]> asedeno.scripts.mit.edu Git - PuTTY.git/log
PuTTY.git
15 years agoColin also suggests that we can replace the backticks in the GTK
Simon Tatham [Thu, 8 Jan 2009 18:19:52 +0000 (18:19 +0000)]
Colin also suggests that we can replace the backticks in the GTK
makefile with make's own $(shell ...) function, which means that
gtk-config and krb5-config and so on only get run once per make
invocation instead of once per gcc invocation.

[originally from svn r8400]

15 years agoColin Watson reports a strange problem whereby krb5-config
Simon Tatham [Thu, 8 Jan 2009 18:18:14 +0000 (18:18 +0000)]
Colin Watson reports a strange problem whereby krb5-config
incorporates the environment variable CFLAGS into its output. Avoid
exporting our version of it from the Makefile (which actually causes
build failures, since quoting phase issues mean that the backticks
in our version end up unexpanded).

[originally from svn r8399]

15 years agoRejig windlg.c:verify_ssh_host_key() to silence a warning.
Jacob Nevins [Tue, 6 Jan 2009 00:25:07 +0000 (00:25 +0000)]
Rejig windlg.c:verify_ssh_host_key() to silence a warning.

[originally from svn r8394]

15 years agoRemove a couple of unused variables.
Jacob Nevins [Tue, 6 Jan 2009 00:16:35 +0000 (00:16 +0000)]
Remove a couple of unused variables.

[originally from svn r8393]

15 years agoIt's a new year (and there have even been checkins).
Jacob Nevins [Mon, 5 Jan 2009 23:49:19 +0000 (23:49 +0000)]
It's a new year (and there have even been checkins).

[originally from svn r8392]
[this svn revision also touched putty-website]

15 years agosk_address_is_local() failed to cope when presented with a Unix-domain socket.
Jacob Nevins [Mon, 5 Jan 2009 23:36:14 +0000 (23:36 +0000)]
sk_address_is_local() failed to cope when presented with a Unix-domain socket.
This could cause Unix PuTTY to segfault when X forwarding over an SSH session
through a proxy.
(sk_getaddr() wouldn't cope either -- in that case, add an assertion to make it
more obvious; I don't think it should ever happen.)

[originally from svn r8391]

15 years agoBe prepared for x11_setup_display() to return NULL (for instance, if DISPLAY
Jacob Nevins [Mon, 5 Jan 2009 22:53:14 +0000 (22:53 +0000)]
Be prepared for x11_setup_display() to return NULL (for instance, if DISPLAY
has invalid contents). Simply refuse to try X forwarding in this circumstance.

[originally from svn r8390]

15 years agor5080 didn't go far enough -- since 0.58, anyone on Unix who didn't previously
Jacob Nevins [Mon, 5 Jan 2009 19:39:27 +0000 (19:39 +0000)]
r5080 didn't go far enough -- since 0.58, anyone on Unix who didn't previously
have default settings for "bell overload" mode will have inherited a twitchy
set of defaults where bells are disabled after 2ms and enabled after 5ms,
rather than 2s and 5s as intended. This error has probably propagated into some
people's saved sessions by now, but there's not much to be done.

[originally from svn r8389]
[r5080 == 7647f57dc43155606132afc75f354b99c41a899a]

15 years agoAdd some hard-coded textual literal-IP representations of localhost to
Jacob Nevins [Mon, 5 Jan 2009 02:45:38 +0000 (02:45 +0000)]
Add some hard-coded textual literal-IP representations of localhost to
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]

15 years ago...and fix an unlikely memory leak.
Jacob Nevins [Mon, 5 Jan 2009 01:15:06 +0000 (01:15 +0000)]
...and fix an unlikely memory leak.

[originally from svn r8384]

15 years agoCope with a (non-standard) ENAMETOOLONG return from gethostname(); glibc will
Jacob Nevins [Mon, 5 Jan 2009 01:01:58 +0000 (01:01 +0000)]
Cope with a (non-standard) ENAMETOOLONG return from gethostname(); glibc will
do this if the supplied buffer isn't big enough, which shouldn't lead to
complete abandonment of X11 auth. (Would only have bitten with hostnames
>255 chars anyway.)

