View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0002182 | unreal | ircd | public | 2004-11-15 00:42 | 2007-04-27 03:49 |
| Reporter | tabrisnet | Assigned To | |||
| Priority | normal | Severity | minor | Reproducibility | always |
| Status | closed | Resolution | suspended | ||
| Product Version | 3.2.2 | ||||
| Summary | 0002182: continuation of 2073, tokens.txt vs msg.h | ||||
| Description | msg.h and tokens.txt still disagree, and i'm unsure what definitions to be using. and I am writing a services package that uses the unreal token system. which file is authoritative? Seems to me, neither are. esp with comments from bug 2073, it seems that the only authoritative source is each individual C file. I can't do it right now, but maybe I or someone else can regenerate the files. | ||||
| Additional Information | in msg.h, SVSJOIN has been deleted. in tokens.txt, the following have been deleted ADDMODE ADDOMODE ADMIN AWAY CONNECT ERROR GNOTICE GOPER HASH HELP HELPOP INFO INVITE KILL KLINE LINKS LIST NAMES NICK OPER PASS PING PONG PRIVMSG QUIT SERVER SQUIT STATS SUMMON SVSMODE TIME TOPIC USER USERS VERSION WALLOPS WHO WHOIS WHOWAS ZLINE in tokens.txt, the following have been added ADDMOTD ADDOMOTD CHGNAME CYCLE DCCDENY EOS MODULE NEWJOIN SENDSNO SVS2SNO SVSLUSERS SVSMOTD SVSNLINE SVSSILENCE SVSSNO SVSWATCH TEMPSHUN UNDCCDENY The following tokens have been redefined in tokens.txt SVSJOIN | ||||
| 3rd party modules | |||||
|
|
Does this bug still apply? (since tokens.txt was updated on February 28 2005) |
|
|
Unfortunately, yes. I just wrote a perl-script to parse msg.h (btw, that file really needs to be regenerated. it's a nightmare to parse) I'll email you the files I generated to do this. the bugs system isn't letting me upload atm. Won't allow me to upload a perl script, and it's taking forever to accept my TXT file (and I have three of them) |
|
|
Looking at the email I got, walking trough the changes, I don't see many bad items left. Note that all commands that should not be used remotely, such as WHOWAS, USERHOST, WHO, etc. should not be included in tokens.txt [now don't ask me why tokens for such cmds exist ;).] List of removed tokens & comments attached: -ADCHAT x <== to be added -ADDLINE z -BOTSERV BS -CHANSERV j -DIE R -GLINE } -GNOTICE Z -GOPER [ -HASH S -HELPSERV s -IDENTIFY g -INFOSERV BO -ISON K -KLINE W -LIST ( -MEMOSERV l -MKPASSWD y -NAMES ? -NICKSERV i -OPER ; -OPERSERV k -POST BN -PRIVMSG ! <== is called 'PRIVATE' in tokens.txt, renamed to PRIVMSG -REMGLINE BG -SERVICE I -SERVICES m -STATSERV AH -SUMMON 1 -USER % -USERHOST J -USERS 3 -WALLOPS = <== to be added -WHOWAS $ -ZLINE q Actually I see some other cmds that are still in tokens.txt, but should be removed, such as WATCH. |
|
|
Hmm. not sure that DIE or RESTART shouldn't have tokens. I remember that there used to be an effort to allow remote-restart. It was discontinued due to limited syntax making it very difficult to do safely. Further, this is, if nothing else, proof that msg.h and tokens.txt need to be regenerated. This may also include going through all the other files as well and removing the unneeded/obsolete/inane TOK_ decls. Meanwhile, I don't know what POST is, nor what the point of having a TOK for SUMMON is (it always was optional, and no modern ircd seems to have it). Seems there are a lot of weird tokens. like why one for CS anyway? It's an alias. Is this an anachronism from 3.1 or 3.0? I never followed the old codebase. |
|
|
by the way, please don't make USERHOST local-only. I admit that we probably don't have to have a token for it, but services may need it to know the users current cloakhost. I can't think of any other way for it to be grabbed if the user has /umode -x, /umode +x. If you want to provide a better way (xref bugs 2565 & 2613), please do, but that is the only aparrent method available at present. |
