When a winhandl.c input thread returns EOF to the main thread, the
latter might immediately delete the input thread's context. I
carefully wrote in a comment that in that case we had to not touch ctx
ever again after signalling to the main thread - but the test for
whether that was true, which also touched ctx, itself came _after_ the
SetEvent which sent that signal. Ahem.
Spotted by Minefield, which it looks as if I haven't run for a while.
I had set up an event object for signalling incoming connections to
the named pipe, and then called handle_add_foreign_event to get that
event object watched for connections - but when I closed down the
listening pipe, I deleted the event object without also cancelling
that foreign-event handle, so that winhandl.c would potentially call
the callback for a destroyed object.
This was an old bug, fixed around 0.59, which apparently regressed
when I rewrote the main event loop using the toplevel_callback
mechanism.
Investigation just now suggests that it has to do with my faulty
assumption that Windows PeekMessage would deliver messages in its
message queue in FIFO order (i.e. that the thing calling itself a
message queue is actually a _queue_). In fact my WM_NETEVENT seems to
like to jump the queue, so that once a steady stream of them starts
arriving, we never do anything else in the main event loop (except
deal with handles).
Worked around in a simple and slightly bodgy way, namely, we don't
stop looping on PeekMessage and run our toplevel callbacks until we've
either run out of messages completely or else seen at least one that
_isn't_ a WM_NETEVENT. That way we should reliably interleave NETEVENT
processing with processing of other stuff.
To understand the handle leak bug that I fixed in git commit
7549f2da40, I had to think fairly hard
to remind myself what all this code was doing, which means the
comments weren't good enough. Expanded and rewritten some of them in
the hope that things will be clearer next time.
If (say) a read handle returns EOF, and its gotdata function responds
by calling handle_free(), then we want the handle to have already had
its defunct flag set so that the handle can be destroyed. Otherwise
handle_free will set the 'done' flag to ask the subthread to
terminate, and then sit and wait for it to say it's done so -
forgetting that it signalled termination already by returning EOF, and
hence will not be responding to that signal.
Ditto for write errors on write handles, though that should happen
less often.
The code for cleaning up handle structures works by the main thread
asking the per-handle subthread to shut down by means of setting its
'done' flag, and then once the subthread signals back through its
event object that it's done so, the main thread frees all its
resources and removes the event object from the list of things being
checked in the program's event loop.
But read threads were not sending back that final event acknowledging
a request to shut down, so their event objects were never being
cleaned up.
Bug spotted by Ronald Weiss.
We were checking the return value of CreateThread for validity, but
not keeping it to free afterwards if it _was_ valid. Also, we weren't
closing ctx->event in the valid case either. Patch due to Tim Kosse.
I don't think anyone has ever actually called it that, colloquially
_or_ formally, and if anyone ever did (in a bug report, say) I'd
probably have to stop and think to work out what they meant. It's
universally called Plink, and should be officially so as well :-)
I'm not actually sure why we've always had back ends notify ldisc of
changes to echo/edit settings by giving ldisc_send(ldisc,NULL,0,0) a
special meaning, instead of by having a separate dedicated notify
function with its own prototype and parameter set. Coverity's recent
observation that the two kinds of call don't even have the same
requirements on the ldisc (particularly, whether ldisc->term can be
NULL) makes me realise that it's really high time I separated the two
conceptually different operations into actually different functions.
While I'm here, I've renamed the confusing ldisc_update() function
which that special operation ends up feeding to, because it's not
actually a function applying to an ldisc - it applies to a front end.
So ldisc_send(ldisc,NULL,0,0) is now ldisc_echoedit_update(ldisc), and
that in turn figures out the current echo/edit settings before passing
them on to frontend_echoedit_update(). I think that should be clearer.
If (Msg)WaitForMultipleObjects returns WAIT_TIMEOUT, we expect 'next'
to have been initialised. This can occur without having called
run_timers(), if a toplevel callback was pending, so we can't expect
run_timers to have reliably initialised 'next'.
I'm not actually convinced this could have come up in either of the
affected programs (Windows PSFTP and Plink), due to the list of things
toplevel callbacks are currently used for, but it certainly wants
fixing anyway for the future.
Spotted by Coverity.
Pageant's list box needs its tab stops reorganised a little for new
tendencies in string length, and also has to cope with there only
being one prefix space in the output of the new string fingerprint
function. PuTTYgen needs to squash more radio buttons on to one line.
This provides support for ECDSA public keys, for both hosts and users,
and also ECDH key exchange. Supported curves are currently just the
three NIST curves required by RFC 5656.
The OpenSSH key importer and exporter were structured in the
assumption that the strong commonality of format between OpenSSH RSA
and DSA keys would persist across all key types. Moved code around so
it's now clear that this is a peculiarity of those _particular_ two
key types which will not apply to others we add alongside them.
Also, a boolean 'is_dsa' in winpgen.c has been converted into a more
sensible key type enumeration, and the individually typed key pointers
have been piled on top of each other in a union.
This is a pure refactoring change which should have no functional
effect.
I've shifted away from using the SVN revision number as a monotonic
version identifier (replacing it in the Windows version resource with
a count of days since an arbitrary epoch), and I've removed all uses
of SVN keyword expansion (replacing them with version information
written out by Buildscr).
While I'm at it, I've done a major rewrite of the affected code which
centralises all the computation of the assorted version numbers and
strings into Buildscr, so that they're all more or less alongside each
other rather than scattered across multiple source files.
I've also retired the MD5-based manifest file system. A long time ago,
it seemed like a good idea to arrange that binaries of PuTTY would
automatically cease to identify themselves as a particular upstream
version number if any changes were made to the source code, so that if
someone made a local tweak and distributed the result then I wouldn't
get blamed for the results. Since then I've decided the whole idea is
more trouble than it's worth, so now distribution tarballs will have
version information baked in and people can just cope with that.
[originally from svn r10262]
The winegcc hack I use for my Coverity builds is currently using a
version of wincrypt.h that's missing a couple of constants I use.
Ensure they're defined by hand, but (just in case I defined them
_wrong_) also provide a command-line define so I can do that only in
the case of Coverity builds.
[originally from svn r10234]
This option is available from the command line as '-hostkey', and is
also configurable through the GUI. When enabled, it completely
replaces all of the automated host key management: the server's host
key will be checked against the manually configured list, and the
connection will be allowed or disconnected on that basis, and the host
key store in the registry will not be either consulted or updated.
The main aim is to provide a means of automatically running Plink,
PSCP or PSFTP deep inside Windows services where HKEY_CURRENT_USER
isn't available to have stored the right host key in. But it also
permits you to specify a list of multiple host keys, which means a
second use case for the same mechanism will probably be round-robin
DNS names that select one of several servers with different host keys.
Host keys can be specified as the standard MD5 fingerprint or as an
SSH-2 base64 blob, and are canonicalised on input. (The base64 blob is
more unwieldy, especially with Windows command-line length limits, but
provides a means of specifying the _whole_ public key in case you
don't trust MD5. I haven't bothered to provide an analogous mechanism
for SSH-1, on the basis that anyone worrying about MD5 should have
stopped using SSH-1 already!)
[originally from svn r10220]
A user points out that the person who writes a REG_SZ into the
registry can choose whether or not to NUL-terminate it properly, and
if they don't, RegQueryValueEx will retrieve it without the NUL. So if
someone does that to PuTTY's saved session data, then PuTTY may
retrieve nonsense strings.
Arguably this is the fault of whoever tampered with the saved session
data without doing it the same way we would have, but even so, there
ought to be some handling at our end other than silently returning the
wrong data, and putting the NUL back on seems more sensible than
complaining loudly.
[originally from svn r10215]
The IDM_RECONF handler unconditionally calls ldisc_configure to
reconfigure the line discipline for the new echo/edit settings, but in
fact ldisc can be NULL if no session is currently active. (Indeed, the
very next line acknowledges this, by testing it for NULL before
calling ldisc_send!) Thanks to Alexander Wong for the report.
[originally from svn r10214]
We now expect that after the server has sent us CHANNEL_CLOSE, we
should not expect to see any replies to our outstanding channel
requests, and conversely after we have sent CHANNEL_CLOSE we avoid
sending any reply to channel requests from the server. This was the
consensus among implementors discussing the problem on ietf-ssh in
April 2014.
To cope with current OpenSSH's (and perhaps other servers we don't
know about yet) willingness to send request replies after
CHANNEL_CLOSE, I introduce a bug-compatibility flag which is detected
for every OpenSSH version up to and including the current 6.6 - but
not beyond, since https://bugzilla.mindrot.org/show_bug.cgi?id=1818
promises that 6.7 will also implement the new consensus behaviour.