[originally from svn r8383]

15 years agor8305 made platform_x11_best_transport[] obsolete, but there still seem to be a
Jacob Nevins [Sun, 4 Jan 2009 23:36:24 +0000 (23:36 +0000)]
r8305 made platform_x11_best_transport[] obsolete, but there still seem to be a
few instances around; expunge them.

[originally from svn r8382]
[r8305 == ca6fc3a4daf51166a15693feffc967bee9e3f59a]

15 years agor8305 seems to have made Unix PuTTY rather over-keen on Unix-domain sockets;
Jacob Nevins [Sun, 4 Jan 2009 23:28:25 +0000 (23:28 +0000)]
r8305 seems to have made Unix PuTTY rather over-keen on Unix-domain sockets;
unless a protocol is explicitly specified with "tcp/foovax:0", it assume a
Unix-domain socket, thus not allowing a remote display on a machine other than
the client.

[originally from svn r8381]
[r8305 == ca6fc3a4daf51166a15693feffc967bee9e3f59a]

15 years agoBah, I've bumped into this often enough. Change one unsatisfactory cast for
Jacob Nevins [Sun, 4 Jan 2009 22:24:08 +0000 (22:24 +0000)]
Bah, I've bumped into this often enough. Change one unsatisfactory cast for
another to shut up "warning: cast from pointer to integer of different size"
(and hence a -Werror compile failure) when compiling for Unix with DEBUG
defined on atreus (x86_64). Minimally checked that it doesn't introduce upset
elsewhere (i386).

[originally from svn r8380]

15 years agoLog reason for failure to create a connection to a local X display.
Jacob Nevins [Sun, 4 Jan 2009 21:43:20 +0000 (21:43 +0000)]
Log reason for failure to create a connection to a local X display.

[originally from svn r8379]

15 years agor8338 broke pasting with Shift-Ins and from the context menu on Windows.
Jacob Nevins [Mon, 29 Dec 2008 20:04:42 +0000 (20:04 +0000)]
r8338 broke pasting with Shift-Ins and from the context menu on Windows.
Divert these to use the request_paste() interface.

[originally from svn r8377]
[r8338 == 54835a9838b7122eaa53d61966ffb8537c9362b0]

15 years ago"Derek" reports that the DECSLPP report of window-size-in-pixels (CSI 14 t)
Jacob Nevins [Sat, 20 Dec 2008 19:43:20 +0000 (19:43 +0000)]
"Derek" reports that the DECSLPP report of window-size-in-pixels (CSI 14 t)
has width and height swapped. Since both a random xterm I have and
<http://invisible-island.net/xterm/ctlseqs/ctlseqs.txt> agree with him, I've
changed ours. (This stuff appears to originate in dtterm, but I can't check the
behaviour of that right now.)

While I'm here, the are-we-iconified report (CSI 11 t) looks to have the
wrong sense compared to the same sources, so swap that too.

(All this has been this way since it was originally implemented in r1414,
which doesn't cite a source. all-escapes is silent too.)

[originally from svn r8376]
[r1414 == bb1f5cec3188fd1e0329425212c149959151450e]

15 years ago"Derek" points out that reporting of wheel event coordinates to the host on
Jacob Nevins [Sat, 20 Dec 2008 19:02:09 +0000 (19:02 +0000)]
"Derek" points out that reporting of wheel event coordinates to the host on
Windows was relative to the screen origin, not the window origin.

[originally from svn r8375]

15 years agoCosmetic: clarify that term->xterm_mouse is not boolean, etc.
Jacob Nevins [Sat, 20 Dec 2008 18:52:09 +0000 (18:52 +0000)]
Cosmetic: clarify that term->xterm_mouse is not boolean, etc.

[originally from svn r8374]

15 years agogtk_selection_clear_targets() does not exist on GTK 1, so ifdef it.
Simon Tatham [Wed, 3 Dec 2008 00:06:38 +0000 (00:06 +0000)]
gtk_selection_clear_targets() does not exist on GTK 1, so ifdef it.

