In order to list cross-certifiable host keys in the GUI specials menu,
the SSH backend has been inventing new values on the end of the
Telnet_Special enumeration, starting from the value TS_LOCALSTART.
This is inelegant, and also makes it awkward to break up special
handlers (e.g. to dispatch different specials to different SSH
layers), since if all you know about a special is that it's somewhere
in the TS_LOCALSTART+n space, you can't tell what _general kind_ of
thing it is. Also, if I ever need another open-ended set of specials
in future, I'll have to remember which TS_LOCALSTART+n codes are in
which set.
So here's a revamp that causes every special to take an extra integer
argument. For all previously numbered specials, this argument is
passed as zero and ignored, but there's a new main special code for
SSH host key cross-certification, in which the integer argument is an
index into the backend's list of available keys. TS_LOCALSTART is now
a thing of the past: if I need any other open-ended sets of specials
in future, I can add a new top-level code with a nicely separated
space of arguments.
While I'm at it, I've removed the legacy misnomer 'Telnet_Special'
from the code completely; the enum is now SessionSpecialCode, the
struct containing full details of a menu entry is SessionSpecial, and
the enum values now start SS_ rather than TS_.
Now there's a centralised routine in misc.c to do the sanitisation,
which copies data on to an outgoing bufchain. This allows me to remove
from_backend_untrusted() completely from the frontend API, simplifying
code in several places.
Two use cases for untrusted-terminal-data sanitisation were in the
terminal.c prompts handler, and in the collection of SSH-2 userauth
banners. Both of those were writing output to a bufchain anyway, so
it was very convenient to just replace a bufchain_add with
sanitise_term_data and then not have to worry about it again.
There was also a simplistic sanitiser in uxcons.c, which I've now
replaced with a call to the good one - and in wincons.c there was a
FIXME saying I ought to get round to that, which now I have!
'struct draw_ctx' has a structure tag inside gtkwin.c, so as per this
week's standard practice, let's expose the tag elsewhere so that
pointers declared that way can't be confused with anything else.
This is another major source of unexplained 'void *' parameters
throughout the code.
In particular, the currently unused testback.c actually gave the wrong
pointer type to its internal store of the frontend handle - it cast
the input void * to a Terminal *, from which it got implicitly cast
back again when calling from_backend, and nobody noticed. Now it uses
the right type internally as well as externally.
Nearly every part of the code that ever handles a full backend
structure has historically done it using a pair of pointer variables,
one pointing at a constant struct full of function pointers, and the
other pointing to a 'void *' state object that's passed to each of
those.
While I'm modernising the rest of the code, this seems like a good
time to turn that into the same more or less type-safe and less
cumbersome system as I'm using for other parts of the code, such as
Socket, Plug, BinaryPacketProtocol and so forth: the Backend structure
contains a vtable pointer, and a system of macro wrappers handles
dispatching through that vtable.
Same principle again - the more of these structures have globally
visible tags (even if the structure contents are still opaque in most
places), the fewer of them I can mistake for each other.
That's one fewer anonymous 'void *' which might be accidentally
confused with some other pointer type if I misremember the order of
function arguments.
While I'm here, I've made its pointer-nature explicit - that is,
'Ldisc' is now a typedef for the structure type itself rather than a
pointer to it. A stylistic change only, but it feels more natural to
me these days for a thing you're going to eventually pass to a 'free'
function.
This is a cleanup I started to notice a need for during the BinarySink
work. It removes a lot of faffing about casting things to char * or
unsigned char * so that some API will accept them, even though lots of
such APIs really take a plain 'block of raw binary data' argument and
don't care what C thinks the signedness of that data might be - they
may well reinterpret it back and forth internally.
So I've tried to arrange for all the function call APIs that ought to
have a void * (or const void *) to have one, and those that need to do
pointer arithmetic on the parameter internally can cast it back at the
top of the function. That saves endless ad-hoc casts at the call
sites.
In commit 528513dde I absentmindedly replaced a write to the local
variable 'need_size' of drawing_area_setup with a write to
inst->drawing_area_setup_needed, imagining that they had the same
effect. But actually, need_size was doing two jobs and I only replaced
one of them: it was also the variable that indicated that the logical
terminal size had changed and so we had to call term_size() to make
the terminal.c data structures resize themselves appropriately. The
loss of that call also inhibited generation of SIGWINCH.
NFC for the moment, because the bufchain is always specially
constructed to hold exactly the same data that would have been passed
in to the function as a (pointer,length) pair. But this API change
allows get_userpass_input to express the idea that it consumed some
but not all of the data in the bufchain, which means that later on
I'll be able to point the same function at a longer-lived bufchain
containing the full stream of keyboard input and avoid dropping
keystrokes that arrive too quickly after the end of an interactive
password prompt.
Changing the window's font size with Alt-< or Alt-> was not setting
any of the flags that make drawing_area_setup consider itself to have
been non-spuriously called, so the real window would enlarge without
the backing surface also doing so.
In GTK 3.10 and above, high-DPI support is arranged by each window
having a property called a 'scale factor', which translates logical
pixels as seen by most of the GTK API (widget and window sizes and
positions, coordinates in the "draw" event, etc) into the physical
pixels on the screen. This is handled more or less transparently,
except that one side effect is that your Cairo-based drawing code had
better be able to cope with that scaling without getting confused.
PuTTY's isn't, because we do all our serious drawing on a separate
Cairo surface we made ourselves, and then blit subrectangles of that
to the window during updates. This has two bad consequences. Firstly,
our surface has a size derived from what GTK told us the drawing area
size is, i.e. corresponding to GTK's _logical_ pixels, so when the
scale factor is (say) 2, our drawing takes place at half size and then
gets scaled up by the final blit in the draw event, making it look
blurry and unpleasant. Secondly, those final blits seem to end up
offset by half a pixel, so that a second blit over the same
subrectangle doesn't _quite_ completely wipe out the previously
blitted data - so there's a ghostly rectangle left behind everywhere
the cursor has been.
