1 TODO for 1.1 And Beyond
2 =======================
4 lib/silccrypt ****PARTLY DONE****
7 o Implement the defined SilcDH API. The definition is in
8 lib/silccrypt/silcdh.h.
10 o SSH2 public keys support, allowing the use of SSH2 public keys in
13 o OpenPGP certificate support, allowing the use of PGP public keys
16 o SILC PKCS (silcpkcs.h) reorganizing when other PK supports added.
17 Move the SILC Public Key routines away from the crypto library into
18 the core library (silccore). silc_pkcs_public/private_key_* routines
19 to silc_public/private_key_* routines. The silc_public_key_* routines
20 should also automatically handle SILC Public Keys, and other keys
21 and certificates as well. Add fe. silcpk.h into silccore. It should
22 also include the Public Key Payload encoding and decoding routines.
26 o Cipher optimizations (asm, that this) at least for i386 would be nice.
31 lib/silccore/silcpacket.[ch] ****DONE****
32 ============================
36 o New SILC Packet API.
39 lib/silccore/silcpacket.[ch]
40 ============================
42 o IV Included flag support, UDP transport support
45 lib/silccore/silcid.[ch] ****DONE****
46 ========================
48 o Add silc_id_str2id to accept the destination buffer as argument
49 and thus not require any memory allocation. Same will happen
50 with silc_id_payload_* functions.
52 o silc_id_str2id, silc_id2str to non-allocating routines.
58 o All payload encoding routines should take SilcStack as argument.
61 lib/silcutil ****PARTLY DONE****
64 o Compression routines are missing. The protocol supports packet
65 compression thus it must be implemented. SILC Zip API must be
68 o Add builtin SOCKS and HTTP Proxy support, well the SOCKS at least.
69 SILC currently supports SOCKS4 and SOCKS5 but it needs to be compiled
76 #define SilcBit(b) unsigned int b : 1
79 lib/silcutil/silcbuffer.h ****DONE****
80 =========================
82 o Remove the `truelen' field from SilcBuffer as it is entirely
83 redundant since we can get the true length of the buffer by
84 doing buffer->end - buffer->header. Add silc_buffer_truelen
85 macro instead. Consider also removing `len' field too since
86 it effectively is buffer->tail - buffer->data, and adding
87 silc_buffer_len macro can do the same. These would save
88 totally 8 bytes of memory per buffer.
91 lib/silcutil/silcbuffmt.[ch] ****PARTLY DONE****
92 ============================
94 o SilcStack aware silc_buffer_unformat
96 o silc_buffer_format reallocates automatically
101 lib/silcutil/silcstack.[ch] ****DONE****
102 ===========================
104 o Data stack implementation
107 lib/silcutil/silcstream.[ch] ****DONE****
108 ============================
110 o Add abstract SilcStream.
113 lib/silcutil/silcsocketstream.[ch] ****PARTY DONE****
114 ==================================
116 o Add SilcSocketStream.
124 o lib/silcutil/epoc routines missing or not completed.
126 o The PKCS#1 also calls global RNG (even though it is not used
127 currently in SILC, the interface allows its use).
129 o Something needs to be thought to the logging globals as well,
130 like silc_debug etc. They won't work on EPOC. Perhaps logging
131 and debugging is to be disabled on EPOC.
134 lib/silcutil/silcschedule*.[ch] ****DONE****
135 ===============================
137 o Scheduler can be optimized for FD tasks by changing the fd_queue
138 to SilcHashTable instead of using linked list. We need to do
139 one-to-one mapping of FD to task and hash table is more efficient
142 Also redefine the silc_select to perhaps return a separate
143 structure of the events that actually occurred, instead of
144 returning the events in the fd_list which is then traversed
145 in the generic code to find the changed events. This can be
146 made faster by having own struct which includes only the
147 changed events, thus the tarversing is faster since the whole
148 fd_list is not traversed anymore (it is still traversed in the
149 silc_select but at least it removes one extra tarversing later
152 Other task queues should be changed to use SilcList.
154 o Add SILC scheduler's internal routines into a table of implementation
155 function pointers, that the generic code then takes as extern from
156 implementation. These are the silc_schedule_internal_* routines.