[originally from svn r8371]

15 years agoAt some point recently, Unix PuTTY started suffering build errors
Simon Tatham [Tue, 2 Dec 2008 18:18:32 +0000 (18:18 +0000)]
At some point recently, Unix PuTTY started suffering build errors
using -DNO_GSSAPI. Move some ifdefs around so it stops.

[originally from svn r8370]

15 years agoAdd missing call to gtk_selection_clear_targets(), without which the
Simon Tatham [Mon, 1 Dec 2008 23:03:11 +0000 (23:03 +0000)]
Add missing call to gtk_selection_clear_targets(), without which the
list of selection targets offered by GTK PuTTY/pterm grows an extra
copy of each of the three supported text formats every time the user
makes a selection!

[originally from svn r8364]

15 years agoChange the Unix version of Ssh_gss_name to be a gss_name_t rather than
Ben Harris [Mon, 1 Dec 2008 21:18:29 +0000 (21:18 +0000)]
Change the Unix version of Ssh_gss_name to be a gss_name_t rather than
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]

15 years agoAs far as I can tell, "simple@putty.projects.tartarus.org" has always been
Jacob Nevins [Sun, 30 Nov 2008 21:35:33 +0000 (21:35 +0000)]
As far as I can tell, "simple@putty.projects.tartarus.org" has always been
implemented as a channel request, not a global one. Change documentation to
match implementation.

[originally from svn r8355]

15 years agoMove the code that reads the Windows clipboard into a trivial
Simon Tatham [Fri, 28 Nov 2008 18:28:23 +0000 (18:28 +0000)]
Move the code that reads the Windows clipboard into a trivial
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]

15 years agoJacob correctly points out that I accidentally lost a clearing of
Ben Harris [Wed, 26 Nov 2008 14:30:58 +0000 (14:30 +0000)]
Jacob correctly points out that I accidentally lost a clearing of
s->gss_sndtok in r8326.  I'm not sure it was strictly necessary, since
even if there's no send token, gss_init_sec_context() is meant to explicitly
make it empty, but it wasn't an intentional change.

[originally from svn r8337]
[r8326 == 81dafd906eb4817393fabe4d408c16de8586de8b]

15 years agoFix a Windows-specific GSS crash introduced in r8326, which (again) apparently
Jacob Nevins [Wed, 26 Nov 2008 14:11:49 +0000 (14:11 +0000)]
Fix a Windows-specific GSS crash introduced in r8326, which (again) apparently
triggers in failure cases. Patch by Iain Patterson.

[originally from svn r8336]
[r8326 == 81dafd906eb4817393fabe4d408c16de8586de8b]

15 years agoMitigation for VU#958563: When using a CBC-mode server-to-client cipher
Ben Harris [Wed, 26 Nov 2008 12:49:25 +0000 (12:49 +0000)]
Mitigation for VU#958563:  When using a CBC-mode server-to-client cipher
under SSH-2, don't risk looking at the length field of an incoming packet
until we've successfully MAC'ed the packet.

This requires a change to the MAC mechanics so that we can calculate MACs
incrementally, and output a MAC for the packet so far while still being
able to add more data to the packet later.

[originally from svn r8334]

15 years agoNow that we use real gss_buffer_ts, there's no need to muck about with
Ben Harris [Tue, 25 Nov 2008 22:11:17 +0000 (22:11 +0000)]
Now that we use real gss_buffer_ts, there's no need to muck about with
casts when passing them to GSS-API functions.  Removing them makes the code
more readable and allows better type-checking.

[originally from svn r8333]

15 years agoHave wingss.c include <windows.h> by way of putty.h rather than
Simon Tatham [Tue, 25 Nov 2008 18:54:05 +0000 (18:54 +0000)]
Have wingss.c include <windows.h> by way of putty.h rather than
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]

15 years agoFixed a PuTTYtel startup assertion failure introduced in r8305.
Jacob Nevins [Tue, 25 Nov 2008 18:43:52 +0000 (18:43 +0000)]
Fixed a PuTTYtel startup assertion failure introduced in r8305.
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 == ca6fc3a4daf51166a15693feffc967bee9e3f59a]

15 years agoI think sshgss.h needs puttyps.h for Ssh_gss_buf. Makes it compile for Windows,
Jacob Nevins [Mon, 24 Nov 2008 23:56:55 +0000 (23:56 +0000)]
I think sshgss.h needs puttyps.h for Ssh_gss_buf. Makes it compile for Windows,
anyway.