[originally from svn r10200]
Philippe Maupertuis reports that on one particular machine, Windows
causes the named pipe created by upstream PuTTY to be owned by the
Administrators group SID rather than the user's SID, which defeats the
security check in the downstream PuTTY. No other machine has been
reported to do this, but nonetheless it's clearly a thing that can
sometimes happen, so we now work around it by specifying explicitly in
the security descriptor for the pipe that its owner should be the user
SID rather than any other SID we might have the right to use.
[originally from svn r10188]
winshare.c includes ssh.h, but if you defined NO_SECURITY it then
decides to fall back to including the stub noshare.c, which includes
ssh.h again. Fix by moving a block of includes inside the ifdef.
[originally from svn r10184]
On Windows (X mouse reporting of the mouse wheel isn't currently done
by the Unix front end, though I'm shortly about to fix that too) a
mouse wheel event is translated into a virtual button, and we send
both a press and a release of that button to terminal.c, which encodes
both in X mouse reporting escape sequences and passes them on to the
server. This isn't consistent with what xterm does - scroll-wheel
events are encoded _like_ button presses, but differ semantically in
that they don't have matching releases. So we're updating to match
xterm.
[originally from svn r10138]
There's been a long-standing FIXME in Windows's sk_newlistener which
says that in IPv6 mode, an explicit source address (e.g. from a
command-line option of the form -L srcaddr:12345:dest:22) is ignored.
Now it's honoured if possible.
[originally from svn r10122]
Both GUI PuTTY front ends have a piece of logic whereby a string is
interpreted as host:port if there's _one_ colon in it, but if there's
more than one colon then it's assumed to be an IPv6 literal with no
trailing port number. This permits the PuTTY command line to take
strings such as 'host', 'host:22' or '[::1]:22', but also cope with a
bare v6 literal such as '::1'.
This logic is also required in the two Plink front ends and in the
processing of CONF_loghost for host key indexing in ssh.c, but was
missing in all those places. Add it.
[originally from svn r10121]
I've gone through everywhere we handle host names / addresses (on
command lines, in PuTTY config, in port forwarding, in X display
names, in host key storage...) and tried to make them handle IPv6
literals sensibly, by using the host_str* functions I introduced in my
previous commit. Generally it's now OK to use a bracketed IPv6 literal
anywhere a hostname might have been valid; in a few cases where no
ambiguity exists (e.g. no :port suffix is permitted anyway)
unbracketed IPv6 literals are also acceptable.
[originally from svn r10120]
Mike Edenfield points out that modern versions of the Windows SDK have
decided that 'INPUT' is a sensible name for an OS data structure
(sigh), and provided a patch to add a disambiguating prefix to
winhandl.c's enum values INPUT, OUTPUT and FOREIGN.
[originally from svn r10109]
Daniel Meidlinger reports that at least one Windows machine which is
not obviously otherwise misconfigured will respond to our
SetEntriesInAcl call with odd errors like ERROR_NONE_MAPPED or
ERROR_TRUSTED_RELATIONSHIP_FAILURE. This is apparently to do with
failure to convert the names "EVERYONE" and "CURRENT_USER" used in the
ACL specification to SIDs. (Or perhaps only one of them is the problem
- I didn't investigate in that direction.)
If we instead construct a fully SID-based ACL, using the well-known
world SID in place of EVERYONE and calling our existing get_user_sid
routine in place of CURRENT_USER, he reports that the problem goes
away, so let's do that instead.
While I'm here, I've slightly simplified the function prototype of
make_private_security_descriptor(), by turning 'networksid' into an
internal static that we can reuse in subsequent calls once we've set
it up. (Mostly because I didn't fancy adding another two pointless
parameters at every call site for the two new SIDs.)
[originally from svn r10096]
This will be useful if someone gets a mysterious Windows error on a
system configured into a language we don't speak - if they cut and
paste the error message to send to us, then we won't have to try to
translate it.
[originally from svn r10092]
XP doesn't have it, and I think having connection sharing work without
its privacy enhancement is better than having it not work at all.
[originally from svn r10087]
The basic strategy is described at the top of the new source file
sshshare.c. In very brief: an 'upstream' PuTTY opens a Unix-domain
socket or Windows named pipe, and listens for connections from other
PuTTYs wanting to run sessions on the same server. The protocol spoken
down that socket/pipe is essentially the bare ssh-connection protocol,
using a trivial binary packet protocol with no encryption, and the
upstream has to do some fiddly transformations that I've been
referring to as 'channel-number NAT' to avoid resource clashes between
the sessions it's managing.
This is quite different from OpenSSH's approach of using the Unix-
domain socket as a means of passing file descriptors around; the main
reason for that is that fd-passing is Unix-specific but this system
has to work on Windows too. However, there are additional advantages,
such as making it easy for each downstream PuTTY to run its own
independent set of port and X11 forwardings (though the method for
making the latter work is quite painful).
Sharing is off by default, but configuration is intended to be very
easy in the normal case - just tick one box in the SSH config panel
and everything else happens automatically.
[originally from svn r10083]
There's now a winsecur.[ch], which centralises helper functions using
the Windows security stuff in advapi.h (currently just get_user_sid),
and also centralises the run-time loading of those functions and
checking they're all there.
[originally from svn r10082]
It was only actually used in X11 and port forwarding, to find internal
state structures given only the Socket that ssh.c held. So now that
that lookup has been reworked to be the sensible way round,
private_ptr is no longer used for anything and can be removed.
[originally from svn r10075]
This commit adds two new support modules, winnpc.c and winnps.c, which
deal respectively with being a client and server of a Windows named
pipe (which, in spite of what Unix programmers will infer from that
name, is actually closer to Windows's analogue of a Unix-domain
socket). Each one provides a fully featured Socket wrapper around the
hairy Windows named pipe API, so that the rest of the code base should
be able to use these interchangeably with ordinary sockets and hardly
notice the difference.
As part of this work, I've introduced a mechanism in winhandl.c to
permit it to store handles of event objects on behalf of other Windows
support modules and deal with passing them to applications' main event
loops as necessary. (Perhaps it would have been cleaner to split
winhandl.c into an event-object tracking layer analogous to uxsel, and
the handle management which is winhandl.c's proper job, but this is
less disruptive for the present.)
[originally from svn r10069]
The mechanism for constructing a new connection-type Socket when a
listening one receives an incoming connection previously worked by
passing a platform-specific 'OSSocket' type to the plug_accepting
function, which would then call sk_register to wrap it with a proper
Socket instance. This is less flexible than ideal, because it presumes
that only one kind of OS object might ever need to be turned into a
Socket. So I've replaced OSSocket throughout the code base with a pair
of parameters consisting of a function pointer and a context such that
passing the latter to the former returns the appropriate Socket; this
will permit different classes of listening Socket to pass different
function pointers.
In deference to the reality that OSSockets tend to be small integers
or pointer-sized OS handles, I've made the context parameter an
int/pointer union that can hold either of those directly, rather than
the usual approach of making it a plain 'void *' and requiring a
context structure to be dynamically allocated every time.
[originally from svn r10068]
That's been a FIXME in the code for ages, because it's difficult to
get winhandl.c to stop an already-started read from a handle (since
the read is a blocking system call running in a separate thread). But
I now realise it isn't absolutely necessary to do so - you can just
buffer one lot of data from winhandl and _then_ tell it to stop.
[originally from svn r10067]
It's now kept in a separate module, where it can be reused
conveniently for other kinds of Windows HANDLE that I want to wrap in
the PuTTY Socket abstraction - for example, the named pipes that I
shortly plan to use for the Windows side of connection-sharing IPC.
[originally from svn r10066]
This restores PuTTY's backward compatibility to versions of Windows
too old to have ToUnicodeEx in their system libraries, which was
accidentally broken in 0.63.
[originally from svn r10061]
A couple of users report that my recent reworking of the Windows
top-level message loop has led to messages occasionally being lost,
and MsgWaitForMultipleObjects blocking when it ought to have been
called with a zero timeout. I haven't been able to reproduce this
myself, but according to one reporter, PeekMessage(PM_NOREMOVE) is
effective at checking for a non-empty message queue in a way that
GetQueueStatus is not. Switch to using that instead. Thanks to Eric
Flumerfelt for debugging and testing help.
[originally from svn r10057]
Jochen Erwied points out that once you've used PeekMessage to remove
_one_ message from the message queue, MsgWaitForMultipleObjects will
consider the whole queue to have been 'read', or at least looked at
and deemed uninteresting, and so it will block until a further message
comes in. Hence, my change in r10040 which stops us from looping on
PeekMessage until the queue is empty has the effect of causing the
rest of the message queue not to be acted on until a new message comes
in to unblock it. Fix by checking if the queue is nonempty in advance
of calling MsgWaitForMultipleObjects, and if so, giving it a zero
timeout just as we do if there's a pending toplevel callback.
