f3ae

Fix more errors

f5d6c32

And another one

9c043f9

And another one

79b74e9

hannahbast force-pushed the master branch from 48f8e0f to 676c694 Compare June 13, 2026 22:27

Merge branch 'master' into improve-wasm

6c5ea4c

RobinTF mentioned this pull request Jun 23, 2026

Change code to be more compatible with libc++ #3001

Merged

RobinTF added 11 commits July 6, 2026 18:08

Merge branch 'master' into improve-wasm

1ca26bb

Merge branch 'master' into improve-wasm

a70a10b

Unify and simplfy char_traits

04fc270

Try a better WASM config

5dcc49f

Add separator comments

255f127

Fix include

6bb6ab3

Exclude HttpTest from WASM build

8e92ca9

Merge remote-tracking branch 'ad-freiburg/master' into improve-wasm

b9e4ca5

Remove redundant c flags

6b7163d

Drop redundant -m64 flags

1f621bf

Merge remote-tracking branch 'ad-freiburg/master' into improve-wasm

c13e566

Fix more unit tests

a8656d8

RobinTF changed the title ~~Enhance WASM build pipeline~~ Modernize and fix the WASM (Emscripten) build Jul 7, 2026

RobinTF changed the title ~~Modernize and fix the WASM (Emscripten) build~~ Modernize and fix the Emscripten build Jul 7, 2026

RobinTF marked this pull request as ready for review July 7, 2026 23:51

RobinTF requested a review from joka921 July 7, 2026 23:51

joka921 requested changes Jul 8, 2026

View reviewed changes

Address PR comments

b9e4468

joka921 approved these changes Jul 8, 2026

View reviewed changes

hannahbast added 2 commits July 9, 2026 05:03

Merge remote-tracking branch 'origin/master' into improve-wasm

9e34ac6

Fix indentation

40a0e7d

hannahbast merged commit e850615 into ad-freiburg:master Jul 9, 2026
21 checks passed

RobinTF deleted the improve-wasm branch July 9, 2026 06:34

Navigation Menu

Uh oh!

Modernize and fix the Emscripten build - #2849

Modernize and fix the Emscripten build#2849
hannahbast merged 28 commits into
ad-freiburg:masterfrom
RobinTF:improve-wasm

RobinTF commented Apr 29, 2026 •

edited by hannahbast

Loading

Uh oh!

codecov Bot commented Apr 29, 2026 •

edited

Loading

Uh oh!

sonarqubecloud Bot commented Apr 30, 2026

Uh oh!

joka921 left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

joka921 left a comment

Uh oh!

Uh oh!

sparql-conformance Bot commented Jul 9, 2026

Uh oh!

sonarqubecloud Bot commented Jul 9, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

RobinTF commented Apr 29, 2026 • edited by hannahbast Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov Bot commented Apr 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

sonarqubecloud Bot commented Apr 30, 2026

Quality Gate failed

Uh oh!

joka921 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

joka921 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

sparql-conformance Bot commented Jul 9, 2026

Overview

Conformance check passed ✅

Uh oh!

sonarqubecloud Bot commented Jul 9, 2026

Quality Gate passed

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

RobinTF commented Apr 29, 2026 •

edited by hannahbast

Loading

codecov Bot commented Apr 29, 2026 •

edited

Loading