It's not that GTK doesn't _let_ you find out the scale factor; it's
just that it's in an out-of-the-way piece of API that you have to call
specially. So now we do: our backing surface is now created at a pixel
resolution matching the screen's real pixels, and we translate GTK's
scale factor into an ordinary cairo_scale() before we commence
drawing. So we still end up drawing the same text at the same size -
and this strategy also means that non-text elements like cursor
outlines and underlining will be scaled up with the screen DPI rather
than stubbornly staying one physical pixel thick - but now it's nice
and sharp at full screen resolution, and the subrectangle blits in the
draw event are back to affecting the exact set of pixels we expect
them to.
One silly consequence is that, immediately after removing the last
one, I've installed a handler for the GTK "configure-event" signal!
That's because the GTK 3 docs claim that that's how you get notified
that your scale factor has changed at run time (e.g. if you
reconfigure the scale factor of a whole monitor in the GNOME settings
dialog). Actually in practice I seem to find out via the "draw" event
before "configure" bothers to tell me, but now I've got a usefully
idempotent function for 'check whether the scale factor has changed
and sort it out if so', I don't see any harm in calling it from
anywhere it _might_ be useful.
I've been using that signal since the very first commit of this source
file, as a combined way to be notified when the size of the drawing
area changes (typically due to user window resizing actions) and also
when the drawing area is first created and available to be drawn on.
Unfortunately, testing on Ubuntu 18.04, I ran into an oddity, in which
the call to gtk_widget_show(inst->window) in new_session_window() has
the side effect of delivering a spurious configure_event on the
drawing area with size 1x46 pixels. This causes the terminal to resize
itself to 1 column wide, and the mistake isn't rectified until a
followup configure-event arrives after new_session_window returns to
the GTK main loop. But that means terminal output can occur between
those two configure events (the connection-sharing "Reusing a shared
connection to host.name" is a good example), and when it does, it gets
embarrassingly wrapped at one character per line down the left column.
I briefly tried to bodge around this by trying to heuristically guess
which configure events were real and which were spurious, but I have
no faith in that strategy continuing to work. I think a better
approach is to abandon configure-event completely, and move to a
system in which the two purposes I was using it for are handled by two
_different_ GTK signals, namely "size-allocate" (for knowing when we
get resized) and "realize" (for knowing when the drawing area
physically exists for us to start setting up Cairo or GDK machinery).
The result seems to have fixed the silly one-column wrapping bug, and
retained the ability to handle window resizes, on every GTK version I
have conveniently available to test on, including GTK 3 both before
and after these spurious configures started to happen.
GTK 3 PuTTY/pterm has always assumed that if it was compiled with
_support_ for talking to the raw X11 layer underneath GTK and GDK,
then it was entitled to expect that raw X11 layer to exist at all
times, i.e. that GDK_DISPLAY_XDISPLAY would return a meaningful X
display that it could do useful things with. So if you ran it over the
GDK Wayland backend, it would immediately segfault.
Modern GTK applications need to cope with multiple GDK backends at run
time. It's fine for GTK PuTTY to _contain_ the code to find and use
underlying X11 primitives like the display and the X window id, but it
should be prepared to find that it's running on Wayland (or something
else again!) so those functions don't return anything useful - in
which case it should degrade gracefully to the subset of functionality
that can be accessed through backend-independent GTK calls.
Accordingly, I've centralised the use of GDK_DISPLAY_XDISPLAY into a
support function get_x_display() in gtkmisc.c, which starts by
checking that there actually is one first. All previous direct uses of
GDK_*_XDISPLAY now go via that function, and check the result for NULL
afterwards. (To save faffing about calling that function too many
times, I'm also caching the display pointer in more places, and
passing it as an extra argument to various subfunctions, mostly in
gtkfont.c.)
Similarly, the get_windowid() function that retrieves the window id to
put in the environment of pterm's child process has to be prepared for
there not to be a window id.
This isn't a complete fix for all Wayland-related problems. The other
one I'm currently aware of is that the default font is "server:fixed",
which is a bad default now that it won't be available on all backends.
And I expect that further problems will show up with more testing. But
it's a start.
Colin Watson reports that on pre-releases of Ubuntu 18.04, configure
events which don't actually involve a change of window size show up
annoyingly often. Our handling of configure events involves throwing
away the backing Cairo surface, making a fresh blank one, and
scheduling a top-level callback to get terminal.c to do a repaint and
populate the new surface; so a draw event before that callback occurs
causes the window contents to flicker off and on again, not to mention
wasting a lot of time.
The simplest solution is to spot spurious configures, and respond by
not throwing away the previous Cairo surface in the first place.
Except in GTK1 (which doesn't have the former), via a gtkcompat.h
workaround.
Up-to-date GTK3 has deprecated gdk_beep(), causing build failures due
to the default -Werror setting.
Looks as if I haven't retried the GTK1 build for a while, and recent
GTK frontend development has broken it. The selection revamp has
pointed out that GTK1 didn't have the accessor function
gtk_selection_data_get_selection(), the standard GdkAtom value
GDK_SELECTION_CLIPBOARD, or keysyms for alphabetic characters; and
also I had an initialisation of one of my own structure fields
(dp->selparams) accidentally not guarded by the same GTK-versioning
ifdef that controls whether or not it was defined.
This still isn't complete: I also need to add the variable collections
of things like mid-session special commands and saved session names,
and also I need to try to grey out menu items when they're not
applicable. But it's a start.
