Flyspray:: http://bugs.irssi.org/ Flyspray:: Irssi core bugs: Recently closed tasks 2015-10-02T16:41:52Z FS#919: usermode is never NULL http://bugs.irssi.org/index.php?do=details&task_id=919 2015-10-02T16:41:52Z zeebaasheDi9 irssi sends `MODE %s %s` to server even if `usermode = "";` in config. This could lead to nickname or activity exposure if irssi used behind bouncer. quickfix attached This could lead to nickname or activity exposure if irssi used behind bouncer.

quickfix attached]]>
FS#781: write_buffer causes awaylog to miss messages (patch included) http://bugs.irssi.org/index.php?do=details&task_id=781 2015-10-02T16:32:43Z Petteri Aimonen Steps to reproduce: /set write_buffer_size 2MB /set write_buffer_timeout 10min /away gone --- send a hilighted message from other client --- /away Irssi prints just the note that there is "1 new messages in awaylog", but the actual message is missing because it is still in the write buffer. The attached patch fixes this problem (applies against 0.8.15 or svn r5202).
/set write_buffer_size 2MB
/set write_buffer_timeout 10min
/away gone

--- send a hilighted message from other client ---

/away

Irssi prints just the note that there is "1 new messages in awaylog", but the actual message is missing because it is still in the
write buffer.

The attached patch fixes this problem (applies against 0.8.15 or svn r5202).
]]>
FS#867: DCC - Download fails to create downloaded file on NTFS mount http://bugs.irssi.org/index.php?do=details&task_id=867 2015-09-24T08:10:26Z MTA When trying to download a file to an NTFS mount, the following error occurs: "DCC can't create file /mnt/ntfs_mount/filename: Text file busy" I think the problem lies in the ``sig_dccget_connected`` function in the 'irc/dcc/dcc-get.c' file on line 239 for the following reason: We open the file 'tempfname' earlier in the function and its still open by the time we get to rename(). So rename fails because the file is Locked (not closed). NOTE: We fallback to rename because link() fails (probably because this filesystem does not support hardlinks. This is a virtualbox mount share from windows, mounted using vboxsf).
"DCC can't create file /mnt/ntfs_mount/filename: Text file busy"

I think the problem lies in the ``sig_dccget_connected`` function in the 'irc/dcc/dcc-get.c' file on line 239 for the following reason:

We open the file 'tempfname' earlier in the function and its still open by the time we get to rename(). So rename fails because the file is Locked (not closed).

NOTE: We fallback to rename because link() fails (probably because this filesystem does not support hardlinks. This is a virtualbox mount share from windows, mounted using vboxsf).]]>
FS#905: paste_buffer_join_lines() - Mangled text when line length exceeds 400 http://bugs.irssi.org/index.php?do=details&task_id=905 2015-09-21T22:13:52Z Fabian Kurz When paste_buffer_join_lines() joins lines together (by replacing newlines with spaces), it does so with a maximum length of 400 characters. Once this length limit is reached, it will go back to the last position that used to be a newline (now space), insert a newline there, after moving everything after it one position further. This shifting is done by the memmove() function, which allows moving overlapping areas of memory. However, this appears not to work correctly. The attached text (lorem.txt) will be joined up to the end of the secont to last line ("proident,"), then the new line will read " sunt in clpa qui officia deserunt mo llit anim id est laborum." - The "u" in "culpa" is missing and a space is added in "mollit". Replacing memmove with an (as far as I can tell) equivalent for-loop (see patch, which also includes the changes to fix issue 904) fixes this problem. But I don't really see why memmove() produces different results?
Once this length limit is reached, it will go back to the last position that used to be a newline (now space), insert a newline there, after moving everything after it one position further.

This shifting is done by the memmove() function, which allows moving overlapping areas of memory. However, this appears not to work correctly. The attached text (lorem.txt) will be joined up to the end of the secont to last line ("proident,"), then the new line will read " sunt in clpa qui officia deserunt mo llit anim id est laborum." - The "u" in "culpa" is missing and a space is added in "mollit".

Replacing memmove with an (as far as I can tell) equivalent for-loop (see patch, which also includes the changes to fix issue 904) fixes this problem. But I don't really see why memmove() produces different results?

