gh-156780: Emscripten: add missing EM_JS_DEPS

[clementperon]

Split out of #156781, as suggested by @hoodmane.

EM_JS bodies are emitted verbatim, so Emscripten's dependency tracker cannot see the JS symbols they reference and silently omits them.

Declared here: $FS, $PATH, $FS_getMode, $resolveGlobalSymbol and emscripten_exit_with_live_runtime in emscripten_syscalls.c, and $wasmTable, $wasmMemory, $addFunction, $addOnPreRun in emscripten_trampoline.c.

No change for MAIN_MODULE builds. Without it, the link now fails with

error: undefined symbol: $resolveGlobalSymbol

instead of a ReferenceError during initRuntime.

• Issue: Emscripten: static (non-MAIN_MODULE) builds fail at startup on dynamic-linking-only JS helpers #156780

[clementperon]

@hoodmane first PR that add the EM_JS_DEPS() everywhere

[freakboy3742]

Noting that CI is currently failing because of a duplicated symbol; that seems to suggest there's a need to collate these declarations into a single coherent location so that we can avoid duplication, rather than co-locating them with usage.

Also - can I ask that you please don't force push. We don't care about a "messy" commit history; CPython uses merge commits, so any messy history isn't preserved after merge. Force pushes also mean that we can lose the context for review comments, and it's really easy to force push a merge comment that pings the entire core team for comment.

[hoodmane]

Emscripten's normal style guide is to put these next to the symbol that uses them so I think that's alright. I think we have a conflict with #156335, need to merge a main that includes that commit and fix the duplicate symbol.

[clementperon]

@freakboy3742 "Noting that CI is currently failing because of a duplicated symbol" ?

The build fail no?
Something like

dup.c >>>
EM_JS_DEPS(tag_a, "$FS");

EM_JS_DEPS(tag_a, "$PATH");  

dup.c:3:1: error: redefinition of '__em_lib_deps_tag_a'
dup.c:1:1: note: previous definition is here

The issue here is that my fork was not up to date with main.

Also about the force push well noted, I had to do it here to properly rebase on latest master

[freakboy3742]

Ok - if Emscripten convention is to keep the declarations locally, then it makes sense for us to do the same.

[bedevere-bot]

⚠️⚠️⚠️ Buildbot failure ⚠️⚠️⚠️

Hi! The buildbot x86 Debian Non-Debug with X 3.x (tierless) has failed when building commit 0b3b154.

What do you need to do:

1. Don't panic.
2. Check the buildbot page in the devguide if you don't know what the buildbots are or how they work.
3. Go to the page of the buildbot that failed (https://buildbot.python.org/#/builders/1245/builds/9398) and take a look at the build logs.
4. Check if the failure is related to this commit (0b3b154) or if it is a false positive.
5. If the failure is related to this commit, please, reflect that on the issue and make a new Pull Request with a fix.

You can take a look at the buildbot page here:

https://buildbot.python.org/#/builders/1245/builds/9398

Summary of the results of the build (if available):

==

Click to see traceback logs

Note: switching to '0b3b154060936e1b120f5b9793ed3f8647df8480'.

You are in 'detached HEAD' state. You can look around, make experimental
changes and commit them, and you can discard any commits you make in this
state without impacting any branches by switching back to a branch.

If you want to create a new branch to retain commits you create, you may
do so (now or later) by using -c with the switch command. Example:

  git switch -c <new-branch-name>

Or undo this operation with:

  git switch -

Turn off this advice by setting config variable advice.detachedHead to false

HEAD is now at 0b3b15406093 gh-156780: Emscripten: add missing EM_JS_DEPS (#156798)
Switched to and reset branch 'main'

configure: WARNING: i686-pc-linux-gnu/gcc is not supported
configure: WARNING:

Platform "i686-pc-linux-gnu" with compiler "gcc" is not supported by the
CPython core team, see https://peps.python.org/pep-0011/ for more information.


Objects/moduleobject.c: In function ‘module_from_slots_and_spec’:
Objects/moduleobject.c:412:38: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
  412 |     uint64_t multiple_interpreters = (uint64_t)Py_MOD_MULTIPLE_INTERPRETERS_SUPPORTED;
      |                                      ^
Objects/moduleobject.c:474:44: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
  474 |                     requires_gil = (val != (uint64_t)Py_MOD_GIL_NOT_USED);
      |                                            ^
In file included from Objects/mimalloc/static.c:26,
                 from Objects/obmalloc.c:26:
Objects/mimalloc/arena.c:59:1: note: the alignment of ‘_Atomic long long int’ fields changed in GCC 11.1
   59 | } mi_arena_t;
      | ^
Python/ceval.c: In function ‘_Py_BuildMap_StackRefSteal’:
Python/ceval.c:943:11: warning: ‘args_o_temp’ may be used uninitialized [-Wmaybe-uninitialized]
  943 |     res = _PyDict_FromItems(
      |           ^~~~~~~~~~~~~~~~~~
  944 |         args_o, 2,
      |         ~~~~~~~~~~
  945 |         args_o+1, 2,
      |         ~~~~~~~~~~~~
  946 |         half_args
      |         ~~~~~~~~~
  947 |     );
      |     ~      
In file included from Python/ceval.h:11,
                 from Python/ceval.c:3:
./Include/internal/pycore_dict.h:333:23: note: by argument 3 of type ‘PyObject * const*’ {aka ‘struct _object * const*’} to ‘_PyDict_FromItems’ declared here
  333 | PyAPI_FUNC(PyObject *)_PyDict_FromItems(
      |                       ^~~~~~~~~~~~~~~~~
In file included from Python/ceval.c:702:
Python/ceval.c:938:52: note: ‘args_o_temp’ declared here
  938 |     STACKREFS_TO_PYOBJECTS(arguments, half_args*2, args_o);
      |                                                    ^~~~~~
Python/ceval_macros.h:522:15: note: in definition of macro ‘STACKREFS_TO_PYOBJECTS’
  522 |     PyObject *NAME##_temp[MAX_STACKREF_SCRATCH+1]; \
      |               ^~~~
./Modules/_remote_debugging/binary_io_reader.c: In function ‘binary_reader_replay’:
./Modules/_remote_debugging/binary_io_reader.c:1250:55: warning: comparison of integer expressions of different signedness: ‘int’ and ‘uint32_t’ {aka ‘unsigned int’} [-Wsign-compare]
 1250 |             if (replayed % PROGRESS_CALLBACK_INTERVAL < count) {
      |                                                       ^
./Modules/_sqlite/connection.c: In function ‘deserialize_impl’:
./Modules/_sqlite/connection.c:2352:19: warning: comparison is always false due to limited range of data type [-Wtype-limits]
 2352 |     if (data->len > 9223372036854775807) {  // (1 << 63) - 1
      |                   ^
./Modules/_testlimitedcapi/heaptype_relative.c: In function ‘make_sized_heaptypes’:
./Modules/_testlimitedcapi/heaptype_relative.c:71:28: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
   71 |                            (unsigned long long)data_ptr,
      |                            ^

make: *** [Makefile:2481: buildbottest] Error 3