Just to avoid an endless proliferation of functions too small to see,
I've arranged an enumeration of action ids and a single
app_menu_action function on the receiving end, and in gtkapp.c, a list
macro that means I at least don't have to define the tiny callback
functions and the GActionEntry records by hand and keep them in sync.
This fixes the problem I'd previously noticed, that if you don't
configure the "Command key acts as Meta" setting, then keystrokes like
Command-Q which _ought_ to function as accelerators for the
application menu bar don't.
Turns out that this was for the totally obvious reason: the keyboard
event was still being processed by gtkwin.c's key_event() and
translated via the GTK IM into ordinary keyboard input. If instead I
return FALSE from key_event on detecting that a key event has a
non-Meta-configured Command modifier, then it will go to the next-
level key-event handler inside GTK itself which implements the menu
accelerator behaviour. Another problem ticked off the OS X checklist.
The gtkapp.c menu now has a Copy as well as Paste option; those menu
items, as well as the corresponding ones on the context menu and Copy
All, now address sets of clipboards parametrised between OS X and
ordinary GTK in unix.h. Also I've tweaked the wording of the
context-menu items to not use the X-specific terminology "CLIPBOARD"
on OS X.
I had a segfault on OS X today at Pterm.app shutdown. I wasn't able to
reproduce it in a debugger, but the cause seemed to be that
clipboard_clear called term_deselect (this was from before the patch
series that renamed that function) when inst->term was already NULL.
This must be because a clipboard_data_instance outlived its associated
inst->term, and quite likely its associated inst as well. But we can't
free those structures when a gui_data is freed, because GTK callbacks
will still depend on them; so instead we must have each gui_data keep
a list of active cdis pointing at it, and then at destruction time,
walk along the list nulling out each one's pointer to part of itself.
On all platforms, you can now configure which clipboard the mouse
pastes from, which clipboard Ctrl-Ins and Shift-Ins access, and which
Ctrl-Shift-C and Ctrl-Shift-V access. In each case, the options are:
- nothing at all
- a clipboard which is implicitly written by the act of mouse
selection (the PRIMARY selection on X, CLIP_LOCAL everywhere else)
- the standard clipboard written by explicit copy/paste UI actions
(CLIPBOARD on X, the unique system clipboard elsewhere).
Also, you can control whether selecting text with the mouse _also_
writes to the explicitly accessed clipboard.
The wording of the various messages changes between platforms, but the
basic UI shape is the same everywhere.
All the data fields referring to the selection in 'struct gui_data'
have been pulled out into a separate structure of which there are now
multiple instances, and I've plumbed through what should be the right
pointers and integer ids to everywhere they should go. So now the GTK
front end defines CLIP_PRIMARY and CLIP_CLIPBOARD in place of the
temporary cop-out CLIP_SYSTEM from the previous commit, and copying
and pasting can be done via either one.
The defaults should be the same as before, except that now the non-Mac
versions of the GtkApplication front ends will access CLIP_PRIMARY in
response to most actions but the 'Paste' menu item will paste from
CLIP_CLIPBOARD. (That's mostly just as a demonstration that accessing
multiple clipboards even works.)
This lays some groundwork for making PuTTY's cut and paste handling
more flexible in the area of which clipboard(s) it reads and writes,
if more than one is available on the system.
I've introduced a system of list macros which define an enumeration of
integer clipboard ids, some defined centrally in putty.h (at present
just a CLIP_NULL which never has any text in it, because that seems
like the sort of thing that will come in useful for configuring a
given copy or paste UI action to be ignored) and some defined per
platform. All the front end functions that copy and paste take a
clipboard id, and the Terminal structure is now configured at startup
to tell it which clipboard id it should paste from on a mouse click,
and which it should copy from on a selection.
However, I haven't actually added _real_ support for multiple X11
clipboards, in that the Unix front end supports a single CLIP_SYSTEM
regardless of whether it's in OS X or GTK mode. So this is currently a
NFC refactoring which does nothing but prepare the way for real
changes to come.
Previously, both the Unix and Windows front ends would respond to a
paste action by retrieving data from the system clipboard, converting
it appropriately, _storing_ it in a persistent dynamic data block
inside the front end, and then calling term_do_paste(term), which in
turn would call back to the front end via get_clip() to retrieve the
current contents of that stored data block.
But, as far as I can tell, this was a completely pointless mechanism,
because after a data block was written into this storage area, it
would be immediately used for exactly one paste, and then never
accessed again until the next paste action caused it to be freed and
replaced with a new chunk of pasted data.
So why on earth was it stored persistently at all, and why that
callback mechanism from frontend to terminal back to frontend to
retrieve it for the actual paste action? I have no idea. This change
removes the entire system and replaces it with the completely obvious
alternative: the character-set-converted version of paste data is
allocated in a _local_ variable in the frontend paste functions,
passed directly to term_do_paste which now takes (buffer,length)
parameters, and freed immediately afterwards. get_clip() is gone.
Now 'putty user@host' will do what you wanted on Unix the same way it
always has on Windows.
(Thanks to Geoff Winkless for pointing out this inconsistency. I've
redone his actual patch my way, but he should still be credited for
the inspiration!)
This one's in frontend_keypress(), which is supposed to close the
window on the first keypress after the session inside it terminates
(that is, if your close-on-exit settings haven't made it close already
at that point).
It looks to me as if that behaviour doesn't currently _work_, and
hasn't worked for quite a while (certainly it was broken as of 0.70,
well before I started on this weekend's refactoring), because when the
session terminates we delete inst->ldisc and that's what would
otherwise be calling frontend_keypress. I should probably decide what
to do about that at some point. But for the moment, I'm satisfied to
simply not break this functionality any worse by making it not a
process-global exit :-)
For gtkapp-based tools that will have to stop being a program-fatal
error, so I've turned it into a function called window_setup_error
(which I could in principle reuse for other problems in the long and
tortuous progress of new_session_window), and kept the original
handling in gtkmain.c's implementation of that function while gtkapp.c
does something more sensible with a message box.