[originally from svn r8327]

15 years agoChange how we handle the Ssh_gss_buf type. Previously, we defined it
Ben Harris [Mon, 24 Nov 2008 23:44:55 +0000 (23:44 +0000)]
Change how we handle the Ssh_gss_buf type.  Previously, we defined it
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]

15 years agoTreat lines starting with '#' as comments in PSFTP.
Jacob Nevins [Mon, 24 Nov 2008 18:19:55 +0000 (18:19 +0000)]
Treat lines starting with '#' as comments in PSFTP.

[originally from svn r8325]

15 years agoNew option to allow use of the local OS username for login to the remote side
Jacob Nevins [Mon, 24 Nov 2008 17:51:42 +0000 (17:51 +0000)]
New option to allow use of the local OS username for login to the remote side
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 == de5dd9d65c0e2ffa9bc9cab038bf0be77efbf820]

15 years agoMake key file import robust against a PPK file that has somehow lost its
Jacob Nevins [Sun, 23 Nov 2008 20:11:12 +0000 (20:11 +0000)]
Make key file import robust against a PPK file that has somehow lost its
final newline. From Debian bug #414784, based on a patch by Justin Pryzby.

[originally from svn r8323]

15 years agoAutoconfiscate GSS-API support, including support for manually disabling it.
Ben Harris [Sat, 22 Nov 2008 22:49:27 +0000 (22:49 +0000)]
Autoconfiscate GSS-API support, including support for manually disabling it.

[originally from svn r8318]

15 years agoChanges to make this compile on Solaris 9: use <gssapi/gssapi.h> rather
Ben Harris [Sat, 22 Nov 2008 22:06:42 +0000 (22:06 +0000)]
Changes to make this compile on Solaris 9: use <gssapi/gssapi.h> rather
than <gssapi/gssapi_krb5.h> and provide the OID for Kerberos 5 ourselves
(since it's a known constant).  I'm not sure this actually works on Solaris
yet, mind.

[originally from svn r8317]

15 years agoRevamp of the local X11 connection code. We now parse X display
Simon Tatham [Mon, 17 Nov 2008 18:38:09 +0000 (18:38 +0000)]
Revamp of the local X11 connection code. We now parse X display
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]

15 years agoAvoid freeing the backend in notify_remote_exit(), since that's
Simon Tatham [Mon, 17 Nov 2008 18:36:27 +0000 (18:36 +0000)]
Avoid freeing the backend in notify_remote_exit(), since that's
called from within a backend function which will expect its own
backend pointer to still be valid on return. Instead, move all the
real functionality of notify_remote_exit() out into a GTK idle
function.

[originally from svn r8304]

15 years agoIn SSH packet logging mode, log SSH-2 packet sequence numbers, in
Simon Tatham [Tue, 11 Nov 2008 07:47:27 +0000 (07:47 +0000)]
In SSH packet logging mode, log SSH-2 packet sequence numbers, in
both directions. We had a bug report yesterday about a Cisco router
sending SSH2_MSG_UNIMPLEMENTED and it wasn't clear for which packet;
logging the sequence numbers should make such problems much easier
to diagnose.

(In fact this logging fix wouldn't have helped in yesterday's case,
because the router also didn't bother to fill in the sequence number
field in the SSH2_MSG_UNIMPLEMENTED packet! This is a precautionary
measure against the next one of these problems.)

[originally from svn r8295]

15 years agoImplement sk_addr_dup().
Simon Tatham [Sat, 8 Nov 2008 16:58:55 +0000 (16:58 +0000)]
Implement sk_addr_dup().