[originally from svn r10052]
[r10040 == 5c4ce2fadf]
Martin Prikryl helpfully points out that when I revamped the socket
error mechanism using toplevel callbacks, I also accidentally passed
the error code to the wrong function. Use winsock_error_string instead.
[originally from svn r10048]
Unix GUI programs should not say 'Fatal Error' in the message box
title, and Plink should not destroy its logging context as a side
effect of printing a non-fatal error. Both appear to have been due to
inattentive cut and paste from the pre-existing fatal error functions.
[originally from svn r10044]
This change attempts to reinstate as a universal property something
which was sporadically true of the ad-hockery that came before
toplevel callbacks: that if there's a _very long_ queue of things to
be done through the callback mechanism, the doing of them will be
interleaved with re-checks of other event sources, which might (e.g.)
cause a flag to be set which makes the next callback decide not to do
anything after all.
[originally from svn r10040]
It was one of those things that went in ages ago on Windows and never
got replicated in the Unix front end. And it needn't be: ldisc.c is a
perfect place to put it, since it knows which of the data it's sending
is based on a keystroke and which is automatically generated, and it
also has access to the terminal context. So now a keypress can
interrupt a runaway paste on all platforms.
[originally from svn r10025]
Again, I've removed the special-purpose ad-hockery from the assorted
front end message loops that dealt with deferred handling of socket
errors, and instead uxnet.c and winnet.c arrange that for themselves
by calling the new general top-level callback mechanism.
[originally from svn r10023]
Instead of setting a must_close_session flag and having special code
in the message loop to check it, we'll schedule the call to
close_session using the new top-level callback system.
[originally from svn r10021]
I've removed the ad-hoc front-end bodgery in the Windows and GTK ports
to arrange for term_paste to be called at the right moments, and
instead, terminal.c itself deals with knowing when to send the next
chunk of pasted data using a combination of timers and the new
top-level callback mechanism.
As a happy side effect, it's now all in one place so I can actually
understand what it's doing! It turns out that what all that confusing
code was up to is: send a line of pasted data, and delay sending the
next line until either a CR or LF is returned from the server
(typically indicating that the pasted text has been received and
echoed) or 450ms elapse, whichever comes first.
[originally from svn r10020]
This is a little like schedule_timer, in that the callback you provide
will be run from the top-level message loop of whatever application
you're in; but unlike the timer mechanism, it will happen
_immediately_.
The aim is to provide a general way to avoid re-entrance of code, in
cases where just _doing_ the thing you want done is liable to trigger
a confusing recursive call to the function in which you came to the
decision to do it; instead, you just request a top-level callback at
the message loop's earliest convenience, and do it then.
[originally from svn r10019]
PuTTY does not trim a colon suffix off the hostname if it contains
_more than one_ colon. This allows IPv6 literals to be entered.
(Really we need to do a much bigger revamp of all uses of hostnames to
arrange that square-bracketed IPv6 literals work consistently, but
this at least removes a regression over 0.62.)
[originally from svn r9983]
[r9214 == a1f3b7a358]
palette_set() to be bogus. Fortunately, this isn't exploitable through
the terminal emulator, because the palette escape sequence parser
contains its own bounds check before even calling palette_set().
While I'm at it, fix the same goof in the OS X version! That port is
more or less abandoned, but that's no excuse for leaving obviously
wrong code lying around.
[originally from svn r9965]
support: transform_jumplist_registry should give its caller
dynamically allocated data if and only if it returns JUMPLISTREG_OK,
and get_jumplist_registry_entries should test the return value against
JUMPLISTREG_OK rather than a value from a totally different enum.
[originally from svn r9960]
The most interesting one is printer_add_enum, which I've modified to
take a char ** rather than a char * so that it can both realloc its
input buffer _and_ return NULL to indicate error.
[originally from svn r9959]
(This has also required me to add a currently unused nonfatal() to
PuTTYgen, since although PuTTYgen won't actually try to delete
putty.rnd, it does link in winstore.c as a whole.)
[originally from svn r9957]
strerror as I can arrange, wrapping up all the ugly FormatMessage
nonsense and caching previously looked-up messages for reuse so that
callers can treat them as static.
[originally from svn r9956]
ToAsciiEx, where possible.
This enables support for keys which generate Unicode characters that
aren't in the system code page, which seems to me like a perverse way
for Windows to have set up the system code page but apparently does
happen, e.g. (I'm told) U+0219 and U+021B on Romanian keyboards.
Patch mostly due to Andrei Damian-Fekete.
[originally from svn r9942]
gather extra entropy at Windows PuTTY startup time. (It's only used as
one of the inputs to PuTTY's internal entropy pool, so nobody is
required to trust it.)
[originally from svn r9941]
that the user really ought to know but that are not actually fatal to
continued operation of PuTTY or a single network connection.
[originally from svn r9932]
of the GET_32BIT macros and then used as length fields. Missing bounds
checks against zero have been added, and also I've introduced a helper
function toint() which casts from unsigned to int in such a way as to
avoid C undefined behaviour, since I'm not sure I trust compilers any
more to do the obviously sensible thing.
[originally from svn r9918]
character set configuration to UTF-8, on both Windows and Unix, and
reorganise the dropdown lists in the Translation menu so that UTF-8
appears at the top (and Unix's odd "use font encoding" is relegated to
the bottom of the list like the special-purpose oddity it is).
[originally from svn r9843]
use 32-bit scrollbar position data instead of being limited to the
16-bit version that comes in scrollbar messages' wParam.
[originally from svn r9720]
capture the error code if listen() returned an error, and instead pass
0 (saved from the previous successful bind) to winsock_error_string.
[originally from svn r9708]
FormatMessage to get the OS's text for any error not in our own
translation table. Should eliminate the frustrating 'unknown error'.
(I haven't chosen to use FormatMessage unconditionally, because it
comes out with enormous messages along the lines of "No connection
could be made because the target machine actively refused it" in place
of "Connection refused" and I'm Unixy enough to prefer the latter.
Also, on older Windowses, Winsock error codes are in a separate API
segment and don't work with FormatMessage anyway.)
[originally from svn r9704]
IPv6 addresses, because I'd mistakenly cast an ai_addr to the low-
level 'struct in6_addr' instead of the correct 'struct sockaddr_in6'.
[originally from svn r9690]
localhost connections, and also enable X forwarding in such a way that
it will attempt to connect to a Unix-domain X server socket, an
assertion will fail when proxy_for_destination() tries to call
sk_getaddr(). Fix by ensuring that Unix-domain sockets are _never_
proxied, since they fundamentally can't be.
[originally from svn r9688]
mistakenly rearranged the logic in an if statement in window.c, with
the effect that scroll-wheel events are no longer sent via xterm mouse
tracking. Put it back to the way it was.
[originally from svn r9679]
[r9214 == a1f3b7a358]
Well, at least across all command-line tools on both Windows and Unix,
and the GTK apps on Unix too. The Windows GUI apps fundamentally can't
write to standard output and it doesn't seem sensible to use message
boxes for these purposes :-)
[originally from svn r9673]
First, make absolute times unsigned. This means that it's safe to
depend on their overflow behaviour (which is undefined for signed
integers). This requires a little extra care in handling comparisons,
but I think I've correctly adjusted them all.
Second, functions registered with schedule_timer() are guaranteed to be
called with precisely the time that was returned by schedule_timer().
Thus, it's only necessary to check these values for equality rather than
doing risky range checks, so do that.
The timing code still does lots that's undefined, unnecessary, or just
wrong, but this is a good start.
[originally from svn r9667]
winadj@putty.projects.tartarus.org request. Not currently enabled
automatically, but should be usable as a manual workaround.
[originally from svn r9592]
zero but does it in such a way that over-clever compilers hopefully
won't helpfully optimise the call away if you do it just before
freeing something or letting it go out of scope. Use this for
(hopefully) every memset whose job is to destroy sensitive data that
might otherwise be left lying around in the process's memory.
[originally from svn r9586]
The previous platform-dependent ifdefs, switching between a system
which tried to cope with spurious callbacks (which I'd observed on
Windows) and one which tried to cope with system clock jumps (which
can happen on Unix, if you use gettimeofday) have been completely
removed, and replaced with a much simpler approach which just copes
with system clock jumps by triggering any timers immediately.