Not all gtkwin-based tools use it. Only the ones with one session per
process, which parse a command line describing that session and might
reasonably want to report errors in that command line by writing to
standard error and exiting the program.
In other words, precisely the ones that link in gtkmain.c and not
gtkapp.c. So gtkmain.c is a more sensible place to put that
error-reporting function.
This was one of a handful of remaining places in gtkwin.c where exit()
is called incautiously. Of course, a failure to set up one SSH
connection should only be fatal to that connection, not the whole
process, so really we should be feeding into the connection_fatal
system.
This change requires me to break up the general cleanups in
delete_inst() into two halves: one runs when the error message box is
created, and cleans up the network connection and all the stuff
associated with it, and the other runs when the error message is
dismissed and the window can actually close.
I've also moved it out into gtkwin.c, because it seemed easier to do
the 'find existing instance of this dialog and raise it' dance there
than to split it across source files pointlessly.
Now it has several 'slots', each named for a particular class of
subsidiary dialog box that a session window can have at most one of,
and register_network_prompt_dialog has a more general name and takes
an enum-typed argument identifying a slot. This lets me avoid writing
a zillion annoyingly similar function pairs and corresponding snippets
of cleanup code in delete_inst.
If you close a session window with an associated SSH back end, the
back end may call back to notify_remote_exit() from ssh_free(), which
queues a new top-level callback citing the inst structure we were
about to delete.
We could fix this by introducing a special 'moribund' flag which
inhibits notify_remote_exit from queueing a callback, but far easier
is to move the delete_callbacks_for_context() call to _after_ all
subsidiary things have been cleaned up, so that any last-minute
callbacks they might schedule will be promptly unscheduled again
before they do any damage.
When I switch verify_ssh_host_key() and friends over to creating
non-modal message boxes and returning to the main loop, there will be
a risk that their parent window will need to close for some other
reason while the user hasn't answered the pending question yet. (E.g.
if the user presses the main session window's close button, which will
no longer be a prohibited UI action once the transient dialog is not
modal.)
At that point we need to get rid of the pending dialog box, both for
UI purposes (it would look silly and be confusing to leave it lying
around) and for memory management (if the user subsequently clicks OK
in such a dialog it would probably try to leave its result somewhere
stale).
So now there's a mechanism for gtkwin.c remembering what the current
'network prompt dialog' is, if any (in which category I intend to
include everything triggered from ssh.c's various reasons for asking
crypto-related questions), and cleaning it up when the struct gui_data
it belongs to goes away.
If a dialog box is destroyed by the program before the user has
pressed one of the result-delivering buttons - e.g. because the parent
window closes so the dialog is no longer relevant to anything anyway -
then dlgparam_destroy would never call the client code's provided
callback. That makes sense in terms of the callback wanting to _take
action_ based on the result of the dialog box, but it ignores the
possibility that the callback may simply need to free its own context
structure.
So now dlgparam_destroy always calls the client's callback, even if
the result it passes is negative (meaning 'the user never got round to
pressing any of the dialog-ending buttons'), and all the existing
client callbacks handle the negative-result case by doing nothing
except freeing any allocated memory they might have.
Now, in place of a variadic argument list with four parameters per
button and a terminating NULL, it takes a pointer to a struct which in
turn contains an (array,length) pair of small per-button structures.
In the process I've renamed the function from messagebox() to
message_box(). Partly that was just because it gave me a convenient
way to search the source for calls I hadn't converted yet, but also
I've thought for a while that that missing underscore didn't really
match the rest of my naming.
NFCI. Partly this minor refactor has the virtue that we can reuse the
more common button layouts without having to type them in at multiple
places in the code (and, indeed, I've provided buttons_yn and
buttons_ok for easy reuse, and could easily provide other things like
yesnocancel any time I need them). But mostly it's because I'm about
to split up message_box into multiple functions, and this saves me the
hassle of deciding which ones to make variadic and which to pass an
actual va_list to - particularly since messagebox() used to go over
its variadic argument list twice, which always makes delegating it to
another function that much more annoying.
Now every call to do_config_box is replaced with a call to
create_config_box, which returns immediately having constructed the
new GTK window object, and is passed a callback function which it will
arrange to be called when the dialog terminates (whether by OK or by
Cancel). That callback is now what triggers the construction of a
session window after 'Open' is pressed in the initial config box, or
the actual mid-session reconfiguration action after 'Apply' is pressed
in a Change Settings box.
We were already prepared to ignore the re-selection of 'Change
Settings' from the context menu of a window that already had a Change
Settings box open (and not accidentally create a second config box for
the same window); but now we do slightly better, by finding the
existing config box and un-minimising and raising it, in case the user
had forgotten it was there.
That's a useful featurelet, but not the main purpose of this change.
The mani point, of course, is that now the multi-window GtkApplication
based front ends now don't do anything confusing to the nesting of
gtk_main() when config boxes are involved. Whether you're changing the
settings of one (or more than one) of your already-running sessions,
preparing to start up a new PuTTY connection, or both at once, we stay
in the same top-level instance of gtk_main() and all sessions' top-
level callbacks continue to run sensibly.
This has been logically necessary in principle for ages, but we got
away without it because we just exited the program. But in the multi-
window GtkApplication front ends, we can't get away with that for
ever; we need to be able to free _one_ of our 'struct gui_data'
instances and everything dangling off it (or, at least, everything
that GTK's reference counting system doesn't clean up for us), without
also doing anything global to the process in which that gui_data is
contained.
ignore_sbar is a flag that we set while manually changing the
scrollbar settings, so that when those half-finished changes trigger
GTK event callbacks, we know to ignore them, and wait until we've
finished setting everything up before actually updating the window.
