Process aliases by mat-hek · Pull Request #2027 · atomvm/AtomVM

mat-hek · 2025-12-08T15:13:52Z

TODO:

These changes are made under both the "Apache 2.0" and the "GNU Lesser General
Public License 2.1 or later" license terms (dual license).

SPDX-License-Identifier: Apache-2.0 OR LGPL-2.1-or-later

mat-hek · 2025-12-09T10:05:27Z

@bettio it seems that CI sometimes fails because it cannot find erlang:alias. I added it to erlang.erl, nifs.gperf and nifs.c, am I missing something?

bettio · 2025-12-09T11:33:48Z

src/libAtomVM/term.h

 #define BOXED_FUN_SIZE 3
 #define FLOAT_SIZE (sizeof(float_term_t) / sizeof(term) + 1)
 #define REF_SIZE ((int) ((sizeof(uint64_t) / sizeof(term)) + 1))
+#define TERM_BOXED_PID_REF_SIZE (REF_SIZE + 1)


For consistency with the work we are doing lately I suggest renaming all PID_REF and pid_reference to PROCESS_REF and process_reference

bettio · 2025-12-09T11:38:03Z

src/libAtomVM/term.h

+    term *boxed_value = memory_heap_alloc(heap, TERM_BOXED_PID_REF_SIZE);
+    boxed_value[0] = TERM_BOXED_PID_REF_HEADER;
+
+#if TERM_BYTES == 8


nitpicking/very low priority comment: just for consistency I would suggest moving first the 4 branch:
#if TERM_BYTES == 4 ... #elif TERM_BYTES == 8.
Same applies to the other functions as well.

bettio · 2025-12-09T11:43:07Z

src/libAtomVM/term.c

        return ret;

+    } else if (term_is_pid_reference(t)) {
+        int32_t process_id = term_pid_ref_to_process_id(t);


We are using uint32_t for process id.

Are you sure? It's int32 here: https://github.com/atomvm/AtomVM/blob/main/src/libAtomVM/context.h#L106 and in many other places

bettio · 2025-12-09T11:51:55Z

tests/erlang_tests/test_monitor.erl

+    ok = test_alias(),
+    ok = test_monitor_alias(),
+    ok = test_monitor_alias_explicit_unalias(),
+    ok = test_monitor_down_alias(),


We should also test unhappy paths, such as:

double/triple unalias

alias a dead process

unalias a dead process

In addition multiple aliases should be tested, this is legit but is should be properly tested.

bettio · 2025-12-09T11:57:41Z

src/libAtomVM/nifs.c

        RAISE_ERROR(BADARG_ATOM);
    }

+    while (!term_is_nil(options)) {


We should use term_is_nonempty_list to avoid issues with non proper lists.

bettio · 2025-12-09T11:59:35Z

src/libAtomVM/nifs.c

+            RAISE_ERROR(UNSUPPORTED_ATOM);
+        } else {
+            RAISE_ERROR(BADARG_ATOM);
+        }


here we check that last list item is nil.

bettio · 2025-12-09T12:08:07Z

src/libAtomVM/context.h

+enum ContextMonitorAliasType
+{
+    CONTEXT_MONITOR_ALIAS_EXPLICIT_UNALIAS,
+    CONTEXT_MONITOR_ALIAS_DEMONITOR,


Let's use PascalCase. See AVMCCS-N004 rule in our coding style guide.
(There is still some old code that has to be migrated to updated style).

mat-hek · 2025-12-13T10:26:22Z

@bettio added more tests and missing features. The CI is failing, but it seems unrelated to the changes

bettio

Looks mostly good: there is only one change related to IS_NULL_PTR that is required. I suggest also rebasing on top of main, since main has a lot of fixes for flaky tests.

bettio · 2025-12-15T15:29:45Z

src/libAtomVM/opcodesswitch.h

+                            Context *p = globalcontext_get_process_lock(glb, local_process_id);
+                            if (p) {
+                                struct MonitorAlias *alias = context_find_alias(p, ref_ticks);
+                                if (!IS_NULL_PTR(alias)) {


We should not use !IS_NULL_PTR (x) since it is a shorthand for UNLIKELY(x == NULL). So we must use it only when it is unlikely that the pointer is null (that is mostly for error handling purposes).

