A side-effect of 8deb618 is that `/save` may replace configuration files
that are symlinks with regular files. Fix this by resolving all
symlinks before renaming the temporary file.
This is modeled after glib's g_file_set_contents. It doesn't use that
function directly because the writing is done with GIOChannel
streaming-like writes and g_file_set_contents expects the whole thing to
be in-memory.
Main differences with g_file_set_contents:
- complete lack of win32 special casing (cygwin/WSL should work though)
- no fallocate() (linux only, but we don't know the size upfront, anyway)
- always calls fsync (glib skips it on btrfs or when not overwriting)
Other than that, it's the same old mkstemp + fsync + rename.
Use g_string_append_printf() instead of g_string_sprintfa() (which is considered deprecated.)
git-svn-id: file:///var/www/svn.irssi.org/SVN/irssi/trunk@5003 dbcabf3a-b0e7-0310-adc4-f8d773084564
and this is what I've been doing a few weeks now.. :) You really shouldn't
upgrade to this version without keeping a backup of the working one, since
this will break everything and at least notify list is broken - probably
something else too.
* On the way to 0.8.0 .. Major rewriting/rearranging code. There's
some changes in behaviour because I'm trying to make Irssi a bit
more compatible with EPIC.
* libPropList isn't needed anymore - I'm using my own configuration
library. This is mostly because different proplists worked a bit
differently everywhere and several people had problems with it.
It's also yet another extra library that you needed to compile
Irssi. New configuration library has several advantages:
You can add comments to configuration file and they also stay
there when it's saved.
It's not nearly as vulnerable as proplist. If some error occurs,
instead of just not reading anything it will try to continue if
possible. Also the error messages are written to irssi's text
window instead of stdout.
It can be managed more easily than proplist - setting/getting the
configuration is a lot more easier.
* Coding style changes - I'm not using gint, gchar etc. anymore,
they're just extra pain when moving code to non-glib projects and
syntax hilighting doesn't work by default with most editors ;)
Indentation style was also changed to K&R because of some political
reasons ;) And I'm already starting to like it.. :) It forces me
to split code to different functions more often and the result is
that the code gets more readable.
And finally I'm also using nst' all over the place.
+ /EVAL <commands> - Expand all the special variables from string and
run it. Commands can be split with ; character. See
docs/SPECIAL_VARS for more info.
+ Aliases are parsed just like /EVAL - arguments are in $0..$9.
+ Text formats are also parsed like /EVAL, arguments used to be in
$1..$9, now they're in $0..$8 so it messes up existing themes..
+ /SET [key [value]] - no more the '=' character. Boolean values
also need to be changed with ON/OFF/TOGGLE values (not yes/no).
Settings aren't saved to disk until you use /SAVE.
+ /TOGGLE <key> [ON/OFF] - same as /SET <key> TOGGLE
git-svn-id: http://svn.irssi.org/repos/irssi/trunk@163 dbcabf3a-b0e7-0310-adc4-f8d773084564