But somehow I had managed to leave the functions that actually _have
the effect_ (at least in GTK1) outside the pair of statements that set
and unset the ignore flag.
The effect was that compiling pterm for GTK1, starting it up, and
issuing a command like 'ls -l' that scrolls off the bottom of the
window would lead to the _top_ half of the ls output being visible,
and the scrollbar at the top of the scrollback rather than the bottom.
Apparently I haven't tested this compile mode in a while: I had a
couple of compile errors due to new code not properly #ifdeffed (the
true-colour mode has to be effectively disabled in the palette-based
GTK1 graphics model) and one for an unused static function
(get_monitor_geometry is only used in GTK2 and above, and with -Werror
that means I mustn't even _define_ it in GTK1).
With these changes, I still didn't get a clean compile unless I also
configured CFLAGS=-std=gnu89, due to the GTK1 headers having an
outdated set of ifdefs to figure out the compiler's semantics of
'inline'. (They seem to expect old-style gcc, which inconveniently
treats 'inline' and 'extern inline' more or less the opposite way
round from the version standardised by C99.)
ATTR_REVERSE was being handled in the front ends, and was causing the
foreground and background colours to be switched. (I'm not completely
sure why I made that design decision; it might be purely historical,
but then again, it might also be because reverse video is one effect
on the fg and bg colours that must still be performed even in unusual
frontend-specific situations like display-driven monochrome mode.)
This affected both explicit reverse video enabled using SGR 7, and
also the transient reverse video arising from mouse selection. Thanks
to Markus Gans for reporting the bug in the latter, which when I
investigated it turned out to affect the former as well.
I've done this on a 'where possible' basis: in Windows paletted mode
(in case anyone is still using an old enough graphics card to need
that!) I simply haven't bothered, and will completely ignore the dim
flag.
Markus Gans points out that some applications which (not at all
unreasonably) don't trust $TERM to tell them the full capabilities of
their terminal will use the sequence "OSC 4 ; nn ; ? BEL" to ask for
the colour-palette value in position nn, and they may not particularly
care _what_ the results are but they will use them to decide whether
the right number of colour palette entries even exist.
Otherwise, moving the cursor (at least in active, filled-cell mode) on
to a true-coloured character cell causes it to vanish completely
because the cell's colours override the thing that differentiates the
cursor.
I'm not sure if any X11 monochrome visuals or Windows paletted display
modes are still around, but just in case they are, we shouldn't
attempt true colour on either kind of display.
This is a heavily rewritten version of a patch originally by Lorenz
Diener; it was tidied up somewhat by Christian Brabandt, and then
tidied up more by me. The basic idea is to add to the termchar
structure a pair of small structs encoding 24-bit RGB values, each
with a flag indicating whether it's turned on; if it is, it overrides
any other specification of fg or bg colour for that character cell.
I've added a test line to colours.txt containing a few example colours
from /usr/share/X11/rgb.txt. In fact it makes quite a good demo to run
the whole of rgb.txt through this treatment, with a command such as
perl -pe 's!^\s*(\d+)\s+(\d+)\s+(\d+).*$!\e[38;2;$1;$2;$3m$&\e[m!' rgb.txt
Alamy Liu points out that asking for CONF_host will display the wrong
part of the configuration in the case where serial port setup fails.
The Windows front end's analogous message already got this right, but
I must have forgotten to change this one too when I introduced
conf_dest.
This seems to work around a GTK 3.22 display bug that Colin Watson and
I have both observed on Ubuntu (though I found that proxying the X
server, e.g. by SSH X forwarding or xtruss, inhibited the bug). The
effect of the bug was that the terminal window would appear completely
black and nothing would ever be displayed in it, though the terminal
session was still actually running and keystrokes would be sent to it.
But changing the call to cairo_set_source_surface() to some other
cairo_set_source_foo caused successful drawing of whatever other
source I selected; the problem seemed specific to the image surface.
Also, when I popped up the Ctrl-right-click menu over the terminal
window, the menu didn't disappear when dismissed, i.e. the drawing
area's redraw operation was not drawing in black, but failing to draw
_anything_.
That led me to hypothesise that the draw event handler for the
terminal drawing area might somehow be accidentally inventing 0 rather
than 255 for the implicit alpha channel when using our RGB-type image
surface as a source; so I tried setting the surface type to one with
an explicit alpha channel in the hope that there would no longer be a
need to make up any alpha value at all. And indeed, that seems to
solve the problem for me, so I might as well commit it.
However, I don't know the full details of what the previous problem
was, so this is only an empirical workaround. If it turns out I was
making some other mistake without which a RGB source surface would
have worked for me, then I should probably revert this and do whatever
other fix turns out to be a better plan.
Calling gtk_widget_realize to enforce the existence of an underlying
GdkWindow, followed by gdk_window_ensure_native to enforce an
underlying X window in turn, allows me to get hold of an X window id
on which I can call the Xlib function for setting WM_CLASS, still
before the window is mapped.
With this change, plus Colin's preceding patches, the whole code base
_actually_ compiles and links against GTK 3.22 without any deprecation
warnings. (My claim in commit 8ce237234 that it previously did appears
to have been completely wrong - my guess is that I'd forgotten to
'make clean' before testing against 3.22 and so some source files had
already been compiled against earlier GTK headers.)
