Re: patch for setting pane background color

2015-02-22 Thread Nicholas Marriott
On Sun, Feb 22, 2015 at 01:45:38AM -0600, J Raynor wrote: > > You can try to make tmux itself support BCE ... or add BCE support > > back on top of this like you had before. > > I don't know how to make tmux support terminals without BCE besides > what I already did. Can you give me some detail f

Re: [PATCH 1/2] Fix broken reflowing after insert/delete characters

2015-02-22 Thread Balazs Kezes
On 2015-01-25 00:18 +, Nicholas Marriott wrote: > Unfortunately I don't have the backtrace anymore, I was busy and the > easiest thing at the time was to revert it, it was definitely under > grid_view_insert_cells and I couldn't see anything obvious. I've kept running my main instance with thi

Inconsistencies when resizing panes

2015-02-22 Thread Ramon
Hi, I have an issue when resizing *vertical* panes (I'm using tmux 1.9a). To resize horizontal panes (i.e. those created with %) I hold and either the left or right arrow key, which works fine on both "Guake! 0.5.2" and xfce4-terminal. Now I tried to apply the same approach to resize vertical

Re: Inconsistencies when resizing panes

2015-02-22 Thread Balazs Kezes
On 2015-02-22 13:06 +0100, Ramon wrote: > Now I tried to apply the same approach to resize vertical panes (i.e. > when using and up/down arrow keys) it does *not* work when > using "Guake! 0.5.2" BUT it works great in xfce4-terminal. I've managed to install and reproduce this behavior for you. :)

Re: Inconsistencies when resizing panes

2015-02-22 Thread Ramon
On 02/22/2015 01:29 PM, Balazs Kezes wrote: >> p.s. Guake does not have any key-binding conflicting with >> left/right arrow keys. > But it does with and as you can see it in the > preferences (guake -p). It is bound to increase/decrease the height. If > you rebind these keys to something differ

[PATCH 3/4] tty_term: use an unsigned int for flags

2015-02-22 Thread Ben Boeckel
Signed-off-by: Ben Boeckel --- tmux.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tmux.h b/tmux.h index e296ac7..1ba0e83 100644 --- a/tmux.h +++ b/tmux.h @@ -1093,7 +1093,7 @@ struct tty_term { #define TERM_256COLOURS 0x1 #define TERM_EARLYWRAP 0x2 - int

[PATCH 0/4] GCC5 warnings

2015-02-22 Thread Ben Boeckel
Rawhide now has GCC5 and these warnings popped up when using it. Ben Boeckel (4): fix format specifiers cast format values where necessary tty_term: use an unsigned int for flags getopt: use const for place client.c | 6 +++--- cmd-capture-pane.c | 2 +- cmd-respawn-pane.c |

[PATCH 1/4] fix format specifiers

2015-02-22 Thread Ben Boeckel
GCC5 is warning about these things more it seems. Signed-off-by: Ben Boeckel --- client.c | 6 +++--- cmd-respawn-pane.c | 2 +- colour.c | 2 +- format.c | 30 +++--- input-keys.c | 2 +- input.c| 2 +- server-client.c

[PATCH 4/4] getopt: use const for place

2015-02-22 Thread Ben Boeckel
Prevents a warning here about const char* -> char* assignment tossing away the qualifier. The warning is still there for BSDoptarg, but that is specified by getopt.h to not be const. Also fix a stray whitespace error. Signed-off-by: Ben Boeckel --- compat/getopt.c | 4 ++-- 1 file changed, 2 in

[PATCH 2/4] cast format values where necessary

2015-02-22 Thread Ben Boeckel
The %x and %o specifers require unsigned arguments. Signed-off-by: Ben Boeckel --- cmd-capture-pane.c | 2 +- input-keys.c | 6 +++--- key-string.c | 2 +- layout.c | 2 +- tty-keys.c | 8 5 files changed, 10 insertions(+), 10 deletions(-) diff --git a/cmd

Re: [PATCH 2/4] cast format values where necessary

2015-02-22 Thread Nicholas Marriott
Thanks but: > } else > - xsnprintf(tmp, sizeof tmp, "\\%03o", line[i]); > + xsnprintf(tmp, sizeof tmp, "\\%03o", (unsigned > char)line[i]); %hho and (u_char)? > buf = cmd_capture_pane_append(buf,

Re: [PATCH 1/4] fix format specifiers

2015-02-22 Thread Nicholas Marriott
Looks good but I think PROTOCOL_VERSION should just be made unsigned. On Sun, Feb 22, 2015 at 11:34:23AM -0500, Ben Boeckel wrote: > GCC5 is warning about these things more it seems. > > Signed-off-by: Ben Boeckel > --- > client.c | 6 +++--- > cmd-respawn-pane.c | 2 +- > colour.c