None of the resulting effects should be catastrophic (the worst thing
might be the waste of CPU in a spurious rekey, but as long as the
system clock isn't jumping around _all_ the time that's hardly
critical) and in any case the Unix port has had a long-standing oddity
involving occasional lockups if pterm or PuTTY runs for too long,
which hopefully this should replace with a much less bad failure mode.
And the code is much simpler, which is not to be sneezed at.
[originally from svn r9528]
sequence: since init_fonts sets up ucsdata based on the available
Windows fonts, we should call it before passing ucsdata to term_init.
[originally from svn r9527]
the offset horizontal line characters in the VT100 line-drawing set
(o,p,r,s), so that no trace of it - and hence no pointless performance
hit - is compiled into the cross-platform modules on non-Windows
platforms.
[originally from svn r9467]
UTF-16 support. High Unicode characters in the terminal are now
converted back into surrogates during copy and draw operations, and
the Windows drawing code takes account of that when splitting up the
UTF-16 string for display. Meanwhile, accidental uses of wchar_t have
been replaced with 32-bit integers in parts of the cross-platform code
which were expecting not to have to deal with UTF-16.
[originally from svn r9409]
so we should ensure we treat it the same way as other WM_SIZEs that
show up during that time: set the width and height in conf, and set
the flag to have that width and height enacted on WM_EXITSIZEMOVE.
Fixes a bug in which dragging a PuTTY window directly from the Win7
snapped-to-half-screen position to the snapped-to-maximised state
would leave the terminal in the pre-snapped size.
[originally from svn r9404]
Pageant's IPC mechanism. It's incomplete (he sent a much more
comprehensive set of fixes that I haven't reviewed), but should be
adequate to mitigate a particular issue for Bazaar users.
[originally from svn r9355]
GUIs of Pageant and PuTTYgen. With that and the prompts_t redesign,
there should no longer be any limit on passphrase length other than
the patience of the user.
[originally from svn r9320]
allocated type.
The main reason for this is to stop it from taking up a fixed large
amount of space in every 'struct value' subunion in conf.c, although
that makes little difference so far because Filename is still doing
the same thing (and is therefore next on my list). However, the
removal of its arbitrary length limit is not to be sneezed at.
[originally from svn r9314]
data channels. Should comprehensively fix 'half-closed', in principle,
though it's a big and complicated change and so there's a good chance
I've made at least one mistake somewhere.
All connections should now be rigorous about propagating end-of-file
(or end-of-data-stream, or socket shutdown, or whatever) independently
in both directions, except in frontends with no mechanism for sending
explicit EOF (e.g. interactive terminal windows) or backends which are
basically always used for interactive sessions so it's unlikely that
an application would be depending on independent EOF (telnet, rlogin).
EOF should now never accidentally be sent while there's still buffered
data to go out before it. (May help fix 'portfwd-corrupt', and also I
noticed recently that the ssh main session channel can accidentally
have MSG_EOF sent before the output bufchain is clear, leading to
embarrassment when it subsequently does send the output).
[originally from svn r9279]
policies before and after r9178, and hence able to talk to both
0.60-like and 0.61-like clients.
I had failed to consider that many pieces of code derived from PuTTY
would have imported the Pageant client code, so we shouldn't randomly
stop supporting things just because _we_ aren't using them any more.
[originally from svn r9264]
[r9178 == af78191a9c]
I think I have to consider this to be a separate but related change to
the wishlist item 'pscp-filemodes'; that was written before the Unix
port existed, and referred to the ability to configure the permissions
used for files copied from Windows to Unix - which is still not done.
[originally from svn r9260]
(In an embarrassingly silly way, too. No end of difficult stuff about
Conf serialisation done with great care and working just fine, and
then a trivial goof in using sscanf lets the whole lot down.)
[originally from svn r9237]
'Config' in putty.h, which stores all PuTTY's settings and includes an
arbitrary length limit on every single one of those settings which is
stored in string form. In place of it is 'Conf', an opaque data type
everywhere outside the new file conf.c, which stores a list of (key,
value) pairs in which every key contains an integer identifying a
configuration setting, and for some of those integers the key also
contains extra parts (so that, for instance, CONF_environmt is a
string-to-string mapping). Everywhere that a Config was previously
used, a Conf is now; everywhere there was a Config structure copy,
conf_copy() is called; every lookup, adjustment, load and save
operation on a Config has been rewritten; and there's a mechanism for
serialising a Conf into a binary blob and back for use with Duplicate
Session.
User-visible effects of this change _should_ be minimal, though I
don't doubt I've introduced one or two bugs here and there which will
eventually be found. The _intended_ visible effects of this change are
that all arbitrary limits on configuration strings and lists (e.g.
limit on number of port forwardings) should now disappear; that list
boxes in the configuration will now be displayed in a sorted order
rather than the arbitrary order in which they were added to the list
(since the underlying data structure is now a sorted tree234 rather
than an ad-hoc comma-separated string); and one more specific change,
which is that local and dynamic port forwardings on the same port
number are now mutually exclusive in the configuration (putting 'D' in
the key rather than the value was a mistake in the first place).
One other reorganisation as a result of this is that I've moved all
the dialog.c standard handlers (dlg_stdeditbox_handler and friends)
out into config.c, because I can't really justify calling them generic
any more. When they took a pointer to an arbitrary structure type and
the offset of a field within that structure, they were independent of
whether that structure was a Config or something completely different,
but now they really do expect to talk to a Conf, which can _only_ be
used for PuTTY configuration, so I've renamed them all things like
conf_editbox_handler and moved them out of the nominally independent
dialog-box management module into the PuTTY-specific config.c.
[originally from svn r9214]
information about where to put items that aren't mentioned in the
saved configuration. So far the only nontrivial use I've made of this
facility is to default to placing KEX_RSA just above KEX_WARN in the
absence of any other information, which should fix
'ssh2-rsa-kex-pref'.
While I'm here I've rewritten wprefs() on general principles to remove
the needless length limit, since I was touching it anyway. The length
limit is still in gprefs (but I've lengthened it just in case).
[originally from svn r9181]
code (as introduced in r9043), so that it uses the user SID rather
than the default SID.
This does change the access-control model, in that a Pageant running
with administrator privilege will now serve keys to an unprivileged
PuTTY running as the same user who started Pageant. Owen and I think
this isn't a problem (in particular, it will still not serve keys to a
_different_ user).
More importantly, making the Pageant client and server code work the
same way means that PuTTY and Pageant can still talk to each other
when UAC is turned off, which we've had several reports of r9043
having broken.
[originally from svn r9178]
[r9043 == 05f22632eb]
that we won't keep calling close_session() again the next time we go
round the message loop. Should fix unclean-close-hang. Thanks to Simon
Coleman for debugging.
[originally from svn r9115]
rectangle into smaller ones: it doesn't work any more, since the new
variable-pitch code can now call general_textout() with a larger
clipping rectangle than the text it's meant to be displaying. Instead,
general_textout() now uses the same semantics as the next loop up in
do_text_internal(): the first piece of text it displays uses the
opacity setting passed in, which blanks the entire clipping rectangle
if necessary, and then subsequent overlays are non-opaque. And the
same clipping rectangle is used throughout.
[originally from svn r9067]
array to ExtTextOut:
- move it inside the new big loop (this should fix a potential bug
whereby the DBCS handling altered some elements of it but the loop
did not actually step along it)
- initialise it more sensibly
- rename it to lpDx rather than IpDx, since as far as I can tell the
latter name was derived from a misreading of the former in the
Windows API docs.
[originally from svn r9066]
pass null lpDx, because general_textout depends on it being filled in.
Instead we null it out in the calls to subroutines _from_
general_textout.
[originally from svn r9064]
Done in much the same way as it is in the GTK front end: the character
cell width is determined using the font's digits (which seems to give
generally not-too-offensive spacing in most cases, at the expense of
Ms and Ws typically overhanging a bit into adjacent cells) and each
character is centred in its cell. Overhangs never leave permanent
droppings on the window, because the existing work done in r5003
handles them just fine even in this stressful scenario.
There's a hacky new checkbox in the Appearance panel to make
variable-pitch fonts appear in the font selector (they still don't by
default, because I still think it's _usually_ not What You Want); the
checkbox state is not actually stored as part of a saved session, but
it should be automatically ticked when reloading a session that's got
a variable pitch font selected.
(I'm half-expecting a potential flurry of requests for this feature in
the wake of http://xkcd.com/840/ , so I thought I'd pre-empt them :-)
[originally from svn r9063]
[r5003 == ba470dec5e]
I'm not sure whether I broke this in the recent revamp or whether it
was always broken, but: transitions in and out of full-screen mode
work by first maximising or restoring the window, which triggers a
WM_SIZE, whose handler then fiddles with the window style to disable
or re-enable all the furniture, which in turn triggers a recursive
WM_SIZE. The trouble is, after returning from the handler for the
inner WM_SIZE, the rest of the outer handler runs, and its client area
size is now out of date.
