PS: maybe it makes sense to replace the CRT file functions with empty stubs when NO_FILESYSTEM is defined? That why $FS wouldn't be pulled in even when the code contains such functions...
On Friday, 19 April 2019 20:14:11 UTC+2, Floh wrote: > > Looks like I'm getting linker errors now in the incoming branch for code > which contains calls to fopen() etc, but defines "-s NO_FILESYSTEM=1": > > error: undefined symbol: $FS > warning: To disable errors for undefined symbols use `-s > ERROR_ON_UNDEFINED_SYMBOLS=0` > > I realize that the previous behaviour may have been somewhat accidental > (after all the code contains filesystem calls, even though the code is > never called), but the previous behaviour was quite convenient ;) > > (the specific problem happens with Dear ImGui, this has filesystem > functions for serializing the UI state to disk, this can disabled at > runtime, but not compile time). > > Is this change of NO_FILESYSTEM behaviour going to stay that way, or is it > an unintended regression? > > > > -- You received this message because you are subscribed to the Google Groups "emscripten-discuss" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