Done, but there are many occurrences of !IS_NULL_PTR in the code

pguyot · 2025-12-15T17:18:47Z

esp32 tests are crashing

pguyot

My main concern so far is that monitors are now extended to 2 or 3 words instead of 1 or 2. This may break some code that took monitors and passed it around as references, as the term_to_ref_ticks/term_from_ref_ticks is a common pattern. A similar breakage existed with resources being larger references if one tries to pass a resource where a 1/2 words ref is expected, but it is less likely as resources were not references so far. This may be the crash cause of esp32 tests.

port.erl

call(Port, Message, Timeout) ->
    MonitorRef = monitor(port, Port),
    Port ! {'$call', {self(), MonitorRef}, Message},
    Result =
        receive
            {'DOWN', MonitorRef, port, Port, normal} ->
                {error, noproc};
            {'DOWN', MonitorRef, port, Port, Reason} ->
                {error, Reason};
            out_of_memory ->
                out_of_memory;
            {MonitorRef, Ret} ->
                Ret
        after Timeout ->
            {error, timeout}
        end,
    demonitor(MonitorRef, [flush]),
    Result.

uart_driver.c:

static void uart_driver_do_read(Context *ctx, GenMessage gen_message)
{
    GlobalContext *glb = ctx->global;
    struct UARTData *uart_data = ctx->platform_data;
    term pid = gen_message.pid;
    term ref = gen_message.ref;
    uint64_t ref_ticks = term_to_ref_ticks(ref);

pguyot · 2025-12-15T17:20:19Z

libs/estdlib/src/erlang.erl

 -type raise_stacktrace() ::
    [{module(), atom(), arity() | [term()]} | {function(), arity() | [term()]}] | stacktrace().

+-type monitor_option() :: {'alias', 'explicit_unalias' | 'demonitor' | 'reply_demonitor'}.


This is not enforced by style guide but usually we don't quote atoms when it's not required.

pguyot · 2025-12-15T17:32:01Z

src/libAtomVM/context.c

                }
                // Prepare the message on ctx's heap which will be freed afterwards.
-                term ref = term_from_ref_ticks(monitored_monitor->ref_ticks, &ctx->heap);
+                term ref = term_make_process_reference(target->process_id, monitored_monitor->ref_ticks, &ctx->heap);


This is not what OTP does and there probably is a good reason.
Regular monitors are smaller than aliases.

7> Pid = spawn(fun() -> receive ok -> ok end end). <0.94.0> 8> monitor(process, Pid). #Ref<0.3008598808.1200095248.57512> 9> monitor(process, Pid, [{alias, explicit_unalias}]). #Ref<0.0.11651.3008598808.1200160784.57533> 10> make_ref(). #Ref<0.3008598808.1200095248.57553>

pguyot · 2025-12-15T17:34:47Z

src/libAtomVM/term.h

+    boxed_value[0] = TERM_BOXED_PROCESS_REF_HEADER;
+
+#if TERM_BYTES == 4
+    boxed_value[1] = (ref_ticks >> 4);


This was fixed for regular refs, we probably want ref_ticks >> 32

pguyot · 2025-12-15T17:41:32Z

tests/erlang_tests/test_binary_to_term.erl

    end,
    Term.

+is_atomvm_or_otp_version_at_least(OTPVersion) ->


get_otp_version/0 already exists in this module.

Also there is a poor usage (fixed in #2025 )

OTPVersion = get_otp_version(), if OTPVersion =:= atomvm orelse OTPVersion >= 26 -> B == erlang:term_to_binary(A); true -> true end.

be sure to rather do:

OTPVersion = get_otp_version(), if OTPVersion >= 26 -> B == erlang:term_to_binary(A); true -> true end.

as atoms sort higher than integers.

Also there is a poor usage

I would call it 'a readable way' :P

mat-hek · 2025-12-19T11:22:47Z

@pguyot the reason for the STM32 tests failing was that the size of a regular reference is actually 3 terms there, which means process reference takes 4 terms, and it probably conflicts with resource reference size, which is always 4 terms. I changed the process reference size to 5, but that's not a perfect solution :P So we need to come up with another way of distinguishing references - do you have anything in mind?

I also added RefData as a more universal representation of references than ref_ticks. WYDT? Would it make sense for it to support resource references too?