So I've added a flag which is set when a resize is handled 'properly',
so that after returning from the inner WM_SIZE handler the outer one
knows not to try to redo badly work that's already been done well.
[originally from svn r9056]
MinGW after r9046, and munge the COMPTR() macro to remove a couple of warnings
with my MinGW GCC (3.4.5).
[originally from svn r9049]
[r9046 == 1a03fa9292]
icon, even if the program isn't running at the time, to be presented
with an application-defined collection of helpful links). The current
jump list is updated every time a saved session is loaded, and shows
the last few launchable saved sessions (i.e. not those like Default
Settings) that were loaded. Also, if Pageant or PuTTYgen or both is in
the same directory as the PuTTY binary, the jump list will present
links to launch those too.
Based on a patch sent last year by Daniel B. Roy, though it's barely
recognisable any more...
[originally from svn r9046]
maximised state, we must be sure to disable the window offset used to
centre the terminal in cases where the window is non-negotiably the
wrong size (e.g. maximised). Hence we must call reset_window after our
terminal resize.
[originally from svn r9044]
should solve some of the SID-mismatch issues we've occasionally had
reported. Because it's a modification on the client side, it doesn't
affect the security of Pageant itself.
[originally from svn r9043]
restoring a maximised window.
Failure to do this was noticeable in the following scenario (again
using Aero UI enhancements):
1. resize window using topmost resize handle, and move pointer to top
of screen which 'maximises' the window vertically
2. now maximise the window properly using the maximise button in top
right
3. now restore. Notepad restores to its position before step 1,
because Aero remembers that position for the purpose, but PuTTY
thinks it knows better. Only now it doesn't any more.
[originally from svn r9041]
Firstly, maximise and restore events were expected never to occur
during an interactive resize process (i.e. between WM_ENTERSIZEMOVE
and WM_EXITSIZEMOVE), but in fact Aero now allows this to happen if
you move the pointer to the top of the screen while dragging the
window.
Secondly, plain old WM_SIZE events were expected never to occur
_outside_ interactive resizes, but Aero permits that too (e.g.
Windows-left and Windows-right), and also third-party window
repositioning tools will send these.
[originally from svn r9040]
are now loaded from standard locations (system32 for SSPI, the
registry-stored MIT KfW install location for KfW) rather than using
the risky default DLL search path; I've therefore also added an
option to manually specify a GSS DLL we haven't heard of (which
should in principle Just Work provided it supports proper GSS-API as
specified in the RFC). The same option exists on Unix too, because
it seemed like too useful an idea to reserve to Windows. In
addition, GSSAPI is now documented, and also (unfortunately) its GUI
configuration has been moved out into a sub-subpanel on the grounds
that it was too big to fit in Auth.
[originally from svn r9003]
called load_system32_dll() which constructs a full pathname for the
DLL using GetSystemDirectory.
The only DLL load not covered by this change is the one for
gssapi32.dll, because that one's not in the system32 directory.
[originally from svn r8993]
reorganises the GSSAPI support so that it handles alternative
implementations of the GSS-API. In particular, this means PuTTY can
now talk to MIT Kerberos for Windows instead of being limited to
SSPI. I don't know for sure whether further tweaking will be needed
(to the UI, most likely, or to automatic selection of credentials),
but testing reports suggest it's now at least worth committing to
trunk to get it more widely tested.
[originally from svn r8952]
today reported an SSH2_MSG_UNIMPLEMENTED from a Cisco router which
looks as if it was triggered by SSH2_MSG_IGNORE, so I'm
experimentally putting this flag in. Currently must be manually
enabled, though if it turns out to solve the user's problem then
I'll probably add at least one version string...
[Edited commit message: actually, I also committed in error a piece
of experimental code as part of this checkin. Serve me right for not
running 'svn diff' first.]
[originally from svn r8926]
NameUserPrincipal, use that; this avoids an issue with SSPI/GSSAPI where
the user logged in to the local machine with a different case of username
to the (case-sensitive) Kerberos username. Falls back to GetUserName as
before if that doesn't work (for machines not on a domain, and Win9x).
Based on a patch by SebastianUnger.
[originally from svn r8909]
WspiapiGetAddrInfo wrapper for getaddrinfo() in MSVC. Split GET_WINDOWS_FUNCTION
into two variants, one with the old behaviour (bypassing the preprocessor) and
another with the new behaviour (for ANSI/Unicode, although it's not actually
used anywhere currently).
[originally from svn r8898]
[r8738 == 24b6168c1d]
function in terminal.c, and replace the cloned-and-hacked handling
code in all our front ends with calls to that.
This was intended for code cleanliness, but a side effect is to make
the GTK arrow-key handling support disabling of application cursor
key mode in the Features panel. Previously that checkbox was
accidentally ignored, and nobody seems to have noticed before!
[originally from svn r8896]
linking on Windows into a single global one, which can cope with function
renaming. Intended to enable eventual removal of ANSI-specific DoSomethingA
references (although I've not removed any).
[originally from svn r8738]
uninitialised storage into WSAAddressToString()'s length function (and
presumably getting away with it by luck).
Also improve error handling (exposed by my Wine installation, which returns
an error from WSAAddressToString() for connections to localhost for some
reason).
[originally from svn r8737]
not fazed by being displayed at other than 96DPI; testing on Vista at a range
of DPIs indicates that we cope (with the minor and inevitable exception of the
drag-list control).
This stops pixel scaling and hence fuzzy display on high-resolution displays.
(Hope this is last disastrous than my last set of manifest tweaks! --
<http://support.fogcreek.com/default.asp?copilot.6.26840.1> suggests that this
is an OK thing to do.)
[originally from svn r8661]
plink did not cope gracefully with this -- it was not possible to override that
hostname on the command line (attempts at doing so would be treated as part of
the remote command).
Fix this by applying the principle of r7265: if the user didn't explicitly
specify that they wanted to launch the hostname in the default (for instance
with '-load "Default Settings"', we assume they don't want to, and such a
hostname doesn't count when deciding whether to treat a non-option argument as
hostname or command.
[originally from svn r8651]
[r7265 == 5d76e00dac]
[r7266 == 856ed4ae73]
sessions submenu of the terminal window context menu (as Pageant does), rather
than an empty menu (which often renders poorly).
[originally from svn r8648]
use the explicitly-narrow type LPSTR, not the switchable type LPTSTR. (Since
we currently build without UNICODE this makes no practical difference to us
now.)
[originally from svn r8627]
UTF-16 when exchanging wchar_t strings with the front end. Enabled
by a #define in the platform's header file (one should not
promiscuously translate UTF-16 surrogate pairs on 32-bit wchar_t
platforms since that could give rise to redundant encoding attacks),
which is present on Windows.
[originally from svn r8495]
prompts packet containing no actual prompts (perhaps due to odd
server organisation, or perhaps so it can print a banner message and
do nothing else). Previously, the get_userpass_input functions
always returned failure when in '-batch' mode, even in this case
where no actual input would be required.
[originally from svn r8490]
to a Unix-domain socket. This typically works fine when PuTTY is run on the
same machine as the X server, but it's broken multi-hop X forwarding through
OpenSSH; when OpenSSH creates a proxy X server "localhost:10", it only listens
on TCP, not on a Unix-domain socket.
Instead, when deciding on the details of the display, we actively probe to see
if there's a Unix-domain socket we can use instead, and only use it if it's
there, falling back to the specified IP "localhost" if not.
Independently, when looking for local auth details in Xauthority for a
"localhost" TCP display, we prefer a matching Unix-domain entry, but will fall
back to an IP "localhost" entry (which would be unusual, but we don't trust a
Windows X server not to do it) -- this is a generalisation of the special case
added in r2538 (but removed in r8305, as the automatic upgrade masked the need
for it).
(This is now done in platform-independent code, so a side-effect is that
get_hostname() is now part of the networking abstraction on all platforms.)
[originally from svn r8462]
[r2538 == fda9983243]
[r8305 == ca6fc3a4da]
read at connection setup time, so don't offer it in the Change
Settings dialog box.
(In particular, this fixes an assertion failure when selecting
Change Settings on a non-SSH connection, since wincfg.c would have
added that control to Connection/SSH/X11 when the parent panel
Connection/SSH didn't exist. Making the control conditional on the
selected protocol would have been sufficient to fix that failure,
but I now realise that the setting should never have been presented
in mid-session in any case.)
