areas of the code. Not all back-ends have been tested, but Telnet and SSH
behave reasonably.
Incidentally, almost all of this patch was written through Mac PuTTY,
admittedly over a Telnet connection.
[originally from svn r2615]
just done this the very simple way - bundle all the globals into a
data structure and pass pointers around. One particularly ugly wart
is that wc_to_mb now takes a pointer to this structure as an
argument (optional, may be NULL, and unused in any Unicode layer
that's even marginally less of a mess than the Windows one). I do
need to do this properly at some point, but for now this should just
about be adequate. As usual, the Mac port has not been updated.
[originally from svn r2592]
and term_reconfig() now passes in a new structure which is copied
over the top. This means that the old and new structures can be
compared, and the _current_ as well as default states of auto wrap
mode, DEC origin mode, BCE, blinking text and character classes can
be conveniently reconfigured in mid-session without requiring a
terminal reset.
[originally from svn r2557]
initialise term->paste_len during initialisation, and indeed looking
at the code confirms this. I'm puzzled as to why valgrind didn't
spot this in pterm, though, since it's all in cross-platform code!
[originally from svn r2494]
right-hand half of a CJK wide character; correct handling of cut and
paste when CJK text wraps between lines _irrespective of the parity
of the starting column_; correct handling of wordness values
irrespective of which half of a CJK character the user
double-clicked on; correct handling when any terminal activity
overwrites only one half of a CJK wide character. I think we now
behave marginally better than xterm in this respect (it has a redraw
problem when you overwrite the RH half of a CJK char), so I'm happy.
Also redefined the internal UCSWIDE marker to something in the
surrogate range, while I'm here, so that U+303F is available for use
by actual users.
[originally from svn r2426]
_both_ halves of the character set, rather than flipping the two
halves. My source for this is linux/drivers/char/console.c.
[originally from svn r2394]
ignore when breaking text into runs for display, and implement setting this
on Mac (other ports just use 0xffffffff).
We don't use DeviceLoop for this any more because Apple Technical Q&A
QA1024 says we shouldn't. Unlike their example, we don't depend on the
Display Manager's being present either.
[originally from svn r2264]
we need to update dispcurs as well as disptext.
Add scroll optimisation to scroll(). This makes it rather obvious that scroll
optimisation is breaking with PuTTY's usual policy of lazy updates, since
scrolling is done eagerly. Fixing this so that all the scrolling is saved
up for do_paint would be nice, but non-trivial.
[originally from svn r2252]
This introduces a new front-end function, do_scroll(), which is expected to
scroll a part of the physical display and cause repaint events for any
areas that couldn't be scrolled (e.g. because they were hidden).
scroll_display() is a wrapper around this which also updates disptext to
match.
Currently, scroll_display is only used in response to user scrollback requests
(via term_scroll()), but extending scroll() to use it as well should be
easy.
All of this is conditional on the front end's defining OPTIMISE_SCROLL, since
only the Mac front end currently implements do_scroll().
[originally from svn r2242]
mode==BELL_VISUAL, otherwise taskbar flashing won't happen on visual
bells. It's up to the frontend routine to spot BELL_VISUAL and avoid
making any noise.
[originally from svn r2155]
absent, and also (I think) all the frontend request functions (such
as request_resize) take a context pointer, so that multiple windows
can be handled sensibly. I wouldn't swear to this, but I _think_
that only leaves the Unicode stuff as the last stubborn holdout.
[originally from svn r2147]
each backend now stores all its internal variables in a big struct,
and each backend function gets a pointer to this struct passed to
it. This still isn't the end of the work - lots of subsidiary things
still use globals, notably all the cipher and compressor modules and
the X11 forwarding authentication stuff. But ssh.c itself has now
been transformed, and that was the really painful bit, so from here
on it all ought to be a sequence of much smaller and simpler pieces
of work.
[originally from svn r2127]
terminal.c was apparently relying on implicit initialisation to
zero, and also I've removed the backends' dependency on terminal.h
by having terminal sizes explicitly passed in to back->size().
[originally from svn r2117]
all the global and function-static variables out of terminal.c into
a dynamically allocated data structure. Note that this does not yet
confer the ability to run more than one of them in the same process,
because other things (the line discipline, the back end) are still
global, and also in particular the address of the dynamically
allocated terminal-data structure is held in a global variable
`term'. But what I've got here represents a reasonable stopping
point at which to check things in. In _theory_ this should all still
work happily, on both Unix and Windows. In practice, who knows?