|
|
There are still issues with 3.2.4, like tokens.txt has SVSSNO but msg.h does not. If another update is desired, I could probably do a run, or just upload the script I use to parse everything. I am as yet unclear on ordering in tokens.txt, if it's manual or can be automated (sort by token?). I have a rough perl script that can extract the MSG_ and TOK_ from the various C files and sort by token if desired. (to be more accurate, you grep first, then pipe through the script) |
|
|
[quote]There are still issues with 3.2.4, like tokens.txt has SENDSNO but msg.h does not.[/quote] Huh? $ grep SENDSNO include/ -r include/msg.h:#define MSG_SENDSNO "SENDSNO" include/msg.h:#define TOK_SENDSNO "Ss" $ grep SENDSNO doc/technical/token.txt SENDSNO Ss |
|
|
On a sidenote, I actually did grep tokens from all files and well. what you said.. {except thanks to the power of bash scripting ;)}. But I excluded all the tokens that have no need to be added (which are many). If you see certain tokens that you think should be in tokens.txt though, let me know... I can always have made a mistake of course. |
|
|
Sorry, thinko. I edited the post above, s/SENDSNO/SVSSNO/. Also missing is SVSWATCH from msg.h (is in tokens.txt). Been working with it today while switching off studying for an american-government exam/test. Also potentially of value, just go through and yank all the TOK_ definitions that shouldn't exist out of the code. There's a couple iirc that have MSG_ but no TOK_ anyhow (I only remember this b/c my parse-script had to skip cases where one or the other wasn't defined/existent). I should make note that many of us services-coders might want to see that diff before it gets committed tho, just in case we see something we use, and thus will need an alternative command, method of derivation, or for it to remain. |
|
|
Well, things like USERHOST should for sure not be used by services, I'm pretty sure you know that actually by yourself... EDIT: Also removing tokens from token.txt or msg.h does not mean the command does not accept remote-server stuff, it's not magic.. or something.. Of course the intention is clear however, unless it got out of token.txt (or not in token.txt) by mistake. |
|
|
I admit that USERHOST probably should not be used, but the question is, what is the alternative? That's what the services coders want to know. I believe that's the point of 2565 and 2613. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2004-11-15 00:42 | tabrisnet | New Issue | |
| 2005-08-19 15:54 | syzop | Note Added: 0010372 | |
| 2005-08-20 16:57 | tabrisnet | Note Added: 0010378 | |
| 2005-08-20 16:57 | tabrisnet | Note Edited: 0010378 | |
| 2005-08-23 18:06 | syzop | Note Added: 0010382 | |
| 2005-08-23 18:07 | syzop | Note Edited: 0010382 | |
| 2005-08-23 19:36 | tabrisnet | Note Added: 0010383 | |
| 2005-08-23 23:18 | tabrisnet | Note Added: 0010385 | |
| 2005-08-23 23:20 | tabrisnet | Note Edited: 0010383 | |
| 2006-03-15 15:01 | tabrisnet | Note Added: 0011374 | |
| 2006-03-15 15:09 | syzop | Note Added: 0011375 | |
| 2006-03-15 15:10 | syzop | Note Added: 0011376 | |
| 2006-03-15 15:19 | tabrisnet | Note Edited: 0011374 | |
| 2006-03-15 15:20 | tabrisnet | Note Added: 0011378 | |
| 2006-03-15 15:23 | tabrisnet | Note Edited: 0011378 | |
| 2006-03-15 15:28 | tabrisnet | Note Edited: 0011378 | |
| 2006-03-15 15:54 | syzop | Note Added: 0011381 | |
| 2006-03-15 15:55 | syzop | Note Edited: 0011381 | |
| 2006-03-15 16:08 | tabrisnet | Note Added: 0011382 | |
| 2007-04-27 03:49 |
|
Status | new => closed |
| 2007-04-27 03:49 |
|
Resolution | open => suspended |