GTK+ 3.22 deprecates gdk_screen_{width,height} on the grounds that the
"screen" here actually refers to a virtual screen that may span multiple
monitors, and applications should generally be considering the width and
height of individual monitors. It's not entirely clear to me how this
fits with X geometry specifications, but I've gone with trying to get
hold of the geometry of the monitor that the window in question is on.
gdk_window_set_background was already deprecated, but with GTK+ 3.22
even gdk_window_set_background_rgba is deprecated, so we need a better
approach. The best seems to be to go with the flow and inject a custom
CSS style for the appropriate widgets.
GTK+ 3.22 deprecates gtk_menu_popup in favour of various
gtk_menu_popup_at_* functions. gtk_menu_popup_at_pointer seems most
appropriate, but that requires being able to pass it a GdkEvent rather
than just some elements of it. In order to achieve that, I've
rearranged the scroll_event shim to construct a real GdkEventButton and
pass that down to button_internal.
Each gtkfont back end now provides a routine that will return the name
of a similar font to the current one but one notch larger or smaller.
For Pango, this is just a matter of incrementing the font size field
in a standard way; for X11 server-side fonts, we have to go and do an
XListFonts query with a wildcard that requests fonts that vary only in
the size fields from the current one, and then iterate over the result
looking for the best one.
(I expect this will be more useful to Pango scalable-font users than
to X11 fonts, but it seemed a shame not to give the X11 side my best
shot while I was at it.)
Choice of hotkey: I know I'm being inconsistent with gnome-terminal's
use of Ctrl-plus and Ctrl-minus. I thought that was because I was
already using Ctrl-minus as a more convenient synonym for
Ctrl-underscore (which sends the actual control code 0x1F), but now I
actually try it, apparently I'm not. However, Ctrl-plus and Ctrl-minus
are quite horrible as a keystroke pair anyway (one has to be typed
with shift and one without!), and I feel as if the 'less' and
'greater' signs are more specific anyway, in that they specifically
indicate _size_ rather than just 'unspecified numerical value'.
This is another widget that can appear in the top-level window, in
addition to the drawing area and scrollbar we put there ourselves, and
hence which needs to be accounted for when figuring out the
relationship between the drawing area size in character cells and the
full window size in pixels.
Finding the menu bar widget itself is a bit of a hassle, but having
found it, dealing with it is basically the same as dealing with the
scrollbar, only with x and y swapped.
This function, which parses the X11-style '-geometry WxH+X+Y' option
argument and automatically loads the result into the window, is also
being deprecated.
Fortunately we already had a fallback option for GTK1 (which didn't
have gtk_window_parse_geometry in the first place), calling the Xlib
geometry-parsing function and manually loading the results into GTK.
The method of loading into GTK is not the same between the two
versions, but the basic strategy is still viable.
For the sake of maintaining and testing fewer ifdef branches, I've
removed the use of gtk_window_parse_geometry _completely_, even in
GTK2 which did have it. GTK2 now uses the same strategy that I've
switched to for GTK3.
gtk_window_resize_to_geometry and gtk_window_set_default_geometry are
deprecated as of GTK 3.20, so now we do the geometry -> pixel size
conversion on our side.
This is preparation for dealing with the fact that GTK's geometry-
based API routines for setting the window size are being deprecated:
we'll no longer be able to specify a width/height in characters and
have GTK convert that into a pixel size based on the geometry hints
we'd already fed it. So we'll need to do that conversion ourselves,
and the easiest approach is to make it easy to recompute the geometry
hints on our side whenever we need them.
That function is deprecated as of 3.18, on the basis that GTK doesn't
need telling any more when the adjustment's owning widget needs
updating. So we just need to condition out the call.
Partly this is because the geometry_widget functionality is going away
in a later version of GTK3, so sooner or later we'll need not to be
using it anyway. But also, it turns out that GTK 3's geometry
calculations have the unfortunate effect of setting the window's base
and min heights to values that are not congruent mod height_increment
(because the former is the value we gave, but the latter is based on
the minimum height of the scrollbar), which confuses at least one
window manager (xfwm4) and causes the window to be created one row too
small.
So I've redone all the geometry computations my own way, based on the
knowledge that the only widgets visible in the top-level window are
the drawing area and the scrollbar and I know how both of those
behave, and taking care to keep base_height and min_height congruent
to avoid that xfwm4 bug.
This is a weird thing to have to do, but it is necessary: the OS X
PuTTY will need its top-level windows to be instances of a thing
called GtkApplicationWindow, rather than plain GtkWindow. Hence, the
actual creation of windows needs to be somewhere that isn't
centralised between the two kinds of front end.
This lays further groundwork for the OS X GTK3 port, which is going to
have to deal with multiple sessions sharing the same process. gtkwin.c
was a bit too monolithic for this, since it included some
process-global runtime state (timers, toplevel callbacks), some
process startup stuff (gtk_init, gtk_main, argv processing) and some
per-session-window stuff.
The per-session stuff remains in gtkwin.c, with the top-level function
now being new_session_window() taking a Conf. The new gtkmain.c
contains the outer skeleton of pt_main(), handling argv processing and
one-off startup stuff like setlocale; and the new gtkcomm.c contains
the pieces of PuTTY infrastructure like timers and uxsel that are
shared between multiple sessions rather than reinstantiated per
session, which have been rewritten to use global variables rather than
fields in 'inst' (since it's now clear to me that they'll have to
apply to all the insts in existence at once).
There are still some lurking assumptions of one-session-per-process,
e.g. the use of gtk_main_quit when a session finishes, and the fact
that the config box insists on running as a separate invocation of
gtk_main so that one session's preliminary config box can't coexist
with another session already active. But this should make it possible
to at least write an OS X app good enough to start testing with, even
if it doesn't get everything quite right yet.