Re: [PATCH 3/4] tty_term: use an unsigned int for flags

2015-02-22 Thread Nicholas Marriott
I always use int for flags. On Sun, Feb 22, 2015 at 11:34:25AM -0500, Ben Boeckel wrote: > Signed-off-by: Ben Boeckel > --- > tmux.h | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/tmux.h b/tmux.h > index e296ac7..1ba0e83 100644 > --- a/tmux.h > +++ b/tmux.h > @@ -1093

Re: [PATCH 3/4] tty_term: use an unsigned int for flags

2015-02-22 Thread Ben Boeckel
On Sun, Feb 22, 2015 at 18:46:33 +, Nicholas Marriott wrote: > I always use int for flags. Any reason why? It changes the meaning of some bits, but I can use u_int for its format usage as well (it is %x). --Ben -- Do

Re: [PATCH 3/4] tty_term: use an unsigned int for flags

2015-02-22 Thread Nicholas Marriott
No particular reason but I don't think we should change them, I'd just cast it or leave the warning. On Sun, Feb 22, 2015 at 02:03:17PM -0500, Ben Boeckel wrote: > On Sun, Feb 22, 2015 at 18:46:33 +, Nicholas Marriott wrote: > > I always use int for flags. > > Any reason why? It changes the

[PATCHv2 3/4] cast format values where necessary

2015-02-22 Thread Ben Boeckel
The %x and %o specifers require unsigned arguments. Signed-off-by: Ben Boeckel --- cmd-capture-pane.c | 2 +- cmd-show-messages.c | 2 +- input-keys.c| 6 +++--- key-string.c| 2 +- layout.c| 2 +- tty-keys.c | 8 6 files changed, 11 insertions(+),

[PATCHv2 0/4] GCC5 warnings

2015-02-22 Thread Ben Boeckel
Comments addressed (patches a little reordered to make the protocol switch first). Ben Boeckel (4): make PROTOCOL_VERSION unsigned fix format specifiers cast format values where necessary getopt: use const for place client.c| 4 ++-- cmd-capture-pane.c | 2 +- cmd-respawn-

[PATCHv2 4/4] getopt: use const for place

2015-02-22 Thread Ben Boeckel
Prevents a warning here about const char* -> char* assignment tossing away the qualifier. The warning is still there for BSDoptarg, but that is specified by getopt.h to not be const. Also fix a stray whitespace error. Signed-off-by: Ben Boeckel --- compat/getopt.c | 4 ++-- 1 file changed, 2 in

[PATCHv2 2/4] fix format specifiers

2015-02-22 Thread Ben Boeckel
GCC5 is warning about these things more it seems. Signed-off-by: Ben Boeckel --- client.c| 4 ++-- cmd-respawn-pane.c | 2 +- cmd-show-messages.c | 2 +- colour.c| 2 +- format.c| 30 +++--- input-keys.c| 2 +- input.c

[PATCHv2 1/4] make PROTOCOL_VERSION unsigned

2015-02-22 Thread Ben Boeckel
Signed-off-by: Ben Boeckel --- tmux.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tmux.h b/tmux.h index e296ac7..87ed093 100644 --- a/tmux.h +++ b/tmux.h @@ -19,7 +19,7 @@ #ifndef TMUX_H #define TMUX_H -#define PROTOCOL_VERSION 8 +#define PROTOCOL_VERSION 8u #includ

Re: The history of tmux

2015-02-22 Thread Ben Boeckel
On Mon, 16 Feb, 2015 at 17:27:35 GMT, Christopher Ritsen wrote: > The point I was trying to make was that it's not necessarily the > committer who is doing the work (they might be just applying a patch received > from elsewhere), but it would be expected that they would be attributed as the > autho

patch for bug in window_copy_append_selection

2015-02-22 Thread J Raynor
There's a bug in window_copy_append_selection that causes tmux to free memory that shouldn't be freed. To reproduce the problem: * Set mode-keys to vi so you can use A to append to a buffer * Enter copy mode and copy some text * Use lsb to note that the buffer names look normal * Enter copy m

Re: [tmux:tickets] #174 Copy in copy mode without leaving copy mode

2015-02-22 Thread J Raynor
> You could maybe make copy-selection accept an argument like copy-pipe. I > wouldn't make it do the full getopt dance just yet but just accepting > "-n" or "-x" or something might work. I've attached a patch that allows you to pass -x to the key bindings for append-selection, copy-selection, and