158 o Change SILC_TASK_CALLBACK to non-static, and remove the macro
159 SILC_TASK_CALLBACK_GLOBAL.
162 lib/silcutil/silcasync.[ch] ****DONE****
163 ===========================
165 o Add SilcAsyncOperation to utility library. Any function that takes
166 callback as an argument must/should return SilcAsyncOperation.
169 lib/silcutil/silctime.[ch] ****DONE****
170 ===========================
174 o system time, universal, generalized.
180 o Import TFM. Talk to Tom to add the missing functions. Use TFM in
181 client and client library, but TMA in server, due to the significantly
182 increased memory consumption with TFM, and the rare need for public
183 key operations in server.
185 o The SILC MP API function must start returning indication of success
186 and failure of the operation.
188 o Do SilcStack support for silc_mp_init, silc_mp_init_size and other
189 any other MP function (including utility ones) that may allocate
192 o All utility functions should be made non-allocating ones.
195 lib/silcasn1 ****PARTLY DONE****
200 o Header documentation missing.
202 o Some string encodings missing (copy/paste matter).
204 o Negative integer encoding
210 o PKIX implementation
213 lib/silcutil/silcfsm.[ch] ****DONE****
214 =========================
216 o SILC Finite State Machine API. Replaces SILC Protocol API,
219 lib/silcutil/silcnet*, lib/silcutil/*/silc*net* ****PARTLY DONE****
220 ===============================================
224 o New network interfaces
226 o Other functions should remain as they are since these new functions have
227 to use them. This way we also provide them for applications that want
228 to handle the sockets by themself.
234 o (Re)write commands/command replys.
236 o (Re)write notify handling.
238 o The SERVER_SIGNOFF notify handing is not optimal, because it'll
239 cause sending of multiple SIGNOFF notify's instead of the one
240 SERVER_SIGNOFF notify that the server received. This should be
241 optimized so that the only SERVER_SIGNOFF is sent and not
242 SIGNOFF of notify at all (using SIGNOFF takes the idea about
243 SERVER_SIGNOFF away entirely).
245 o Another SERVER_SIGNOFF opt/bugfix: Currently the signoff is
246 sent to a client if it is on same channel as the client that
247 signoffed. However, the entire SERVER_SIGNOFF list is sent to
248 the client, ie. it may receive clients that was not on the
249 same channel. This is actually against the specs. It must be
250 done per channel. It shouldn't receive the whole list just
251 because one client happened to be on same channel.
253 o MAYBE: The SilcChannelClientEntry can be:
257 where address is SilcClientEntry address XOR SilcChannelEntry.
258 You can get SilcClientEntry by doing client = chl->address XOR channel,
259 and SilcChannelEntry by doing channel = chl->address XOR client.
260 As long as the other pointer is always available when accessing the
261 structure this can be done.
263 o Add reference counters to all Silc*Entry structures
265 o SERVICEs support (plugin, SIM)
267 o If client's public key is saved in the server (and doing public key
268 authentication) then the hostname and the username information could
269 be taken from the public key. Should be a configuration option!
271 o Add a timeout to handling incoming JOIN commands. It should be
272 enforced that JOIN command is executed only once in a second or two
273 seconds. Now it is possible to accept n incoming JOIN commands
274 and process them without any timeouts. THis must be employed because
275 each JOIN command will create and distribute the new channel key
276 to everybody on the channel (Fix this to 0.9.x).
278 o Related to above. If multiple JOINs are received in sequence perhaps
279 new key should be created only once, if the JOINs are handeled at the same
280 time. Now we create multiple keys and never end up using them because
281 many JOINs are processed at the same time in sequence. Only the last
282 key ends up being used.
284 o The CMODE cipher & hmac change problem (#101).
290 o Some form of rewrite to make it more efficient.
292 o Clear up interfaces.
294 o Remove silc_client_run and silc_client_run_one from calling SilcSchedule.
295 Leave the scheduling entirely to application.