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

pguyot · 2026-01-24T11:25:12Z

src/libAtomVM/jit.c

-            }
+    } else if (term_is_local_pid_or_port(recipient_term)) {
+        int local_process_id;
+        if (term_is_local_pid_or_port(recipient_term)) {


This condition was checked on line 704.

pguyot · 2026-01-24T11:32:18Z

src/libAtomVM/term.h

 #define BOXED_INT64_SIZE (BOXED_TERMS_REQUIRED_FOR_INT64 + 1)
 #define BOXED_FUN_SIZE 3
 #define FLOAT_SIZE (sizeof(float_term_t) / sizeof(term) + 1)
-#define REF_SIZE ((int) ((sizeof(uint64_t) / sizeof(term)) + 1))


A lot of downstream code uses REF_SIZE. If we want to rename it, we shoud keep the previous macro with a deprecation warning.

Looks good.

pguyot · 2026-01-24T11:35:12Z

src/libAtomVM/term.h

+#define TERM_BOXED_REFERENCE_RESOURCE_SIZE 5
+#endif
 #define TERM_BOXED_REFERENCE_RESOURCE_HEADER (((TERM_BOXED_REFERENCE_RESOURCE_SIZE - 1) << 6) | TERM_BOXED_REF)
 #define TERM_BOXED_RESOURCE_SIZE TERM_BOXED_REFERENCE_RESOURCE_SIZE


We can distinguish reference types by size (instead of a common header) and make resources larger than process aliases since aliases should be more commonly created. However, I would welcome static asserts to ensure sizes and headers are indeed different.

Also, we probably want the same prefix TERM_BOX_REFERENCE_*

In this current implementation, we have:

type 32 bits 64 bits

short refs 3 words 2 words

process aliases 4 words 3 words

resources 5 words (2 unused) 4 words (1 unused)

We could also have:

type 32 bits 64 bits

short refs 3 words 2 words

process aliases 4 words 4 words

resources 4 words 4 words

and distinguish process aliases from resources with last word being INVALID_PROCESS_ID for resources.

I'd prefer sticking to the current option for now if it's acceptable, and change it in a separate PR

Looks good to me.

src/libAtomVM/term.h

pguyot · 2026-01-24T12:33:29Z

src/libAtomVM/term.h


+typedef struct RefData RefData;
+
+struct RefData


RefData is an excellent initiative, howerer it should be a way to serialize/deserialize any ref (local or external). And also we would need a TERM_BOXED_REFERENCE_MAX_SIZE to always be able to deserialize it.

pguyot · 2026-01-24T12:40:10Z

src/libAtomVM/context.c

            }
+            case CONTEXT_MONITOR_ALIAS: {
+                free(monitor);
+            }


We may want a break here line what was there on line 800.

pguyot · 2026-01-24T12:41:26Z

src/libAtomVM/nifs.c

    // msg is not in the port's heap
    NativeHandlerResult result = NativeContinue;
-    if (UNLIKELY(memory_ensure_free_opt(ctx, 12, MEMORY_CAN_SHRINK) != MEMORY_GC_OK)) {
+    if (UNLIKELY(memory_ensure_free_opt(ctx, 13, MEMORY_CAN_SHRINK) != MEMORY_GC_OK)) {


12 was a mistake, we want a sum of constants.

pguyot · 2026-01-24T12:42:05Z

src/libAtomVM/nifs.c

+    *is_alias = false;
+    while (term_is_nonempty_list(monitor_opts)) {
+        term option = term_get_list_head(monitor_opts);
+        if (term_is_tuple(option) && term_get_tuple_element(option, 0) == ALIAS_ATOM) {


We need to check the tuple size.

pguyot · 2026-01-24T12:42:37Z

src/libAtomVM/nifs.c

+                default:
+                    RAISE_ERROR(BADARG_ATOM);
+            }
+        } else if (term_is_tuple(option) && term_get_tuple_element(option, 0) == TAG_ATOM) {


We need to check the tuple size.

pguyot · 2026-01-24T12:44:11Z

src/libAtomVM/term.h

+// If you change a reference size, make sure it doesn't
+// conflict with other reference sizes on all architectures.
+#define SHORT_REF_SIZE ((int) ((sizeof(uint64_t) / sizeof(term)) + 1))
+#define TERM_BOXED_PROCESS_REF_SIZE SHORT_REF_SIZE + 1


Macro value should be parenthesized (see AVMCCS-L016)

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

bettio · 2026-01-30T10:58:24Z

Which clang-format version are you using?
clang-format has slightly changed style (for better) in recent versions and now this is formatted in a different way:

-            ref_data = (RefData) {
+            ref_data = (RefData){
                 .type = RefTypeProcess,
                 .process = { .ref_ticks = globalcontext_get_ref_ticks(ctx->global), .process_id = ctx->process_id }
             };
@@ -4389,7 +4389,7 @@
     RefData ref_data;
     struct Monitor *alias_monitor = NULL;
     if (is_alias) {
-        ref_data = (RefData) {
+        ref_data = (RefData){

