- the expando values need to be stored now that the lines are
reformattable, otherwise the old values are lost (and they depend on
context only available at the time the line is initially printed)
- the cache is collected from the special-vars evaluation code
- the cache is controlled by the textbuffer-formats code, and stored in
the text_buffer_format_rec
Hi folks!
I tried clang-static-analyzer on irssi 1.1.1, it seems like it
finds some things. Here's a diff, but there might be more that you would
want to check, or choose to work differently.
(in special-vars.c, ret is commands->data sometime)
I hope it's not too much trouble if reported as a confidential bug.
Thanks.
Originally found by oss-fuzz (issue 525) in get_ansi_color using ubsan.
After a lot of analysis I'm 99% sure this isn't security relevant so
it's fine to handle this publicly.
The fix is mainly adding a function that does it right and use it
everywhere. This is harder than it seems because the strtol() family of
functions doesn't have the friendliest of interfaces.
Aside from get_ansi_color(), there were other pieces of code that used
the same (out*10+(*in-'0')) pattern, like the parse_size() and
parse_time_interval() functions, which are mostly used for settings.
Those are interesting cases, since they multiply the parsed number
(resulting in more overflows) and they write to a signed integer
parameter (which can accidentally make the uints negative without UB)
Thanks to Pascal Cuoq for enlightening me about the undefined behavior
of parse_size (and, in particular, the implementation-defined behavior
of one of the WIP versions of this commit, where something like signed
integer overflow happened, but it was legal). Also for writing
tis-interpreter, which is better than ubsan to verify these things.
server.. Also added a FIXME: window item would need reference counting as
well, eg. "/EVAL win close;say hello" wouldn't work now.. But that's probably
job for the rewrite irssi. don't go use things like that :)
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@2381 dbcabf3a-b0e7-0310-adc4-f8d773084564
/alias echo foo; echo bar works properly). Also if there was any empty
commands they're now ignored.
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@2105 dbcabf3a-b0e7-0310-adc4-f8d773084564
of signals the expandos use. Also added "time changed" signal which gets
emitted when $Z changes.
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@1814 dbcabf3a-b0e7-0310-adc4-f8d773084564
you can use $C etc. in themes to print the active channel. printformat()
also expands it now correctly to the server/target where the message was
printed, not to the active window's server/target.
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@1802 dbcabf3a-b0e7-0310-adc4-f8d773084564
with % char. Used this with statusbar items - now for example "{error xxx}"
topic won't print it with error color.
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@1564 dbcabf3a-b0e7-0310-adc4-f8d773084564
treated as color code. So color codes can now be used in /SET prompt
string itself, but in none of the $variables it uses.
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@1324 dbcabf3a-b0e7-0310-adc4-f8d773084564
'prompt' is used when channel or query is active in window and
'prompt_window' is used with empty windows.
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@1005 dbcabf3a-b0e7-0310-adc4-f8d773084564
eval_special_string() : if any of the commands separated with ; used
the arguments ($0, $1, etc.) don't anymore append all the arguments
after other commands.
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@903 dbcabf3a-b0e7-0310-adc4-f8d773084564
If you /SET expand_escapes ON and type \n or \t to text line, they
will be expanded.
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@326 dbcabf3a-b0e7-0310-adc4-f8d773084564
- GNOME version isn't anymore build here so you don't need all that GTK and
GNOME crap to compile irssi-text.
- Some fixes to compile with -ansi -pedantic
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@200 dbcabf3a-b0e7-0310-adc4-f8d773084564
