John Skaller

Yep .. missing switch: ``` Dynamic Linking library build/release/trial/lib/rtl/libflx_gc_dynamic.so [system] "g++" "-shared" "build/rtl-tmp/flx_collector_dynamic.o" "build/rtl-tmp/flx_serialisers_dynamic.o" "build/rtl-tmp/flx_judy_scanner_dynamic.o" "build/rtl-tmp/flx_gc_dynamic.o" -o "build/release/trial/lib/rtl/libflx_gc_dynamic.so" "-Lbuild/release/trial/lib/rtl" "-ljudy_dynamic" "-lflx_exceptions_dynamic" ``` Now the question is how did the CI...

But here is a test case as an example: ``` [system] "g++" "-shared" "/home/runner/.felix/cache/text/home/runner/work/felix/felix/build/release/test/tut/tut_12_dynamic.o" -o "/home/runner/.felix/cache/binary/home/runner/work/felix/felix/build/release/test/tut/tut_12.so" "-Lbuild/release/host/lib/rtl" "-lflx_dynamic" "-lflx_pthread_dynamic" "-lpthread" "-lflx_dynlink_dynamic" "-ldl" "-lflx_strutil_dynamic" "-lflx_gc_dynamic" "-ljudy_dynamic" "-lflx_exceptions_dynamic" ``` See? That one...

Can you try this: ``` ~/felix>cat build/release/host/config/pthread.fpc Generated_from: 539 "/Users/skaller/felix/src/packages/rtl-threads.fdoc" Description: pthread support defaults to no requirements ``` That's on MacOS. On Linux is should be this file: ``` Generated_from:...

You should probably build again from scratch. I just don't understand what's happening considering it builds on Ubuntu with g++ version 9 on the GitHub CI server. You could try:...

What do you mean "go into errors"? The repl requires a Felix program. For example ``` ~/felix>flx --repl > println$ 1,2,3; (1, 2, 3) ``` It reloads definitions every time....

GOT IT I think! It's a bug in C++ thread library construction. ``` void flx_collector_t::mark_multi(pthread::memory_ranges_t *px,int reclimit, int nthreads) { //fprintf(stderr, "starting %d mark threads\n", nthreads); j_tmp_waiting = 0; mark_thread_context_t...

Reopening. Stupid github decided a reference to this in a message should close it. Assigning to razetime.

> > Is the idea here to provide this mostly for informational purposes, or would we try to enforce that a procedure behaves according to these declarations? > > It...

Here is the result of the analyser. When it shows n..m it means that's the range of the stack depth at that point. Join occurs when two control flows merge....

I guess the analyser could calculate that. But i'm not sure how useful it would be: miden procedures aren't functions. In fact the don't have coherent semantics because they can...