[originally from svn r8294]

15 years agoMove out of the SockAddr structure the mutable fields "ai" and
Simon Tatham [Sat, 8 Nov 2008 16:45:45 +0000 (16:45 +0000)]
Move out of the SockAddr structure the mutable fields "ai" and
"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]

15 years agoPatch from Iain Patterson: fix crash on Windows when GSSAPI auth is attempted
Jacob Nevins [Fri, 17 Oct 2008 20:55:08 +0000 (20:55 +0000)]
Patch from Iain Patterson: fix crash on Windows when GSSAPI auth is attempted
but fails for some reason (such as not having a tgt for the server's realm).

[originally from svn r8210]

15 years agoRejig the Translation panel controls and documentation to remove the emphasis
Jacob Nevins [Mon, 13 Oct 2008 22:34:57 +0000 (22:34 +0000)]
Rejig the Translation panel controls and documentation to remove the emphasis
on received data. Experiment and suggestion suggest that the character set
configuration applies equally to keystrokes sent to the server, or at least
that that's close enough to being true that we should document it as a first
approximation.

[originally from svn r8209]

15 years agoErroneously invisible index term.
Jacob Nevins [Mon, 13 Oct 2008 22:16:25 +0000 (22:16 +0000)]
Erroneously invisible index term.

[originally from svn r8208]

15 years agoFix inability to save session on Unix when ~/.putty doesn't exist introduced
Jacob Nevins [Sun, 12 Oct 2008 11:32:23 +0000 (11:32 +0000)]
Fix inability to save session on Unix when ~/.putty doesn't exist introduced
in r7934.

[originally from svn r8204]
[r7934 == 087adb167e9e25a09839ca802707c11565468e11]

15 years agoAhem. Make sure I NULL out any unused pointer field that I later
Simon Tatham [Wed, 8 Oct 2008 18:09:56 +0000 (18:09 +0000)]
Ahem. Make sure I NULL out any unused pointer field that I later
plan to free if it isn't NULL.

[originally from svn r8202]

15 years agosshrsa.c now obeys the RFC793 Robustness Principle when it comes to
Simon Tatham [Tue, 7 Oct 2008 17:48:59 +0000 (17:48 +0000)]
sshrsa.c now obeys the RFC793 Robustness Principle when it comes to
the ordering of the primes in a fully specified RSA private key:
when the key format typically has p > q, it will always output p > q
but be willing to tolerate p < q on input. (Inspired by seeing an
OpenSSH-format key file in the wild which had p < q, which I've
never seen before; I suspect a third-party application incautiously
generating the format.)

[originally from svn r8201]

15 years agoThe "server refused our key" bit didn't link to the public-key auth chapter,
Jacob Nevins [Thu, 25 Sep 2008 19:00:42 +0000 (19:00 +0000)]
The "server refused our key" bit didn't link to the public-key auth chapter,
which has some hints on server-side setup. Now it does.

[originally from svn r8194]

15 years agoDocument that PSFTP's "open" command allows an optional port number.
Jacob Nevins [Sun, 21 Sep 2008 20:39:21 +0000 (20:39 +0000)]
Document that PSFTP's "open" command allows an optional port number.

[originally from svn r8193]

15 years agoFurther correction: on reflection, after examining all the call
Simon Tatham [Tue, 16 Sep 2008 22:56:08 +0000 (22:56 +0000)]
Further correction: on reflection, after examining all the call
sites for uint64_make(), we _shouldn't_ attempt to preserve high
bits in the low-order argument; it turns out not to be what the call
sites want.

[originally from svn r8185]

15 years ago64-bit cleanliness in int64.c. This is all hideous; really I ought
Simon Tatham [Tue, 16 Sep 2008 18:21:33 +0000 (18:21 +0000)]
64-bit cleanliness in int64.c. This is all hideous; really I ought
to be trying harder to find a 32-bit type rather than making a
uint64 structure out of two potentially 64-bit unsigned longs. And
really I ought to be using the C99 64-bit integers anyway if they're
available. But this should do for the moment.

[originally from svn r8184]

15 years agoTim Kosse points out that xfer_upload_gotpkt can return a negative
Simon Tatham [Tue, 16 Sep 2008 12:48:38 +0000 (12:48 +0000)]
Tim Kosse points out that xfer_upload_gotpkt can return a negative
value on error, so the check at the call site shouldn't test for
exactly zero.