[originally from svn r2115]
of the alternate-screen and save-cursor control sequences, with
subtly different semantics and entertaining interactions with the
usual ones. No thanks to xterm for doing so in the first place :-(
This checkin should sort it all out.
[originally from svn r2112]
it's automatically deactivated by any keypress, so that command-line
beeps from (e.g.) filename completion don't suddenly stop occurring,
but it still provides a rapid response to an accidental spewing of a
binary to your terminal.
[originally from svn r2107]
The current pty.c backend is temporarily a loopback device for
terminal emulator testing, the display handling is only just enough
to show that terminal.c is functioning, the keyboard handling is
laughable, and most features are absent. Next step: bring output and
input up to a plausibly working state, and put a real pty on the
back to create a vaguely usable prototype. Oh, and a scrollbar would
be nice too.
In _theory_ the Windows builds should still work fine after this...
[originally from svn r2010]
beginning of a Unix port. It's nowhere near done, and currently it
won't even compile on Unix. But this represents the start of the
process of separating out platform-specific code, and also contains
the mkfiles.pl changes required to support a Unix makefile and a
non-flat source tree.
[originally from svn r1993]
mode ever failed to do this, and only Plink actually had a problem
with it, so this didn't become obvious for a while. rlogin mode is
fixed, and all implementations of from_backend() now contain an
assertion so that we should spot errors of this type more quickly in
future.
[originally from svn r1571]
sick of recompiling to enable packet dumps. SSH packet dumping is
now provided as a logging option, and dumps to putty.log like all
the other logging options. While I'm at it I cleaned up the format
so that packet types are translated into strings for easy browsing.
POSSIBLE SIDE EFFECT: in the course of this work I had to re-enable
the SSH1 packet length checks which it turns out hadn't actually
been active for some time, so it's possible things might break as a
result. If need be I can always disable those checks for the 0.52
release and think about it more carefully later.
[originally from svn r1493]
uses to manipulate the window (minimise, maximise, front, back,
move, resize) and report things about the window (is it minimised or
maximised, how big is it, what's its title). Missing are ESC[4;X;Yt
(resize to a specified pixel size; our resize code doesn't like it)
and ESC[19;X;Yt (report size of _screen_ in _characters_, which it
isn't even obvious how to do when you've got a variable font size).
[originally from svn r1414]
line break was created by wrapping. (Equivalently, if the selection
would _paste_ as a single word without a newline in the middle, then
it will _select_ in the same way.)
[originally from svn r1347]
9, then did ANSI Delete Line on line 10, the selection highlight
would move up a line even though it wasn't over any text that
actually moved. Easy to reproduce in the likes of vi. Trivial fix.
[originally from svn r1334]
wraparound, not referencing vbell_timeout if in_vbell==FALSE, that
sort of thing. I doubt it'll fix the reported problems with screen
vbells, since none of the failure modes I've just prevented looked
all that probable to me, but it's nice to have extra robustness
anyway.
[originally from svn r1314]
term_out() can in turn call ldisc_send() which calls back to
from_backend() when local echo is enabled. This was giving rise to
crazy re-entrancy stuff and stack overflows. Instead from_backend()
deposits its data in a bufchain which term_out() empties the next
time it's called.
[originally from svn r1276]
fiddle with the widths of characters in DBCS screen fonts, and (the
big one) one to enable a mode in which resizing the window locks the
terminal size and lets the font change, instead of vice versa. That
should shut up a few feature requests!
[originally from svn r1269]
characters that failed the UTF-8 canonicality rules were being sent
to the session log twice. Sounds trivial, but I bet it'd have
confused anyone who turned on session logging precisely to track
down a canonicality bug :-)
[originally from svn r1244]
mouse tracking is enabled. (This can be turned off if your app
really wants Shift+mouse, but it defaults to on for general
usefulness.)
[originally from svn r1235]
by ceasing to listen on input channels if the corresponding output
channel isn't accepting data. Has had basic check-I-didn't-actually-
break-anything-too-badly testing, but hasn't been genuinely tested
in stress conditions (because concocting stress conditions is non-
trivial).
[originally from svn r1198]
it out. A line was removed from the scrollback, cleared, and placed
at the bottom of the screen. Fine, except that the clearing process
assumed the line was the right length already, and thanks to lazy
resizing this wasn't necessarily the case. Segfaults and memory
corruption ensued.)
