runMetaWithFS always use XHR and fails in nodejs environment

[jeffguorg]

Please include the following in your bug report:

Version of emscripten/emsdk: emcc (Emscripten gcc/clang-like replacement + linker emulating GNU ld) 3.1.64 (a1fe3902bf73a3802eae0357d273d0e37ea79898) clang version 19.0.0git (https:/github.com/llvm/llvm-project 4d8e42ea6a89c73f90941fd1b6e899912e31dd34) Target: wasm32-unknown-emscripten Thread model: posix InstalledDir: /home/[myname]/Workspace/github.com/emscripten-core/emsdk/upstream/bin

Failing command line in full: Full link command and output with -v appended:

I was compiling libreoffice to wasm and emscripten + libreoffice already did a great job. but when i used the generated output, i found that it works in browser but not in nodejs. it says XMLHttpRequest is not defined. i tracked down the whole build process and found that libreoffice is using --separate-metadata and it generated a function called runMetaWithFS

https://github.com/emscripten-core/emscripten/blob/54c570094fa5bf48151e4a79d78a50a2b3c612c2/tools/file_packager.py#L1113-L1118

i temporarily work around it by modify locally installed emscripten like this:

  1113   function runMetaWithFS() {
  1114     Module['addRunDependency']('%(metadata_file)s');
  1115     var REMOTE_METADATA_NAME = Module['locateFile'] ? Module['locateFile']('%(metadata_file)s', '') : '%(metadata_file)s';
  1116     if (                
  1117       typeof process === "object" &&
  1118       typeof process.versions === "object" &&
  1119       typeof process.versions.node === "string"
  1120     ) {                 
  1121       require("fs").readFile(REMOTE_METADATA_NAME, function (err, contents) {
  1122         if (err) {      
  1123           console.log(err)
  1124         } else {        
  1125           loadPackage(JSON.parse(contents.toString()))
  1126         }               
  1127       });               
  1128       return;           
  1129     }  

[sbc100]

Are you set in using --preload-file for some reason over --embed-file? That latter is much more effecient and I would only really recommend the former if you want to be able to separately update your data without updating the binary. (sorry if I've asked this already).