| Age | Commit message (Collapse) | Author |
|---|
|
new_api
|
|
|
|
messaging function.
This removes code duplication and allows us to easily add new message
types to the api without having to add new functions.
|
|
friend_get_connection_status
|
|
This is now documented in the API.
Ported programs in testing/ to this behaviour.
|
|
receive to recv in file receive functions.
Added TOX_MAX_FILENAME_LENGTH define.
|
|
offline.
|
|
|
|
|
|
Make sure some assumptions are always correct.
|
|
|
|
This function can be used to seek an incoming file tranfer right
before accepting it.
It is meant to be used to resume incomplete file tranfers by clients.
|
|
Pointer in callback will be NULL if length is 0.
|
|
file_id is a 32byte identifier that can be used by users to identify
file tranfers across core/client restarts in order to resume broken
file tranfers.
In avatar tranfers it corresponds to the hash of the avatar.
Added tox_file_get_file_id() function to api to obtain the file_id
of an ongoing file transfer.
If not set, core will generate a random one.
|
|
|
|
|
|
first 32 bytes.
Enforce length of filename in core when transfer is an avatar type
transfer to make things more safe.
|
|
|
|
contents pointed by proxy_address.
|
|
|
|
|
|
Removed time variable in packet struct and replaced it with sent
variable.
|
|
|
|
|
|
|
|
This allows clients to agree on what numbers mean what without having
it be set in core.
|
|
in the requested chunk callback.
For zero size transfers if the data sent is not the same length, the
file is assumed to be done.
|
|
|
|
|
|
Use correct double to integer conversion.
|
|
|
|
callback.
|
|
This is done so that the function now has the same parameters as the
request chunk callback.
|
|
|
|
|
|
If data is NULL and length non zero, TOX_ERR_NEW_NULL is set.
error is set to TOX_ERR_NEW_LOAD_BAD_FORMAT when load fails.
|
|
The first commit's message is:
TOX_STATUS -> TOX_USER_STATUS, is more specific
This is the 2nd commit message:
I pretty strongly believe tox_iteration needs to be renamed to a verb
There are several other noun functions in the API, but none of them *do* things.
I think even tox_do is better than tox_iteration.
tox_do_interval is one possibility, but I'm open to suggestions.
This is the 3rd commit message:
private_key -> secret_key
This is more consistent with modern/NaCl/elliptic cryptography, and also "pk", meaning public key, is all over other toxcore code and documentation. This will eliminate ambiguity.
This is the 4th commit message:
Rename some functions for pseudo-namespace consistency
The enum change results in long enum types, but I think consistency trumps
having a few less characters.
This is the 5th commit message:
TOX_FILE_KIND -> TOX_FILE_TYPE
This is more of an English thing than a consistency thing, but
TOX_FILE_KIND sounds funnier/stranger to me than TOX_FILE_TYPE.
This is the 6th commit message:
More specific file control function names
This is the 7th commit message:
Rename custom packet functions for pseudo-namespace consistency
This also has the issue with long enums... but I still think consistent enum names are better
This is the 8th commit message:
Rename functions for pseudo-namespace consistency
This is the 9th commit message:
Consistency with https://github.com/sonOfRa/tox4j/blob/master/doc/core-design.md#naming-conventions and the rest of the api
This is the 10th commit message:
Fix errors in previous function rename commits
This is the 11th commit message:
Shorten one error enum name
|
|
Attempted fix of connection checking to make it more stable.
|
|
|
|
Added tox_get_udp_port() to tests.
|
|
Since the functions are not new api like I put them in tox_old.{c,h}
|
|
Everything should work except resuming.
|
|
|
|
recv file numbers are now (file num in packet + 1) << 16
|
|
|
|
|
|
|
|
|
|
[uint8_t packet_id (PACKET_ID_FILE_SENDREQUEST)][uint8_t file_number]
[uint32_t type][uint64_t file_size][filename (currently max 255 bytes)]
recv file numbers are << by 16 to distinguish them from sending files.
|
|
|