]]>
FS#331: layout save does not handle !channel properly http://bugs.irssi.org/index.php?do=details&task_id=331 2015-09-19T11:08:19Z Juhapekka Tolvanen If I give commands "/layout save" and "/save", !-channels are not handled right: (Major clip) { name = "!ABCDEFGcoolchan"; chatnet = "ircnet"; autojoin = "yes"; }, (Major clip) windows = { 1 = { immortal = "yes"; name = "(status)"; level = "ALL"; }; 15 = { items = ( { type = "CHANNEL"; chat_type = "IRC"; name = "!coolchan"; tag = "IRCNet"; } ); }; I need to add that ABCDEFG to layout manually.
(Major clip)

{ name = "!ABCDEFGcoolchan"; chatnet = "ircnet"; autojoin = "yes"; },

(Major clip)

windows = {
1 = { immortal = "yes"; name = "(status)"; level = "ALL"; };
15 = {
items = (
{
type = "CHANNEL";
chat_type = "IRC";
name = "!coolchan";
tag = "IRCNet";
}
);
};

I need to add that ABCDEFG to layout manually.

]]>
FS#733: Please don't translate capital letter in channel name into small letter. http://bugs.irssi.org/index.php?do=details&task_id=733 2015-09-19T11:05:47Z AYANOKOUZI, Ryuunosuke irssi logs messages into $0.log file. But it seems to change capital letter in channel name into small letter. For example, #Japanese is converted into #japanese and $0 is set to #japanese. This behavior is not so much major issue, but in the case of Japanese(ISO-2022-JP) channel name it is a big problem. When convert #日本語 into iso-2022-jp code, and translate it into hex number. it is 231b2442467c4b5c386c1b2842 Generated file name($0) is 231b2442467c4b5c384c1b2842 As you can see, all capital letters 0x41-0x5a is translated into small letters 0x61-0x7a. Because of this, it is diffcult to restore original channel name from $0 string. So if you can, Please keep $0 oroginal channel name without any change.
When convert #日本語 into iso-2022-jp code, and translate it into hex number. it is

231b2442467c4b5c386c1b2842

Generated file name($0) is

231b2442467c4b5c384c1b2842

As you can see, all capital letters 0x41-0x5a is translated into small letters 0x61-0x7a. Because of this, it is diffcult to restore original channel name from $0 string.

So if you can, Please keep $0 oroginal channel name without any change.]]>
FS#780: irssi-proxy breaks RFC http://bugs.irssi.org/index.php?do=details&task_id=780 2015-07-28T16:47:35Z Shane Mc Cormack Recently when trying to use irssi-proxy with an RFC 1459 compliant irc parser, the lines output to the parser from irssi-proxy were considered invalid. Specifically, irssi-proxy incorrectly only uses "\n" line terminators rather than the "\r\n" terminators specified by the protocol - http://www.irchelp.org/irchelp/rfc/chapter2.html#c2_3_1
Specifically, irssi-proxy incorrectly only uses "\n" line terminators rather than the "\r\n" terminators specified by the protocol - http://www.irchelp.org/irchelp/rfc/chapter2.html#c2_3_1]]>
FS#882: /hilight (mispelled) and /highlight (absent) http://bugs.irssi.org/index.php?do=details&task_id=882 2014-07-01T19:29:15Z Andrew All documentation notes are spelled highlight, but the command is /hilight. Can be confusing. Suggestion to add /highlight and /dehighlight as permanent aliases. FS#538: Support for server redirecting (numeric 010) http://bugs.irssi.org/index.php?do=details&task_id=538 2014-06-30T16:20:59Z Quis I was trying to create some user-balancer, and found out UnrealIRCd used this to redirect users to other servers. tried it, and found out irssi does not support it: :server.domain.net 010 Schaap other.domain.net 6667 :Please use this Server/Port instead ERROR :Closing Link: TestUser[123.123.123.123] (This server is full.) I know this isn't in RFC 1459, but it would be nice to support. and found out UnrealIRCd used this to redirect users to other servers.

tried it, and found out irssi does not support it:

:server.domain.net 010 Schaap other.domain.net 6667 :Please use this Server/Port instead
ERROR :Closing Link: TestUser[123.123.123.123] (This server is full.)

I know this isn't in RFC 1459, but it would be nice to support.]]>
FS#888: Gentoo 12.1 emerge QA Notice: Package may exhibit random runtime failures http://bugs.irssi.org/index.php?do=details&task_id=888 2014-06-30T16:04:08Z Edhs4b QA Notice: Package triggers severe warnings which indicate that it may exhibit random runtime failures. dcc-send.c:296:2: warning: dereferencing type-punned pointer will break strict-aliasing rules Printout of Portage following the command: emerge irssi It does installs and seems to work properly. dcc-send.c:296:2: warning: dereferencing type-punned pointer will break strict-aliasing rules

Printout of Portage following the command: emerge irssi
It does installs and seems to work properly.]]>