Finish the messenger state plugin system

[e0ff]

I have wrote all the plugins for them and set them up. but it doesn't work yet, there is an issue with load and/or saving.

It seems to be working now, it passes the tests and I've tested it with uTox.

Depends on #1051.

---

This change is 

[iphydf]

Why doesn't this use write_cookie?

[iphydf]

Messenger_State_Plugin *temp = (Mess..

[e0ff]

Reviewable status: 0 of 1 LGTMs obtained

---

toxcore/DHT.c, line 2819 at r2 (raw file):

Previously, iphydf wrote…

Why doesn't this use write_cookie?

The DHT cookie is written differently, it has less padding. This is how the messenger cookie is written:

memset(data, 0, size32);
data += size32;
host_to_lendian32(data, MESSENGER_STATE_COOKIE_GLOBAL);
data += size32;

I was going to ask about this. Do you want me to get rid of the cookie function or I could add the extra padding before writing the messenger cookie.

---

toxcore/Messenger.c, line 2933 at r2 (raw file):

Previously, iphydf wrote…

Messenger_State_Plugin *temp = (Mess..

Done.

[iphydf]

Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @iphydf and @endoffile78)

---

toxcore/Messenger.c, line 2940 at r4 (raw file):

    const uint8_t index = m->options.state_plugins_length - 1;
    memset(&m->options.state_plugins[index], 0, sizeof(Messenger_State_Plugin));

memset on structures containing pointers does not necessarily yield a valid structure. The pointers inside may be messed up. You probably don't want a memset here at all, since you initialise all members, and you want to know about it if you don't.

---

toxcore/Messenger.c, line 2993 at r4 (raw file):

static uint32_t nospam_keys_size(const Messenger *m)
{
    (void) m;

No need for this. We don't warn on unused params.

---

toxcore/Messenger.c, line 2994 at r4 (raw file):

{
    (void) m;
    return (sizeof(uint32_t) + CRYPTO_PUBLIC_KEY_SIZE + CRYPTO_SECRET_KEY_SIZE);

Remove outer parentheses.

---

toxcore/Messenger.c, line 3003 at r4 (raw file):

        load_secret_key(m->net_crypto, (&data[sizeof(uint32_t)]) + CRYPTO_PUBLIC_KEY_SIZE);

        if (public_key_cmp((&data[sizeof(uint32_t)]), nc_get_self_public_key(m->net_crypto)) != 0) {

Replace (&data[sizeof(uint32_t)]) with data + sizeof(uint32_t). The former expresses "pointer to that one element" while the latter expresses "data at offset ...".

---

toxcore/Messenger.c, line 3007 at r4 (raw file):

        }
    } else {
        return STATE_LOAD_STATUS_ERROR;    /* critical */

If you invert the condition, you can use early error return.

---

toxcore/Messenger.c, line 3167 at r4 (raw file):

static State_Load_Status load_name(Messenger *m, const uint8_t *data, uint32_t length)
{
    if ((length > 0) && (length <= MAX_NAME_LENGTH)) {

Remove superfluous parentheses.

---

toxcore/Messenger.c, line 3191 at r4 (raw file):

static State_Load_Status load_status_message(Messenger *m, const uint8_t *data, uint32_t length)
{
    if ((length > 0) && (length <= MAX_STATUSMESSAGE_LENGTH)) {

Same here.

---

toxcore/Messenger.c, line 3201 at r4 (raw file):

static uint32_t status_size(const Messenger *m)
{
    (void) m;

Remove void-cast.

---

toxcore/Messenger.c, line 3226 at r4 (raw file):

static uint32_t tcp_relay_size(const Messenger *m)
{
    (void) m;

Remove.

---

toxcore/Messenger.c, line 3260 at r4 (raw file):

static uint32_t path_node_size(const Messenger *m)
{
    (void) m;

Remove.

---

toxcore/Messenger.c, line 3270 at r4 (raw file):

    data = state_write_section_header(data, MESSENGER_STATE_COOKIE_TYPE, 0, MESSENGER_STATE_TYPE_PATH_NODE);
    memset(nodes, 0, sizeof(nodes));
    unsigned int num = onion_backup_nodes(m->onion_c, nodes, NUM_SAVED_PATH_NODES);

const

---

toxcore/Messenger.c, line 3271 at r4 (raw file):

    memset(nodes, 0, sizeof(nodes));
    unsigned int num = onion_backup_nodes(m->onion_c, nodes, NUM_SAVED_PATH_NODES);
    int l = pack_nodes(data, NUM_SAVED_PATH_NODES * packed_node_size(net_family_tcp_ipv6), nodes, num);

const

---

toxcore/Messenger.c, line 3287 at r4 (raw file):

    if (length != 0) {
        int num = unpack_nodes(nodes, NUM_SAVED_PATH_NODES, nullptr, data, length, 0);

const

---

toxcore/Messenger.c, line 3300 at r4 (raw file):

static uint32_t end_size(const Messenger *m)
{
    (void) m;

Remove all these (I'm going to stop commenting on them now).

