Make toxcore code C++ compatible.

[iphydf]

It is still C code, so still compatible with C compilers as well. This
change lets us see more clearly where implicit conversions occur by
making them explicit.

---

This change is NOT reviewable. Use Github reviews instead.

---

This change is

[nbraud]

Reviewed 45 of 45 files at r1.
Review status: all files reviewed at latest revision, 7 unresolved discussions, some commit checks failed.

---

auto_tests/dht_test.c, line 531 at r1 (raw file):

                ck_assert_msg(count == 1, "Nodes in search don't know ip of friend. %u %u %u", i, j, count);
#if 0

Why is #if 0 required for C++ compat?

---

other/bootstrap_daemon/src/config.c, line 367 at r1 (raw file):

    while (config_setting_length(node_list)) {
        int address_resolved;

Any specific reason for moving the definition site (and losing cons)?

---

toxcore/list.c, line 116 at r1 (raw file):

static int resize(BS_LIST *list, uint32_t new_size)
{
    uint8_t *data = (uint8_t *)realloc(list->data, list->element_size * new_size);

The pointer could be const, same for ids.

---

toxcore/Messenger.c, line 2532 at r1 (raw file):

        uint32_t myfriend, dhtfriend;

It's strange for myfriend to be a uint32.
Perhaps friend_idx?

---

toxcore/TCP_connection.c, line 41 at r1 (raw file):

        ? (temp_pointer = (element_type *)realloc( \
              array, \
              (num) * sizeof(element_type)), \

This can overflow, and such a macro would actually be a nice place to catch it.
(Obviously out of scope for this PR)

---

toxcore/tox.c, line 845 at r1 (raw file):

    if (ret == USERSTATUS_INVALID) {
        SET_ERROR_PARAMETER(error, TOX_ERR_FRIEND_QUERY_FRIEND_NOT_FOUND);
        return (TOX_USER_STATUS)(TOX_USER_STATUS_BUSY + 1);

If the convention is to return an invalid enum value, I think there should be a macro defined for it (rather than risk pandemonium once another enum value gets added).

---

toxcore/tox.c, line 1352 at r1 (raw file):

}

/**********GROUP CHAT FUNCTIONS: WARNING Group chats will be rewritten so this might change ************/

Is moving the code from tox_group.c needed for C++ compat?

---

Comments from Reviewable

[iphydf]

I removed reviewable, so I'm going to reply in here:

• Why is if 0 required? It's not, that's just a drive-by cleanup. #if 0 is preferred to comments to comment out blocks of code in C. The main reason is that code formatters can see them, a second reason is that enabling it is a single bit change in the source code.
• Why move the definition and lose const? Because goto jumps over the initialisation, which is ill-formed in C++.
• Pointers should be const. That's not required for C++, and none of the other code does this. Perhaps we can do a transformation over the code at some point and make all non-changing names const.
• myfriend to friend_idx. Done.
• Overflow: I don't know if it actually can, but yes, by itself that could happen. Toxcore code tends not to check for this, because "we know that it can't overflow because we have 30% of the code in our heads at all times" is how most of the code is supposed to be read. We'll be looking into those things at some point.
• Returning invalid values: it should not return an invalid value. The value should be unspecified, and no code should rely on it being something known, but it should return valid values. Out of scope for this PR though.
• Moving code: no, not required, but another PR was going to do that, and I wanted this PR to be compatible with it. That PR is merged now.

[nurupo]

Was that rename really necessary? Might make it harder to merge iru's new branch later on.

[nbraud]

@nurupo Yes, friend is a C++ keyword.

[GrayHatter]

Again, yes. Friend in the context of DHT is problematic.

Breaking merges or otherwise this is a positive change.

[mannol]

LGTM

[nbraud]

@iphydf OK for me, I will open issues for the out-of-scope comments