This causes some friction, but I'm not sure yet if just updating clang-format is the right thing to do.

bettio

This has to be changed:

/home/runner/work/AtomVM/AtomVM/src/libAtomVM/term.h:3148:12: error: ‘ref_data’ may be used uninitialized [-Werror=maybe-uninitialized]
 3148 |     return ref_data;
      |            ^~~~~~~~
/home/runner/work/AtomVM/AtomVM/src/libAtomVM/term.h: In function ‘process_console_mailbox’:
/home/runner/work/AtomVM/AtomVM/src/libAtomVM/term.h:3129:13: note: ‘ref_data’ declared here
 3129 |     RefData ref_data;
      |             ^~~~~~~~
In function ‘term_to_ref_data’,
    inlined from ‘process_console_message’ at /home/runner/work/AtomVM/AtomVM/src/libAtomVM/nifs.c:1163:28,
    inlined from ‘process_console_mailbox’ at /home/runner/work/AtomVM/AtomVM/src/libAtomVM/nifs.c:1229:18:
/home/runner/work/AtomVM/AtomVM/src/libAtomVM/term.h:3148:12: error: ‘ref_data’ may be used uninitialized [-Werror=maybe-uninitialized]
 3148 |     return ref_data;
      |            ^~~~~~~~
/home/runner/work/AtomVM/AtomVM/src/libAtomVM/term.h: In function ‘process_console_mailbox’:
/home/runner/work/AtomVM/AtomVM/src/libAtomVM/term.h:3129:13: note: ‘ref_data’ declared here
 3129 |     RefData ref_data;

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

mat-hek · 2026-02-02T14:50:49Z

@bettio I have clang-format version 18.1.8. I disabled the SpaceAfterCStyleCast rule, let's see if it helps.

bettio · 2026-02-03T14:45:50Z

src/libAtomVM/term.h

+#define TERM_BOXED_REFERENCE_RESOURCE_SIZE 5
+#endif
 #define TERM_BOXED_REFERENCE_RESOURCE_HEADER (((TERM_BOXED_REFERENCE_RESOURCE_SIZE - 1) << 6) | TERM_BOXED_REF)
 #define TERM_BOXED_RESOURCE_SIZE TERM_BOXED_REFERENCE_RESOURCE_SIZE


Looks good to me.

src/libAtomVM/term.h

bettio · 2026-02-03T14:49:47Z

src/libAtomVM/term.h

 #define BOXED_INT64_SIZE (BOXED_TERMS_REQUIRED_FOR_INT64 + 1)
 #define BOXED_FUN_SIZE 3
 #define FLOAT_SIZE (sizeof(float_term_t) / sizeof(term) + 1)
-#define REF_SIZE ((int) ((sizeof(uint64_t) / sizeof(term)) + 1))


Looks good.

bettio · 2026-02-03T15:20:29Z

src/libAtomVM/term.c

+        uint64_t ref_ticks = term_to_ref_ticks(t);
+
+        // Update also REF_AS_CSTRING_LEN when changing this format string
+        return fun->print(fun, "#Ref<%" PRId32 ".%" PRIu32 ".%" PRIu32 ">", process_id, (uint32_t) (ref_ticks >> 32), (uint32_t) ref_ticks);


I think that term.c require more changes than this, namely I think that term_compare logic has to be updated. @pguyot was the last who updated it, so @pguyot let us know if this is a concern after your changes.

Good catch! We should update tests as well to ensure comparison works. I added some tests when I made resources references.

I'll add tests, but monitors are still uniquely identified by ref_ticks, so comparison should work as expected

Ok, so it worked as I expected, but apparently not the way the BEAM does it. Should be fine now.

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

bettio · 2026-02-03T16:43:36Z

src/platforms/esp32/components/avm_builtins/network_driver.c

    uint32_t port_process_id;
    uint32_t owner_process_id;
-    uint64_t ref_ticks;
+    RefData ref_data;


This comment applies to all the other port drivers here around.
In port.erl we have:

call(Port, Message, Timeout) -> MonitorRef = monitor(port, Port),

by default erlang:monitor/2 doesn't make aliases, and all the port calls are going through port:call so actually I don't think this change is really required.
TL;DR: I think this change adds some extra complexity for something it's not supposed to happen.
On the other side, supporting aliases in port drivers would require changing every 3rd party port driver as well (with no benefits in terms of performances)

@mat-hek let me know if there is anything wrong in my reasoning.