---

toxcore/state.c, line 73 at r4 (raw file):

}

uint8_t *state_write_cookie(uint8_t *data, uint32_t cookie, uint32_t length)

This function is used exactly once. Perhaps better inline it.

[e0ff]

Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @iphydf)

---

toxcore/Messenger.c, line 2940 at r4 (raw file):

Previously, iphydf wrote…

memset on structures containing pointers does not necessarily yield a valid structure. The pointers inside may be messed up. You probably don't want a memset here at all, since you initialise all members, and you want to know about it if you don't.

Done.

---

toxcore/Messenger.c, line 2993 at r4 (raw file):

Previously, iphydf wrote…

No need for this. We don't warn on unused params.

Done.

---

toxcore/Messenger.c, line 2994 at r4 (raw file):

Previously, iphydf wrote…

Remove outer parentheses.

Done.

---

toxcore/Messenger.c, line 3003 at r4 (raw file):

Previously, iphydf wrote…

Replace (&data[sizeof(uint32_t)]) with data + sizeof(uint32_t). The former expresses "pointer to that one element" while the latter expresses "data at offset ...".

Done.

---

toxcore/Messenger.c, line 3007 at r4 (raw file):

Previously, iphydf wrote…

If you invert the condition, you can use early error return.

Done.

---

toxcore/Messenger.c, line 3167 at r4 (raw file):

Previously, iphydf wrote…

Remove superfluous parentheses.

Done.

---

toxcore/Messenger.c, line 3191 at r4 (raw file):

Previously, iphydf wrote…

Same here.

Done.

---

toxcore/Messenger.c, line 3201 at r4 (raw file):

Previously, iphydf wrote…

Remove void-cast.

Done.

---

toxcore/Messenger.c, line 3226 at r4 (raw file):

Previously, iphydf wrote…

Remove.

Done.

---

toxcore/Messenger.c, line 3260 at r4 (raw file):

Previously, iphydf wrote…

Remove.

Done.

---

toxcore/Messenger.c, line 3270 at r4 (raw file):

Previously, iphydf wrote…

const

Done.

---

toxcore/Messenger.c, line 3271 at r4 (raw file):

Previously, iphydf wrote…

const

Done.

---

toxcore/Messenger.c, line 3287 at r4 (raw file):

Previously, iphydf wrote…

const

Done.

---

toxcore/Messenger.c, line 3300 at r4 (raw file):

Previously, iphydf wrote…

Remove all these (I'm going to stop commenting on them now).

Done.

---

toxcore/state.c, line 73 at r4 (raw file):

Previously, iphydf wrote…

This function is used exactly once. Perhaps better inline it.

Done.

[iphydf]

Reviewed 5 of 6 files at r5.
Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @iphydf and @endoffile78)

---

toxcore/Messenger.c, line 2940 at r4 (raw file):

Previously, endoffile78 (Endoffile) wrote…

Done.

What is done?

---

toxcore/Messenger.c, line 3003 at r4 (raw file):

Previously, endoffile78 (Endoffile) wrote…

Done.

Ok. When I put a comment, I generally assume that you'll fix similar issues around the code. Please also change the one 2 lines above this one.

---

toxcore/Messenger.c, line 3191 at r4 (raw file):

Previously, endoffile78 (Endoffile) wrote…

Done.

It doesn't seem so.

---

toxcore/Messenger.c, line 48 at r5 (raw file):

// TODO(endoffile78): Remove these, currently they need to be up here so they can be called in new_messenger
// Decide whether to move new_messenger or the state plugins + what they need
static uint32_t nospam_keys_size(const Messenger *m);

You could also remove all of these and have a single m_register_default_plugins or something below all the plugin functions.

---

toxcore/Messenger.c, line 2164 at r5 (raw file):

    friendreq_kill(m->fr);

    if (m->options.state_plugins) {

No need for the if. free(0) is valid.

---

toxcore/Messenger.c, line 2930 at r5 (raw file):

                             m_state_save_cb save_callback)
{
    Messenger_State_Plugin *temp = (Messenger_State_Plugin *) realloc(m->options.state_plugins,

Remove space after cast.

---

toxcore/Messenger.c, line 3010 at r5 (raw file):

    load_secret_key(m->net_crypto, (&data[sizeof(uint32_t)]) + CRYPTO_PUBLIC_KEY_SIZE);

    if (public_key_cmp((data + sizeof(uint32_t)), nc_get_self_public_key(m->net_crypto)) != 0) {

Remove parentheses around data + sizeof(uint32_t).