This change is almost entirely rearranging existing code, so it
shouldn't be seriously destabilising. But two noticeable actual
changes have happened, both pleasantly simplifying:
Firstly, the global-variables rewrite of gtkcomm.c has allowed the
post_main edifice to become a great deal simpler. Most of its
complexity was about remembering what 'inst' it had to call back to,
and in fact the right answer is that it shouldn't be calling back to
one at all. So now the post_main() called by gtkdlg.c has become the
same function as the old inst_post_main() that actually did the work,
instead of the two having to be connected by a piece of ugly plumbing.
Secondly, a piece of code that's vanished completely in this
refactoring is the temporary blocking of SIGCHLD around most of the
session setup code. This turns out to have been introduced in 2002,
_before_ I switched to using the intra-process signal pipe strategy
for SIGCHLD handling in 2003. So I now expect that we should be robust
in any case against receiving SIGCHLD at an inconvenient moment, and
hence there's no need to block it.
If you run something like 'seq 2000000000' in a GTK3 pterm, the window
never actually updates, because pterm always considers reading more
data from the pty to have higher priority than delivering the "draw"
event. Using g_io_add_watch_full instead of g_io_add_watch allows us
to explicitly lower the priority of the I/O sources, so that window
redraws will take precedence.
This is quite a pain, since it involves inventing an entire new piece
of infrastructure to install a custom Xlib error handler and give it a
queue of things to do. But it fixes a bug in which Unix pterm/PuTTY
crash out at startup if one of the root window's CUT_BUFFERn
properties contains something of a type other than STRING - in
particular, UTF8_STRING is not unheard-of.
For example, run
xprop -root -format CUT_BUFFER3 8u -set CUT_BUFFER3 "thingy"
and then pterm without this fix would have crashed.
I've made the licence text, the About box, and the host key dialog
into GTK selectable edit controls. (The former because it contains a
lot of text; the About box because pasting version numbers into bug
reports is obviously useful; the host key because of the fingerprint.)
TOOLTYPE_NONNETWORK (i.e. pterm) already has "-log" (as does Unix
PuTTY), so there's no sense suppressing the synonym "-sessionlog".
Undocumented lacunae that remain:
plink accepts -sessionlog, but does nothing with it. Arguably it should.
puttytel accepts -sshlog/-sshrawlog (and happily logs e.g. Telnet
negotiation, as does PuTTY proper).
I had originally planned to implement a Compose-type key locally in
GTK PuTTY, as I did in Windows PuTTY. But in fact we've done this for
some time by delegating to the GTK IM system, which is a far better
idea anyway. So there's no point any more having the FIXME comment
that mentions Compose keys.
Also, there was a comment worrying about what I was going to do about
double-width characters in Pango, which is long since sorted out.
The top-level loop in gtkwin.c which draws text was expecting that the
right way to draw a printing character plus combining characters was
to overprint them one by one on top of each other. This is an OK
assumption for X bitmap fonts, but in Pango, it works very badly -
most obviously because asking Pango to display a combining char on its
own causes it to print a dotted circle representing the base char, but
also because surely there will be character combinations where Pango
wants to do something more sophisticated than just printing them each
at a standard offset, and it would be a shame not to let it.
So I've moved the previous overprinting loop down into the x11font
subclass of the unifont mechanism. The top-level gtkwin.c drawing code
now calls a new method unifont_draw_combining, which in the X11 case
does the same loop as before, but in the Pango case, just passes a
whole base+combinings string to Pango in one go and lets it do the
best job it can.
Touchpad gestures can generate much smoother scrolling events than the
discrete increments of mouse wheels. GDK3 supports this by means of a
new kind of scroll event, with direction GDK_SCROLL_SMOOTH and a
floating-point delta value. Added support for this; so in GTK3 mode,
you can now touchpad-scroll at a granularity of one line rather than
five, but in mouse tracking mode, scroll events are still grouped into
5-line chunks for purposes of turning them into escape sequences to
send to the server.
Apparently, if you don't specify GDK_SMOOTH_SCROLL_MASK in a widget's
event mask, then you don't receive "scroll_event" signals at all, even
of the non-smooth variety that was all GTK2 had. Hence, neither mouse
scroll wheels nor touchpad scroll gestures seem to generate any
response.
Adding GDK_SMOOTH_SCROLL_MASK brings the old scroll events back again,
so this is at least no worse than GTK2 was. But in GTK3 we _ought_ to
be able to do better still, by supporting smooth scrolling from
touchpads; this commit doesn't do that.
In GTK 3, it was impossible to resize the window smaller than it
started off, because the size request on the drawing area was taken as
a minimum. (I can't actually remember how the GTK 2 version doesn't
have this problem too.)
Fixed by instead setting the initial window size using
gtk_window_set_default_geometry() (having first set up the geometry
hints to reflect the character cell size).
In any DRAWTYPE_CAIRO mode, we now do all our Cairo drawing to a Cairo
image surface which lives on the client; then we either blit directly
from that to the window (if we're in GTK3 mode, or GTK2 without
deprecated pieces of API), or else we blit from the Cairo surface to
the server-side pixmap and then from there to the actual window.
In DRAWTYPE_GDK mode, nothing much has changed: we still draw directly
to the server-side pixmap using the GDK drawing API, then blit from
there to the window. But there is one change, namely that the blit is
no longer done proactively - instead, we queue a redraw of the
affected rectangle, and wait until we're called back by the expose
handler.
The main aim of all this is to arrange that the only time we ever draw
to the real window is in response to expose/draw events. The
experimental GTK3 OS X port stopped working a week or two ago (I
presume in response to an OS update) with the symptoms that attempts
to draw on the window outside the context of a "draw" event handler
just didn't seem to work any more; this change fixes it.