You're right ;) These changes were needed because I changed all monitor refs to be process refs. Now I changed that back. The reason I kept RefData in drivers is that I assumed that we could still get a process ref there. If that's not the case, I'll revert it.

From my perspective, though, it's hard to figure out what kind of reference we can expect where, and keeping ref_ticks alone when ref_data is needed wouldn't fail immediately, but lead to hard-to-debug issues.

bettio · 2026-02-03T17:02:26Z

Also, there is one more issue, that is not just about this PR, but more general.
[...]

~~@pguyot @mat-hek let's discuss how to fix this.~~

I completely overlooked a test I did, ignore this comment.

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

bettio

Review is still in progress, but meanwhile I submit comments for what I reviewed.

bettio · 2026-02-06T11:36:35Z

libs/estdlib/src/erlang.erl

+%% @end
+%%-----------------------------------------------------------------------------
+-spec monitor
+    (Type :: process, Pid :: pid() | atom(), [monitor_option()]) -> reference();


monitor can be also used with atoms or with tuples for processes on remote nodes.
I suggest introducing monitor_process_identifier() as Erlang does.

This is copy-pasted from the typespec above, I think it should be fixed in a separate PR

bettio · 2026-02-06T11:45:29Z

src/libAtomVM/opcodesswitch.h

+                                }
+                                globalcontext_get_process_unlock(glb, p);
+                            }
+                        } else if (!term_is_reference(recipient_term)) {


Personally I find more readable having just } else { at the end, instead of the previous condition negated.

This is not the previous condition negated

bettio · 2026-02-06T11:53:37Z

src/libAtomVM/opcodesswitch.h

+                            TRACE_SEND(ctx, x_regs[0], x_regs[1]);
+                            globalcontext_send_message(ctx->global, local_process_id, x_regs[1]);
+                        } else if (term_is_process_reference(recipient_term)) {
+                            int32_t local_process_id = term_process_ref_to_process_id(recipient_term);


This same code block is duplicated in jit.c. I suggest creating a helper function to avoid code duplication, such as: globalcontext_send_message_to_alias(GlobalContext *, term alias, term message);

Actually, the whole jit_send function is a duplicate of the case OP_SEND in opcodesswitch, so I believe it all should be unified.

bettio · 2026-02-06T11:54:31Z

src/libAtomVM/jit.c

        globalcontext_send_message(ctx->global, local_process_id, ctx->x[1]);
        ctx->x[0] = ctx->x[1];
+    } else if (term_is_process_reference(recipient_term)) {
+        int32_t process_id = term_process_ref_to_process_id(recipient_term);


See my comment about duplicated code in opcodesswitch.h.

bettio · 2026-02-06T11:59:08Z

tests/erlang_tests/test_refs_ordering.erl

-    check(Sorted, [A, B, C, D, E]) +
-        bool_to_n(Sorted < [make_ref()]) * 2 +
+    erlang:display([A, C, E, B, D]),
+    erlang:display(Sorted),


I think erlang:display() here is a leftover.

bettio · 2026-02-06T12:08:49Z

src/libAtomVM/term.h

    return ret;
 }

+static inline RefData term_to_ref_data(term t)


As soon as the last port change is reverted, I think this will be an orphan function. If so, I suggest removing it.

(I might be wrong since I'm still doing the review)

Actually as an additional feedback, I'm not sure this struct return is a good pattern, usually is more common in C using pointers when passing around structs.
See also AVMCCS-A007 in C_CODING_STYLE.md guide.

bettio · 2026-02-06T12:09:44Z

src/libAtomVM/term.h

+    const uint32_t *words;
+};
+
+typedef struct RefData RefData;


See my comment about term_to_ref_data(term t).

I've seen, how does it affect this line?

src/libAtomVM/nifs.c

bettio · 2026-02-06T12:15:42Z

src/libAtomVM/nifs.c

+    RefData ref_data = {
+        .type = RefTypeProcess,
+        .process = { .ref_ticks = globalcontext_get_ref_ticks(ctx->global), .process_id = ctx->process_id }
+    };


As a feedback I suggest a shorter lines here:

RefData ref_data = { .type = RefTypeProcess, .process = { .ref_ticks = globalcontext_get_ref_ticks(ctx->global), .process_id = ctx->process_id } };

Maybe a line-length limit in clang-format would be a good idea?

bettio · 2026-02-06T12:21:04Z

src/libAtomVM/term.h

+    return ref_data;
+}
+
+static inline term term_from_ref_data(RefData ref_data, Heap *heap)


See AVMCCS-A007 in C_CODING_STYLE.md guide.