---

toxcore/Messenger.c, line 3022 at r5 (raw file):

    assert(sizeof(get_nospam(m->fr)) == sizeof(uint32_t));
    data = state_write_section_header(data, MESSENGER_STATE_COOKIE_TYPE, len, MESSENGER_STATE_TYPE_NOSPAMKEYS);
    *(uint32_t *)data = get_nospam(m->fr);

Maybe while we're at it, we can make this unportable code portable.

---

toxcore/Messenger.c, line 3323 at r5 (raw file):

    for (uint8_t i = 0; i < m->options.state_plugins_length; ++i) {
        const Messenger_State_Plugin plugin = m->options.state_plugins[i];

Copying it is probably not necessary. A ptr-to-const would be fine here.

[e0ff]

Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @iphydf and @endoffile78)

---

toxcore/Messenger.c, line 48 at r5 (raw file):

Previously, iphydf wrote…

You could also remove all of these and have a single m_register_default_plugins or something below all the plugin functions.

Done.

---

toxcore/Messenger.c, line 2164 at r5 (raw file):

Previously, iphydf wrote…

No need for the if. free(0) is valid.

Done.

---

toxcore/Messenger.c, line 2930 at r5 (raw file):

Previously, iphydf wrote…

Remove space after cast.

Done.

---

toxcore/Messenger.c, line 3010 at r5 (raw file):

Previously, iphydf wrote…

Remove parentheses around data + sizeof(uint32_t).

Done.

---

toxcore/Messenger.c, line 3022 at r5 (raw file):

Previously, iphydf wrote…

Maybe while we're at it, we can make this unportable code portable.

How would I do that?

---

toxcore/Messenger.c, line 3323 at r5 (raw file):

Previously, iphydf wrote…

Copying it is probably not necessary. A ptr-to-const would be fine here.

Is that what you wanted?

[codecov]

Codecov Report

Merging #1046 into master will increase coverage by <.1%.
The diff coverage is 95.8%.

@@           Coverage Diff            @@
##           master   #1046     +/-   ##
========================================
+ Coverage    82.8%   82.8%   +<.1%     
========================================
  Files          82      82             
  Lines       14594   14666     +72     
========================================
+ Hits        12084   12158     +74     
+ Misses       2510    2508      -2

Impacted Files	Coverage Δ
toxcore/DHT.c	76.6% <100%> (-0.2%)	⬇️
toxcore/state.c	64.7% <83.3%> (-6.8%)	⬇️
toxcore/Messenger.c	86.7% <96.1%> (+0.7%)	⬆️
toxcore/LAN_discovery.c	83.9% <0%> (-1.9%)	⬇️
toxcore/onion_client.c	95.5% <0%> (-0.7%)	⬇️
toxcore/tox.c	65.7% <0%> (-0.1%)	⬇️
toxav/toxav.c	69.1% <0%> (+0.9%)	⬆️
toxav/msi.c	65.6% <0%> (+1.9%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 1cd1836...08cc418. Read the comment docs.

[iphydf]

Remove "WIP" from the title if this is ready for review.

[iphydf]

Reviewed 1 of 6 files at r5, 1 of 2 files at r6.
Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @iphydf)

---

toxcore/Messenger.c, line 3022 at r5 (raw file):

Previously, endoffile78 wrote…

How would I do that?

Use lendian_from_host or something. Right now this code makes saves on big endian systems behave differently on little endian systems.

---

toxcore/Messenger.c, line 3323 at r5 (raw file):

Previously, endoffile78 wrote…

Did I do what you are talking about?

Yes, that's correct, but you can keep the const at the beginning (also keep the second const at the end).

[e0ff]

Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @iphydf)

---

toxcore/Messenger.c, line 3022 at r5 (raw file):

Previously, iphydf wrote…

Use lendian_from_host or something. Right now this code makes saves on big endian systems behave differently on little endian systems.

Done.

---

toxcore/Messenger.c, line 3323 at r5 (raw file):

Previously, iphydf wrote…

Yes, that's correct, but you can keep the const at the beginning (also keep the second const at the end).

Done.

[iphydf]

Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @endoffile78)

---

toxcore/Messenger.c, line 2963 at r7 (raw file):

    }

    set_nospam(m->fr, *(const uint32_t *)data);

Also fix this one.

[iphydf]

Reviewed 1 of 2 files at r7.
Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @endoffile78)

[e0ff]

Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @iphydf)

---

toxcore/Messenger.c, line 2963 at r7 (raw file):

Previously, iphydf wrote…

Also fix this one.

Done.

[iphydf]

Reviewed 1 of 1 files at r8.
Reviewable status: 1 change requests, 0 of 1 approvals obtained

[iphydf]

Let's hold off on this one until @ingvar1995 made a PR for NGC.

[iphydf]

Ingvar is not making a PR anytime soon, so let's go ahead with this.

[iphydf]

Reviewed 3 of 3 files at r9.
Reviewable status: 0 of 1 approvals obtained

[iphydf]

@sudden6 please still take a look when you have the chance.

[iphydf]

Reviewable status: complete! 1 of 1 approvals obtained