[originally from svn r1129]
possible and we have a single unified means of trying to display any
Unicode code point. Instead of the various ad-hoc translation modes
we had before, we now have a single `codepage' option which allows
us to treat the incoming (and outgoing) text as any given character
set, and locally we map that to Unicode and back.
[originally from svn r1110]
requires fix_cpos() to be called after it (otherwise cpos might point
to a line that isn't where you remember it being), and a mis-aimed
incpos() was causing forward selection dragging not to include the
char under the mouse. Both fixed.
[originally from svn r1063]
configurable bell overload handling. Thanks to Robert de Bath for
galvanising me into doing this, but I've had to rip most of his code
out and redo it myself...
[originally from svn r1039]
now to translate them into poor man's characters (+--+ and |). We also
have an option to disable this (and map line drawing characters to the
corresponding ASCII code as before). Thanks to Robert de Bath.
[originally from svn r1029]
multiple switchable line disciplines, we now have a single unified
one which changes its behaviour based on option settings. Each
option setting can be suggested by the back end and/or the terminal
handler, and can be forcibly overridden by the configuration. Local
echo and local line editing are separate, independently switchable,
options.
[originally from svn r895]
character if we were wrapping, not whether we _will_ wrap next
character. Makes for saner behaviour with vertical-line cursor and
also when changing autowrap mode while on rightmost column. Does
entail small behavioural changes to backspace and destructive-
backspace when in rightmost column with Auto Wrap off, but I don't
think they should be catastrophic, or indeed that there's a well
defined Right Behaviour.
[originally from svn r872]
advantages:
- protocol modules can call sk_write() without having to worry
about writes blocking, because blocking writes are handled in the
abstraction layer and retried later.
- `Lost connection while sending' is a thing of the past.
- <winsock.h> is no longer needed in most modules, because
"putty.h" doesn't have to declare `SOCKET' variables any more,
only the abstracted `Socket' type.
- select()-equivalent between multiple sockets will now be handled
sensibly, which opens the way for things like SSH port
forwarding.
[originally from svn r744]
- Robert de Bath's Compose key is now off by default and configurable on
- The ages-old controversy over whether ALT by itself should bring the
System menu up is now controllable by a config option
- You can now independently configure whether scrollback resets on a
keypress _and_ whether it resets on screen activity.
[originally from svn r741]
use when they have data from the network. Replaces the utterly daft
inbuf / inbuf_head / term_out() interface, which only made sense
when feeding to terminal.c. (terminal.c now implements
from_backend() as a small function that gateways to the old
interface.)
As a side effect, from_backend() also has an `is_stderr' parameter,
so scp can once again separate the server's pronouncements on stderr
from the actual protocol progress on stdout.
[originally from svn r729]
- cope with strange WinSock wrappers not supporting SIOCATMARK
- define yet more terminal compatibility modes
- support UK-ASCII (just like US-ASCII but # is a sterling sign)
- support connection keepalives at a configurable interval
[originally from svn r692]
ATTR_BLINK (as bold background) and VT52 support. Plus a
compatibility tracking system whereby all escape sequences can be
disabled for a pure-VT102 compatibility mode or other levels.
[originally from svn r421]
apparently defined behaviour for _all_ CSI-type sequences that ESC[Q
should be equivalent to ESC[0Q. Which is a pain in the wossname and
not a sane way to do it, but if the standards say it then I suppose
... :-(
[originally from svn r406]
- ^E answerback is now `PuTTY'.
- The framework is now in place for the scrollback to reset to
bottom on display _or_ keyboard events _or_ both. An actual
configurable option isn't yet present, but most of the code is in
place.
- Try to deal with the problems where incoming data gets dropped
after decoding but before display.
- Scrollback behaviour has changed: instead of keying it off
`scroll' versus `delete top line', things now go into the
scrollback from _either_ of those but only if the primary screen
is selected. Should fix problems with `less' and talkers.
- must_update variable has gone because rdb correctly observed that
it didn't seem to be doing a great deal :-)
[originally from svn r328]
- Stop using the identifier `environ' as some platforms make it a macro
- Fix silly error box at end of connection in FWHACK mode
- Fix GPF on maximise-then-restore
- Use SetCapture to allow drag-selecting outside the window
- Correctly update window title when iconic and in win_name_always mode
[originally from svn r12]