In addition to that benefit, this change also has obvious performance
advantages in principle. No more expensive text rendering in response
to an expose event - just re-copy to the window from the bitmap we
already have, from wherever it's stored that's nearest.
Moreover, this seems to have fixed the significant performance problem
with X server-side fonts under GTK. I didn't expect _that_! I'd
guessed that the approach of downloading character bitmaps and
rendering them manually via Cairo was just inherently slow for some
reason. I've no real idea _why_ this change improves matters so much;
my best guess is that perhaps there's an overhead in each drawing
operation to a GDK Cairo surface, so we win by doing lots of
operations to a much faster image surface and then batching them up
into one GDK Cairo operation. But whyever it is, I'm certainly not
complaining!
(In fact, it now seems to be noticeably _faster_, at least on my usual
local X displays, to draw server-side fonts using that technique than
using the old GDK approach. I may yet decide to switch over...)
When I introduced the KEY_EVENT_DIAGNOSTICS system last month in
commit 769600b22, I somehow didn't notice that it sat next to an
existing system of ifdefs labelled KEY_DEBUGGING, which did some
things worse but some things better.
Now I've expanded both of those into a fairly complete system of
diagnostics (keeping the newer name of KEY_EVENT_DIAGNOSTICS), and
made them use debug() rather than printf() so that in situations where
no standard output is available I can still retrieve the diagnostics
from debug.log.
I turned it into Shift-Return, because I was trying to find out why
the former didn't work on OS X and replaced it with something else
random to see if the code was even being reached. And then, like an
utter doofus, I committed that change as part of a50da0e30.
Rather than trying to get my existing hugely complicated X-style
clipboard code to somehow work with the Quartz GTK back end, I've
written an entirely new and much simpler alternative clipboard handler
usnig the higher-leve GtkClipboard interface. It assumes all clipboard
text can be converted to and from UTF-8 sensibly (which isn't a good
assumption on all front ends, but on OS X I think it's reasonable),
and it talks to GDK_SELECTION_CLIPBOARD rather than PRIMARY, which is
the only clipboard OS X has.
I had to do a fiddly thing to cope with the fact that each call to
gtk_clipboard_set_with_data caused a call to the clipboard clear
function left over from the previous set of data, so I had to avoid
mistaking that for a clipboard-clear for the _new_ data and
immediately deselecting it. I did that by allocating a distinct
placeholder object in memory for each instance of the copy operation,
so that I can tell whether a clipboard-clear is for the current copy
or a previous one.
This is only very basic support which demonstrates successful copying
and pasting is at least possible. For a sensible OS X implementation
we'll need a more believable means of generating a paste UI action
(it's quite easy to find a Mac on which neither Shift-Ins nor the
third mouse button even exists!). Also, after the trouble I had with
the clipboard-clear event, it's a bit annoying to find that it
_doesn't_ seem to get called when another application becomes the
clipboard owner. That may just be something we have to put up with, if
I can't find any reason why it's failing.
If I'm using Option as the Meta key, I want to suppress OS X GTK's
default behaviour of treating it as an AltGr-oid which changes the
keyval and Unicode translation of alphabetic keys. So on OS X I enable
a somewhat bodgy workaround which retranslates from the hardware
keycode as if the Option modifier had not been active at the time, and
use that as the character to prefix Esc to.
This is a bit nasty because I have to hardwire group = 0 in the call
to gdk_keymap_translate_keyboard_state(), whereas in principle what I
wanted was group = (whatever would have resulted from everything else
in the key event other than MOD1). However, in practice, they seem to
be the same, so this will do for the moment.
Personally I like using Command as the Esc-prefixing Meta key in
terminal sessions, because it occupies the same physical keyboard
position as the Alt key that I'm used to using on non-Macs. OS X
Terminal uses Option for that purpose (freeing up Command for the
conventional Mac keyboard shortcuts, of course), so I anticipate
differences of opinion.
Hence, here's a pair of OSX-specific config options which permit a
user to set either, or neither, or both of those modifier keys to
function as the terminal Meta key.
In a UTF-8 pterm, it makes sense to set the IUTF8 flag (on systems
that have one) on the pty device, so that line editing will take
account of UTF-8 multibyte characters.
Several utility functions I've written over the last few weeks were in
rather random places because I didn't have a central gtkmisc.c to put
them in. Now I've got one, put them there!
I'm using a slightly more up-to-date GTK version for testing on MacOS,
and it's marked a few more functions as deprecated, among which is
gdk_color_parse(). So now parsing -fg and -bg options has to be done
by two different calls and an ugly #ifdef, depending on GTK version.
If we're not supporting server-side fonts, it's utterly silly to set
one as the default! Instead, we use Pango's guarantee that some
reasonably sensible monospaced font will be made available under the
name "Monospace", and use that at a reasonable default size of 12pt.
Using GTK to run on OS X is going to require several workarounds and
behaviour tweaks to be enabled at various points in the code, and it's
already getting cumbersome to remember what they all are to put on the
command line. Here's a central #define (OSX_GTK) that enables them all
in one go, and a configure option (--with-quartz) that sets it.
As part of this commit, I've also rearranged the #include order in the
GTK source files, so that they include unix.h (which now might be
where NOT_X_WINDOWS gets defined) before they test NOT_X_WINDOWS to
decide whether to include X11 headers.
The Quartz GDK back end, if you press (say) Ctrl-A, will generate a
GdkKeyEvent with keyval='a' and state=CONTROL, but it'll have a
translated string of "a" where the X back end would have returned
"\001". So we have to do our own translation, which fortunately isn't
hard.
These should dump out all the important parts of the incoming
GdkEventKey, so that if keys aren't being translated right, it should
be possible to work out something about why not.
To enable: make CPPFLAGS="-DKEY_EVENT_DIAGNOSTICS"