[originally from svn r8443]
only call it when the _last_ mouse button comes back up. Otherwise,
xterm mouse tracking will lose a button-up event if you press down
two buttons, move the mouse outside the window, then release them
one at a time.
[originally from svn r8425]
OVERLAPPED structure in output threads, as we already do for input
threads. This apparently sorts out a hanging issue with serial ports
when trying to do simultaneous read and write, because (GJV says,
and it sounds plausible to me) in the absence of that event object
Windows signals the file handle itself to notify GetOverlappedResult
that it can return - and since the file handle might be being
signalled by a read operation instead, that leads to ambiguity.
Using an explicit event object in both directions means Windows
always knows which way the data is going.
Also a trivial fix in handle_output_new(), which was referencing the
wrong element of a union due to a copy and paste error. (Since the
result was address-taken and cast to void *, this wasn't a
functional error, but it was conceptually wrong.)
[originally from svn r8410]
case of double-width text (ESC # 3, ESC # 4, ESC # 6), because the
string passed to it was not truncated to the same width as the
clipping rectangle. (In fact, it _can't_ reliably be, in the case
where the window width is odd.) So instead we just assert that we
managed to _at least_ fill the clipping rectangle, not that we
exactly filled it.
The problem is easily reproduced by sending ESC # 8 (fill the screen
with Es) followed by ESC # 3. It doesn't typically happen, though,
if you _manually_ fill the screen with Es, because in that case
PuTTY's terminal buffer ends up being filled with CSET_ACP | 'E' or
similar, which means that general_textout() never gets called
because one of the other branches of do_text_internal() does the
work instead. ESC # 8 will fill the terminal buffer with genuine
_Unicode_ 'E' characters, which exercises the failing code path.
[originally from svn r8403]
sk_hostname_is_local(), to catch the case where we're doing something like X11
forwarding over SSH through a proxy, and we've thus disabled local lookup of
hostnames.
(I think this is what's behind the report in
<e9a86996-5dc2-4428-9b0c-c65693ca6351@m32g2000hsf.googlegroups.com>
in comp.security.ssh, although I'd like to know more of the circumstances.)
[originally from svn r8385]
void *, and hence eliminate a few casts. The Windows definition is
unchanged, but I daresay I've managed to stop it compiling nonetheless.
[originally from svn r8359]
subthread, so that it won't deadlock if fetching the content of the
clipboard turns out to depend on a network connection forwarded
through PuTTY.
[originally from svn r8338]
directly. Fixes a build failure involving name clashes between
winsock2.h and winsock.h, which had somehow managed to get included
in succession.
[originally from svn r8332]
While I'm here, a cosmetic PuTTYtel change: remove a reference to SSH from the
"logical host name" label in PuTTYtel only.
[originally from svn r8331]
[r8305 == ca6fc3a4da]
ourselves, but on Unix then assumed it was compatible with the system's
gss_buffer_desc, which wasn't the case on LP64 systems. Now, on Unix
we make Ssh_gss_buf into an alias for gss_buffer_desc, though we keep
something similar to the existing behaviour on Windows. This requires
renaming a couple of the fields in Ssh_gss_buf, and hence fixing all
the references.
Tested on Linux (MIT Kerberos) and Solaris. Compiled on NetBSD (Heimdal).
Not tested on Windows because neither mingw32 nor winegcc worked out of the
box for me. I think the Windows changes are all syntactic, though, so
if this compiles it should work no worse than before.
[originally from svn r8326]
if we have no better ideas, with UI shamelessly stolen from Quest PuTTY.
Off by default, which effectively reverts the change to using the local
username by default that came in with GSSAPI support in r8138. Anyone wanting
seamless single sign-on will need to set the new option. (The previous
default behaviour was getting in the way in ad-hoc scenarios.)
Note that the PSCP and Unix-Plink behaviour of using the local username by
default have remained unchanged throughout; they are not affected by the new
option. Not sure if that's the Right Thing.
[originally from svn r8324]
[r8138 == de5dd9d65c]
strings more rigorously, and then we look up the local X authority
data in .Xauthority _ourself_ rather than delegating to an external
xauth program. This is (negligibly) more efficient on Unix, assuming
I haven't got it wrong in some subtle way, but its major benefit is
that we can now support X authority lookups on Windows as well
provided the user points us at an appropriate X authority file in
the standard format. A new Windows-specific config option has been
added for this purpose.
[originally from svn r8305]
"curraddr", and turn "family" into a macro-derived property of the
other fields. The idea is that this renders SockAddrs immutable once
created, which should open up the possibility of duplicating and
reusing one without having to redo the actual DNS lookup.
I _hope_ I haven't broken anything. The new code architecture
contains several rather dubious-looking operations (namely the
arbitrary choice of the first returned address in functions like
sk_getaddr and sk_address_is_local - what if, for instance, a DNS
lookup returned a local and a non-local address?), but I think they
were functionally just as dubious beforehand and all this change has
done is to make them more obviously so to a reader.
[originally from svn r8293]
combination Ctrl + \ as the Ctrl-\ character. All of mine have, but
at least one laptop turns out not to. Do so explicitly.
[originally from svn r8182]
ago, I apparently caused all data received from local proxies to be
unconditionally tagged as TCP Urgent. Most network backends ignore
this, but it's critical to the Telnet backend, which will ignore all
Urgent-marked data in the assumption that there's a SYNCH on its way
that it should wait for. Nobody has noticed in two years, presumably
meaning that nobody has ever tried to do Telnet over a local proxy
in that time.
[originally from svn r8158]
to manually tweak the host name and port number under which the SSH
host key is read and written.
I've put it in the cross-platform Connection panel. Partly under the
flimsy pretext that other backends _can_ use it if they so wish (and
in fact it overrides the host name for title-bar purposes in all
network backends, though it has no other effect in anything but
SSH); but mostly because the SSH panel was too full already :-)
[originally from svn r8033]
addressing X displays. Update PuTTY's display-name-to-Unix-socket-
path translation code to cope with it, thus causing X forwarding to
start working again on Leopard.
[originally from svn r8020]
underlying WinSock SOCKET. Therefore, if we plan to modify the
SOCKET in a socket, we must remove it from the tree before doing so,
and put it back again afterwards. Otherwise it'll violate the tree's
sorting order, and sooner or later someone will try to find it and
get back NULL.
[originally from svn r7795]
port number in the GUI when the connection type is changed if the current
port number is the standard one for the current protocol.
It's not perfect, but it should make the common case of tabbing through the
Session panel easier when starting non-SSH connections on odd ports.
[originally from svn r7635]
Should be no significant change in behaviour.
(Well, entering usernames containing commas on Plink's command line will be
a little harder now.)
[originally from svn r7628]
Instead of passing -1 to its gotdata and sentdata callbacks on
error, winhandl.c will now pass the negation of the Windows error
number; and the Plink front end will now format that into an error
message and pass it on to the user.
[originally from svn r7416]
[r7415 == 702a92ceb8]
from is closed normally from the writing end. This is ludicrous; if
that situation isn't a natural EOF, _nothing_ is. So if we get that
particular error, we pretend it's EOF.
[originally from svn r7415]
simply specifying a hostname on the command line -- this would bring up the
config dialog. Use a slightly more sophisticated notion of whether the user
meant to launch a session.
[originally from svn r7321]
[r7265 == 5d76e00dac]
wouldn't care about except for the fact that it's still used to
implement the Saved Sessions menu item in PuTTY and Pageant.
[originally from svn r7291]
[r7265 == 5d76e00dac]
putting HTML Help into "single-threaded" mode. Furthermore, this requires
extra work from the application (message pumping via HH_PRETRANSLATEMESSAGE).
Thus, remove them and run Help in a secondary thread. This means that keyboard
input into the Index and Search tabs now works.
[originally from svn r7285]
represent a launchable session, unless the user can be construed to
have really meant it. This means:
- starting up PuTTY when the Default Settings are launchable still
brings up the config box, and you have to hit Open to actually
launch that session
- double-clicking on Default Settings from the config box will load
them but not launch them.
On the other hand:
- explicitly loading the Default Settings on the command line using
`-load' _does_ still launch them.
[originally from svn r7265]
parts of the versioning code which might not like them.
As a result of this checkin, bob builds from modified SVN working
copies will still announce themselves as revision nnnnM in the
textual version strings, but their binary version in the Windows
VERSIONINFO will now be 0.0.0.0.
[originally from svn r7231]
a fourth class of PuTTY version tags in addition to release,
snapshot and unidentified: we now have `Custom build r1234',
indicating a build made from that SVN revision in a context other
than that of a dated snapshot. The build script generates these when
it doesn't know what else to do; `unidentified builds' will now only
occur when you run nmake from the command line.
