This is a combination of 18 commits.
The first commit's message is:
Here is the initial banlist branch of RichardHitt/hexchat.
Changed files are only src/fe-gtk/{banlist.c,fe-gtk.h}.
This version works and contains my first efforts at selective
sensitization of radio buttons and control buttons.
From this point I intend to undertake a stepwise redesign.
Step 1 will be to make the existing banlist code work for
multiple simultaneous banlist windows (for different channels,
obviously). It will be a hackathon with the only goal of getting
it working.
Step 2 will be the objectization and alpha-stage tidying-up of
all the terrible looking stuff I will have done in Step 1.
This is the 2nd commit message:
Here's the post-Step-1 commit. It works for multiple banlist windows.
Note particularly what I've done to banlist.h.
Note that for many functions in banlist.c the argument is now
a banlist_info *, rather than a session *.
Note in banlist.c the initialization of array modes[] which
contains driving information for the checkboxes.
Of course those checkboxes aren't yet implemented. Maybe in Step 2
I will change to checkboxes from radio buttons; but definitely I will
change to letting modes[] drive processing.
This is the 3rd commit message:
Converted to checkboxes. Much additional work.
Note that the infrastructure for Auto-invite is not yet
present in the hexchat tree. I'm nearly done with banlist,
I think!
This is the 4th commit message:
Fleshed out 'invite'. Tagged masks uniformly, e.g. (b) (e) (I).
General cleanup, nearly at the point of beta quality.
This is the 5th commit message:
Added fourth mode type: quiet. Did lots and lots of cleanup. Beta-ready?
This is the 6th commit message:
Get the banlist timestamps properly sortable.
This is the 7th commit message:
Redesign the supports_foo() routines. Now they're responsible for setting
the flags in ->capable, ->readable, ->writeable.
This is the 8th commit message:
Deleted a couple of RBH comments.
This is the 9th commit message:
Now the ESC key will close the banlist window.
This is the 10th commit message:
Fix the fe-text occurrence of fe_add_ban_list().
This is the 11th commit message:
Fixed also fe_ban_list_end() and removed fe_is_banwindow().
This is the 12th commit message:
Use old-style initialization for array of structures modes[]
This is the 13th commit message:
Oops, incomplete regression of modes[] initialization. This fixes.
This is the 14th commit message:
Fixed strptime buy implementing a special version here.
Fixed column width concerns by setting resizable and autosize.
This is the 15th commit message:
Get rid of testing line.
This is the 16th commit message:
Changed to gtkutil_destroy_on_esc ()
This is the 17th commit message:
Remove no-longer-used functnion
This is the 18th commit message:
Minor cleanups to banlist.c, banlist.h
Please enter the commit message for your changes. Lines starting
with '#' will be ignored, and an empty message aborts the commit.
Author: RichardHitt <rbh00@netcom.com>
Committer: Richard Hitt <rbh00@f17.rbh00.pacbell.net>
Not currently on any branch.
Changes to be committed:
(use "git reset HEAD <file>..." to unstage)
modified: src/common/fe.h
modified: src/common/hexchat.h
modified: src/common/inbound.c
modified: src/common/modes.c
modified: src/common/proto-irc.c
modified: src/common/server.c
modified: src/fe-gtk/banlist.c
modified: src/fe-gtk/banlist.h
modified: src/fe-gtk/fe-gtk.c
modified: src/fe-gtk/fe-gtk.h
modified: src/fe-gtk/maingui.c
modified: src/fe-text/fe-text.c
So, to fix this an explicit FileAccess had to be set. Package.Open uses
the default FileShare.None option. This causes a conflict and throws the
'UnauthorizedAccessException' exception. So, we just tell it that we are
there to read it and everything is fine.
Remove old iffed-out version of is_del() as well, as a cleanup step; that change
was for not regarding left and right parens as delimiters. With this change,
mousing over a nickname will underline it even if it contains embedded formatting
data; it used to work with all formatting data but bold and reset; now it
works for all.
Unlike the other re_foo() functions it was not checking immediately
and returning if host_ret had already been filled in. This would
causes a memory leak since the previous GRegex would be lost.