However my question is: after reverting all the changes to port drivers, are we just using term_from_ref_data for process references? If so, this function might be way more complex than what we need now in our codebase now.

RefData can store any ref, as requested by @pguyot here: #2027 (comment). Please have a look at @pguyot's review, as I have an impression your preferences differ on this matter

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

mat-hek force-pushed the mf/upstream-aliases branch 8 times, most recently from e2c4216 to f30b1f3 Compare December 9, 2025 09:18

bettio requested changes Dec 9, 2025

View reviewed changes

mat-hek force-pushed the mf/upstream-aliases branch 9 times, most recently from 4b23228 to 07373d0 Compare December 12, 2025 14:47

mat-hek requested a review from bettio December 12, 2025 16:41

bettio requested changes Dec 15, 2025

View reviewed changes

bettio requested a review from pguyot December 15, 2025 15:34

mat-hek force-pushed the mf/upstream-aliases branch 2 times, most recently from 92db4b7 to 1bdef14 Compare December 15, 2025 16:13

pguyot reviewed Dec 15, 2025

View reviewed changes

mat-hek force-pushed the mf/upstream-aliases branch from ce57282 to eac7682 Compare December 19, 2025 09:45

mat-hek requested a review from pguyot December 19, 2025 13:16

mat-hek force-pushed the mf/upstream-aliases branch from f759361 to fe5d3c7 Compare January 21, 2026 10:42

CR

c882f24

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

mat-hek force-pushed the mf/upstream-aliases branch from fe5d3c7 to c882f24 Compare January 21, 2026 10:43

pguyot reviewed Jan 24, 2026

View reviewed changes

mat-hek added 4 commits January 28, 2026 16:25

Support for external and resource refs in RefData

5f8964c

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

CR fixes

78dbdc6

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

rename ref sizes

bf53366

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

fix formatting

1a2f460

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

mat-hek requested a review from pguyot January 28, 2026 15:28

bettio reviewed Jan 30, 2026

View reviewed changes

CR

ab5f3c8

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

mat-hek force-pushed the mf/upstream-aliases branch from d6ce43f to ab5f3c8 Compare February 2, 2026 13:28

mat-hek requested a review from bettio February 2, 2026 14:47

bettio requested changes Feb 3, 2026

View reviewed changes

mat-hek added 2 commits February 3, 2026 17:27

fix reference_process_pid_offset

8cb1807

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

fix parse_monitor_opts

e1aaebf

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

bettio requested changes Feb 3, 2026

View reviewed changes

mat-hek added 4 commits February 4, 2026 11:35

Revert ref_data back to ref_ticks in drivers

f036397

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

REF_SIZE -> TERM_BOXED_SHORT_REFERENCE_SIZE in drivers

303d1e8

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

Add aliases to test_refs_ordering

aef99bd

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

Make comparing process refs compliant with the BEAM

a5b8149

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

mat-hek requested a review from bettio February 4, 2026 17:27

bettio requested changes Feb 6, 2026

View reviewed changes

CR

a0f7887

Signed-off-by: Mateusz Front <mateusz.front@swmansion.com>

mat-hek force-pushed the mf/upstream-aliases branch from dc30e6b to a0f7887 Compare February 6, 2026 14:19

mat-hek requested a review from bettio February 6, 2026 14:30

bettio added this to the v0.7.0 milestone Feb 6, 2026

type	32 bits	64 bits
short refs	3 words	2 words
process aliases	4 words	3 words
resources	5 words (2 unused)	4 words (1 unused)

Conversation

mat-hek commented Dec 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mat-hek commented Dec 9, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mat-hek Dec 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mat-hek commented Dec 13, 2025

Uh oh!

bettio left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pguyot commented Dec 15, 2025

Uh oh!

pguyot left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mat-hek commented Dec 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bettio commented Jan 30, 2026

Uh oh!

bettio left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

mat-hek commented Dec 8, 2025 •

edited

Loading

mat-hek Dec 10, 2025 •

edited

Loading

mat-hek commented Dec 19, 2025 •

edited

Loading

bettio left a comment •

edited

Loading

bettio commented Feb 3, 2026 •

edited

Loading