Also, the build script now generates sensible version data in the
installer to match this. So I _think_ we should now be set to use
bob to generate installer builds of the nightly snapshots, although
of course I'll have to wait until tomorrow to test one.
[originally from svn r7211]
term_provide_resize_fn() was not being broken when the back end was
destroyed on session termination, causing resizing an inactive PuTTY
to be a segfault hazard.
[originally from svn r7143]
which have been broken since r6797.
(At least some versions of Win9x are gratuitously picky about the arguments to
CreateThread(), requiring lpThreadId not to be NULL.)
[originally from svn r7132]
[r6797 == 291533d3f9]
It's specific to the Windows installer, so it seems unnecessarily confusing to
have it in the top level of the source distribution alongside README.
[originally from svn r7125]
quite big and tends to hide the existence of the `Serial' config
panel.
This is implemented by folding up every branch of depth 2 or more,
which with any luck might turn out to be general enough to carry
over unchanged if other branches start expanding. Then again, we may
have to fiddle with it again when that time comes; who knows?
[originally from svn r7117]
about it very hard; it's a plausible fix for the observed cause of the extreme
CPU usage, being that we were asking to be notified of Windows messages and
then not dealing with them, plausibly leading to a loop. Works for me,
anyway.)
[originally from svn r7098]
it'll let you see an identifier (SHGFP_TYPE_CURRENT) referenced since r7082.
(Actually, you need a pretty recent w32api before it's there at all.)
Morally, this should be defined for all toolchains, not just MinGW/Cygwin, but I'll leave that to people who have those toolchains.
<http://msdn2.microsoft.com/en-us/library/aa383745.aspx>
Also add some other comments on our use of this API (since it's a horrible one
that I suspect will come back and haunt us...)
[originally from svn r7087]
[r7082 == dbbd6eb5ec]
[this svn revision also touched putty-wishlist]
takes a third argument which is TRUE if the file is being opened for
writing and wants to be created in such a way that it's readable
only to the owner. This is used when saving private keys.
While I'm here, I also use this option when writing session logs, on
the general principle that they probably contain _something_
sensitive.
The new argument is only supported on Unix, for the moment. (I think
writing owner-accessible-only files is the default on Windows.)
[originally from svn r7084]
Application Data directory in preference to the old-fashioned
attempt to find the user's home directory, and use the _local-
machine_ Application Data directory in preference even to that.
SHGetFolderPath() is called via GetProcAddress, so this degrades
gracefully on old Windowses. (Tested myself on Win95.)
As part of this change, we now search for a location for the seed
file separately for reading and writing, so that installing the new
PuTTY should cause a seamless migration as the old seed file is read
from the old location and then a new one written to the new location.
`putty -cleanup' attempts to delete the seed file from _all_
affected locations.
Naturally, a user-specified seed file path in the Registry still
takes priority over all other means of finding the location.
[originally from svn r7082]
caused by the MessageBox() internal message loop eating WinSock
FD_READ notifications, which then don't reappear afterwards because
you have to explicitly prod a socket in order to get a repeat
notification on it.
Hence, here's a piece of infrastructure which seems to sort it out:
a new winnet.c function called socket_reselect_all(), whose function
is to go through all currently active sockets and re-run
WSAAsyncSelect() on them, causing repeat notifications for anything
we might have missed. I call this after every call to MessageBox(),
and that seems to solve the problem.
(The problem was actually masked in very recent revisions, probably
by the reinstatement of pending_netevent in r7071. However, I don't
believe that was a complete fix. This should be.)
[originally from svn r7077]
[r7071 == 57a763b0ec]
r4906 in the process of adding the new timing code. It seems to have
been what was previously preventing spew-lockup, and still seems to
prevent it now I've put it back in.
[originally from svn r7071]
[r4906 == 7ecf13564a]
[this svn revision also touched putty-wishlist]
match the original icons. (Apparently I managed to introduce errors
while transcribing the originals for detailed analysis.)
While I'm at it, add the obviously useful `make install' target in
icons/Makefile, and fix the svn:ignore property on the icons
directory.
[originally from svn r7068]
suite. In a dramatic break with tradition, I'm actually checking in
the resulting icon files as well as the script that generates them,
because the script requires Python and ImageMagick and I don't think
it's reasonable to require that much extra infrastructure on
everyone checking out from Subversion.
The new icons should be _almost_ indistinguishable from the old
ones, at least at the 32x32 resolution. The immediately visible
change is that all the icons now come in 16x16, 32x32 and 48x48
formats, in both 16 colours and monochrome, instead of an ad-hoc
mixture of whichever ones I could be bothered to draw.
The same code can also be adapted to generate icons for the GTK port
(although icons for the running programs don't seem to be supported
by GTK 1 - another reason to upgrade to GTK 2!).
[originally from svn r7063]
since even the latest version of w32api (3.6) shows no sign of HTMLHelp
support.
(This touches mkfiles.pl because that's where the details of what Cygwin
doesn't support are kept currently. This may be deliberate, so I haven't
changed it.)
[originally from svn r7032]
making the manual shortcut in the Start menu point to one or other
of the two help files depending on the version of Windows;
fortunately Inno Setup has no difficulty doing that.
[originally from svn r7028]
and various calls to WinHelp() have been centralised into a new file
winhelp.c, which in turn has been modified to detect a .CHM file as
well as .HLP and select between them as appropriate. It explicitly
tries to load HHCTRL.OCX and use GetProcAddress, meaning that it
_should_ still work correctly on pre-HTML-Help platforms, falling
gracefully back to WinHelp, but although I tested this by
temporarily renaming my own HHCTRL.OCX I haven't yet been able to
test it on a real HTML-Help-free platform.
Also in this checkin: a new .but file and docs makefile changes to
make it convenient to build the sources for a .CHM. As yet, owing to
limitations of Halibut's CHM support, I'm not able to write a .CHM
directly, more's the pity.
[originally from svn r7000]
easily manage, by adopting a hybrid approach to Unicode text
display. The old approach of simply calling ExtTextOutW provided
font linking without us having to lift a finger, but didn't do the
right thing when it came to bidirectional or Arabic-shaped text.
Arabeyes' replacement exact_textout() supported the latter, but
turned out to break the former (with no warning from the Windows API
documentation, so it's not their fault).
So now I've got a second wrapper layer called general_textout(),
which splits the input string into substrings based on bidi
character class. Any character liable to cause bidi or shaping
behaviour if fed straight to ExtTextOutW is instead fed through
Arabeyes' exact_textout(), but the rest is fed straight to
ExtTextOutW as it used to be.
The effect appears to be that font linking is restored for all
characters _except_ Arabic and other bidi scripts, which means in
particular that we are no longer in a state of regression over 0.57.
(0.57 would have done font linking on Arabic as well, but would also
have misbidied it, so we've merely exchanged one failure mode for
another slightly less harmful one in that situation.)
[originally from svn r6910]
session, we were clearing the new session_closed flag, but failing
to clear must_close_session; with that set, the session was being
opened but immediately re-closed.
[originally from svn r6857]
[r6802 == 0dcdb6c3c1]
required. (I just tried getting rid of them; it worked fine for
serial ports, but not for anything else. The Windows I/O API sucks.)
[originally from svn r6843]
behave like a pointer. In particular, the right thing to set a
HANDLE to to indicate that it's invalid is INVALID_HANDLE_VALUE, not
NULL. Crack down on sloppy use of NULL HANDLEs across all Windows
code.
(There is one oddity, which is that {Create,Open}FileMapping are
documented to return a NULL HANDLE instead of INVALID_HANDLE_VALUE
on failure. Shrug. If MS want to be inconsistent, I suppose I have
to live with it.)
[originally from svn r6833]
values one might expect, which means that GetMessage() was
occasionally blocking the process. That appears to be the last of
the annoying data loss issues, so I think the Windows serial back
end actually looks vaguely reliable now. Phew.
[originally from svn r6830]
unfriendly in an interactive session, because at 19200 baud it takes
nearly two seconds to receive that much data, and as long as the
data is flowing continuously Windows waits until it has a full
buffer. So here's another annoying flag in the winhandl API, which
restricts reads to length 1 so that serial output shows up as it
appears.
(I tried this yesterday, but without the OVERLAPPED fix in r6826 it
behaved very erratically. It now seems solid.)
[originally from svn r6827]
[r6826 == 2aedc83f8d]
there): `plink host -nc host2:port' causes the SSH connection's main
channel to be replaced with a direct-tcpip connection to the
specified destination. This feature is mainly designed for use as a
local proxy: setting your local proxy command to `plink %proxyhost
-nc %host:%port' lets you tunnel SSH over SSH with a minimum of
fuss. Works on all platforms.
[originally from svn r6823]
options, here's a slight change to the API of ser_setup_config_box()
to make it filter its parity and flow control options using
platform-supplied bit masks.
[originally from svn r6820]
in place of making a network connection. This has involved a couple
of minor infrastructure changes:
- New dlg_label_change() function in the dialog.h interface, which
alters the label on a control. Only used, at present, to switch
the Host Name and Port boxes into Serial Line and Speed, which
means that any platform not implementing serial connections (i.e.
currently all but Windows) does not need to actually do anything
in this function. Yet.
- New small piece of infrastructure: cfg_launchable() determines
whether a Config structure describes a session ready to be
launched. This was previously determined by seeing if it had a
non-empty host name, but it has to check the serial line as well
so there's a centralised function for it. I haven't gone through
all front ends and arranged for this function to be used
everywhere it needs to be; so far I've only checked Windows.
- Similarly, cfg_dest() returns the destination of a connection
(host name or serial line) in a text format suitable for putting
into messages such as `Unable to connect to %s'.
[originally from svn r6815]
a serial port backend:
- In order to do simultaneous reading and writing on the same
HANDLE, you must enable overlapped access and pass an OVERLAPPED
structure to each ReadFile and WriteFile call. This would make
sense if it were an optional thing I could do if I wanted to do
the reading and writing in the same thread, but making it
mandatory even if I'm doing them in _different_ threads is just
annoying and arbitrary.
- Serial ports occasionally return length 0 from ReadFile, for no
particularly good reason. Fortunately serial ports also don't
have a real EOF condition to speak of, so ignoring EOFs is
actually a viable response in spite of sounding utterly gross.
Hence, handle_{input,output}_new() now accept a flags parameter,
which includes a flag to enable the OVERLAPPED bureaucracy and a
flag to cause EOFs to be ignored on input handles. The current
clients of winhandl.c do not use either of these.
[originally from svn r6813]
it's NULL. Since we already have one back end (uxpty) which doesn't
in fact talk to a network socket, and may well have more soon, I'm
replacing this TCP/IP-centric function with a nice neutral
`connected' function returning a boolean. Nothing else about its
semantics has currently changed.
[originally from svn r6810]
inherit _our_ ends of its I/O pipes! Otherwise, closing our copy of
those handles does not cause it to see EOF on its stdin, because
it's holding the pipe open itself.
[originally from svn r6808]
to do something, otherwise handle_get_events will forget to tell the
front end to check for that subthread finishing. This applies even
when we're only setting `busy' to tell the subthread to terminate!
[originally from svn r6805]
of the previous ad-hockery which depended on the return value from
select_result() and hence which will not adapt sensibly to a world
in which the primary session is something local rather than a
network connection.
[originally from svn r6802]
because it gets unconditionally sfree()d in sk_addr_free(). This
just bit me when running under the MSVC debugger; not sure how it
hasn't bitten anyone until now!
[originally from svn r6800]
thread-based approach to stdin and stdout, wraps it in a halfway
sensible API, and makes it a globally available service across all
network tools.
There is no direct functionality enhancement from this checkin:
winplink.c now talks to the new API instead of doing it all
internally, but does nothing different as a result.
However, this should lay the groundwork for several diverse pieces
of work in future: pipe-based ProxyCommand on Windows, a serial port
back end, and (hopefully) a pipe-based means of communicating with
Pageant, which should have sensible blocking behaviour and hence
permit asynchronous agent requests and decrypt-on-demand.
[originally from svn r6797]
we set _FILE_OFFSET_BITS to 64 on the compiler command line (via mkfiles.pl),
and on Windows we use SetFilePointer and GetFileSize to cope with 64-bit sizes
where possible. Not tested on Win9x.
[originally from svn r6783]
Shift-hjklyubn for batch movement in NetHack, because they have
subtly different behaviour within the game and the Ctrl-moves are
more useful. Unfortunately, PuTTY's NetHack keypad mode doesn't
support Ctrl-moves. Therefore, it does now :-)
[originally from svn r6593]
- Now we've fixed `win-versioninfo', choose some sensible outcomes from
the installer's comparisons of binary version numbers. Also, give the
installer _itself_ a matching binary version.
In particular, without this change, it would not have been possible
to downgrade PuTTY -- it would have silently left the "newer" files in
place. Now it will make some fuss, but permit it.
- Also remove descriptions from shortcuts, on the grounds that the
binaries have embedded descriptions now. (Although I've not checked
whether those are actually visible in the Start Menu.)
- At the request of various people (e.g., PJB), add flags so that if
files are in use at the time the (un)installer is run, replacement is
deferred to the next restart. (The user may be prompted to restart,
which isn't ideal; see comments).
This is supposed to make centrally-pushed silent upgrades more robust.
- Note some limitations of the installer.
[originally from svn r6585]
Pageant for local authentication. (This is a `don't use Pageant for
authentication at session startup' button rather than a `pretend
Pageant doesn't exist' button: that is, agent forwarding is
independent of this option.)
[originally from svn r6572]
button tend to get disabled on login.
After a suggestion by "Tkil", change the way we handle the specials menu
to be robust against the window menu being externally modified.
[originally from svn r6546]
[this svn revision also touched putty-wishlist]
nicely elsewhere, which should fix `win64' _properly_.
Tested on recent-ish MinGW (with GetWindowLongPtr but not GetClassLongPtr),
and VC++ 6.0 with a recent SDK, but not with vanilla VC++.
[originally from svn r6535]
[this svn revision also touched putty-wishlist]
basis for other terminal-involving applications: a stub
implementation of the printing interface, an additional function in
notiming.c, and also I've renamed the front-end function beep() to
do_beep() so as not to clash with beep() in lib[n]curses.
[originally from svn r6479]
abstracted out; replace loops structured around a single interaction
per loop with less tortuous code (fixes: `ki-multiprompt-crash',
`ssh1-bad-passphrase-crash'; makes `ssh2-password-expiry' and
`proxy-password-prompt' easier).
The new interaction abstraction has a lot of fields that are unused in
the current code (things like window captions); this is groundwork for
`gui-auth'. However, ssh.c still writes directly to stderr; that may
want to be fixed.
In the GUI apps, user interaction is moved to terminal.c. This should
make it easier to fix things like UTF-8 username entry, although I
haven't attempted to do so. Also, control character filtering can be
tailored to be appropriate for individual front-ends; so far I don't
promise anything other than not having made it any worse.
I've tried to test this fairly exhaustively (although Mac stuff is
untested, as usual). It all seems to basically work, but I bet there
are new bugs. (One I know about is that you can no longer make the
PuTTY window go away with a ^D at the password prompt; this should be
fixed.)
[originally from svn r6437]
[this svn revision also touched putty-wishlist]
a VERSIONINFO resource. The versioning scheme is described in
windows/version.rc2.
Some .rc files are now #included in others. In order to keep MSVC
project files working, these have been renamed to .rc2; there may exist
a better solution.
(This checkin also includes the documentation tweak missing from r6367.)
Testing performed:
- MinGW (cross-compiler): works
- VC nmake: works (tested with VC6)
- VC project files: builds with VERSIONINFO resource (no VER variable though)
- Borland: an old version of this patch was tested with it and more or
less worked, except that some of the VERSIONINFO strings were apparently
not terminated properly. Not attempted to work around this.
- LCC: not tested. Some fixes are in there from the last time we tried
this, but then the build ultimately failed and I haven't tried this
since that was fixed.
- Dev-C++: untested. (Haven't done anything special.)
- Unix Gtk/autoconf Makefiles work as before.
[originally from svn r6374]
[r6367 == f86ad059db]
[this svn revision also touched putty-wishlist]
our app-private window messages, which is considerably higher than the
WM_XUSER we arbitrarily chose. (This isn't known to be causing any actual
problems. The fix seems not to have obviously broken anything.)
[originally from svn r6183]
[this svn revision also touched putty-wishlist]
eaten by the trailing "\f0" on the RTF preamble. The RTF spec (1.0 and 1.6)
suggests that adding a space should defuse this situation and be otherwise
harmless, and it works for me (Win98).
[originally from svn r5931]
there are servers which could in principle operate in this mode, although I
don't know if any do in practice. (Hence, I haven't been able to test it.)
[originally from svn r5748]
[this svn revision also touched putty-wishlist]
Unix Plink sends everything sensible it can find, and it's fully configurable
from the GUI.
I'm not entirely sure about the precise set of modes that Unix Plink should
look at; informed tweaks are welcome.
Also the Mac bits are guesses (but trivial).
[originally from svn r5653]
[this svn revision also touched putty-wishlist]