[originally from svn r8183]

15 years agoSome Windows keymaps, it turns out, don't translate the key
Simon Tatham [Sun, 14 Sep 2008 15:11:42 +0000 (15:11 +0000)]
Some Windows keymaps, it turns out, don't translate the key
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]

15 years agoIf we got a keyboard-interactive INFO_REQUEST with an "instruction" string but
Jacob Nevins [Tue, 9 Sep 2008 20:36:40 +0000 (20:36 +0000)]
If we got a keyboard-interactive INFO_REQUEST with an "instruction" string but
no actual prompts, we weren't displaying the former, which was wrong. We
should now (although I haven't found a server to test it against).

[originally from svn r8172]

15 years agoFix punctuation in the large print statements. In particular, one
Simon Tatham [Mon, 1 Sep 2008 17:56:20 +0000 (17:56 +0000)]
Fix punctuation in the large print statements. In particular, one
semicolon which crept in in r8138 was causing a lot of the "make
install" implementation to be missing from Makefile.gtk.

[originally from svn r8159]
[r8138 == de5dd9d65c0e2ffa9bc9cab038bf0be77efbf820]

15 years agoGood grief. When I originally wrote the local proxy code two years
Simon Tatham [Sun, 31 Aug 2008 21:45:39 +0000 (21:45 +0000)]
Good grief. When I originally wrote the local proxy code two years
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]

15 years agoColin Watson points out an apparently erroneous cast: Ssh_gss_name
Simon Tatham [Sun, 31 Aug 2008 19:18:17 +0000 (19:18 +0000)]
Colin Watson points out an apparently erroneous cast: Ssh_gss_name
and gss_name_t are supposed to be congruent types, so a pointer to
one should never be cast to a non-indirect instance of the other.

[originally from svn r8157]

15 years agoMinor code cleanup: I just happened to be looking at this file for
Simon Tatham [Sun, 31 Aug 2008 17:12:37 +0000 (17:12 +0000)]
Minor code cleanup: I just happened to be looking at this file for
other reasons and I noticed that the list of TELOPTs is given twice
and hence needs to be kept in sync. Replace with my now-standard
second-order-macro approach which allows the list to be maintained
in only one place.

[originally from svn r8156]

15 years agoFix for portfwd-addr-family: on Unix, when a tunnel is specified as "Auto"
Jacob Nevins [Wed, 20 Aug 2008 22:21:04 +0000 (22:21 +0000)]
Fix for portfwd-addr-family: on Unix, when a tunnel is specified as "Auto"
(rather than IPv4 or IPv6-only; this is the default), try to open up listening
sockets on both address families, rather than (unhelpfully) just IPv6. (And
don't open one if the other can't be bound, in a nod to CVE-2008-1483.)
Based on a patch from Ben A L Jemmett.

[originally from svn r8150]
[this svn revision also touched putty-wishlist]

15 years agoJames has made the URL syntax for http://svn.tartarus.org/ more
Simon Tatham [Mon, 11 Aug 2008 18:03:34 +0000 (18:03 +0000)]
James has made the URL syntax for http://svn.tartarus.org/ more
sensible. Update the FAQ link to it.

[originally from svn r8142]

15 years agoLink to pTerm, an SSH client for the iPhone that's apparently based on PuTTY.
Jacob Nevins [Sun, 10 Aug 2008 14:28:51 +0000 (14:28 +0000)]
Link to pTerm, an SSH client for the iPhone that's apparently based on PuTTY.
Attempts at damage limitation from the name similarity with pterm.

Also try to refresh the ports section of the FAQ a bit.

[originally from svn r8139]
[this svn revision also touched putty-website]

15 years agoInitial commit of GSSAPI Kerberos support.
Owen Dunn [Sun, 10 Aug 2008 13:10:31 +0000 (13:10 +0000)]
Initial commit of GSSAPI Kerberos support.

[originally from svn r8138]

15 years agowww.winputty.com
Jacob Nevins [Sat, 19 Jul 2008 16:41:52 +0000 (16:41 +0000)]
www.winputty.com

[originally from svn r8121]
[this svn revision also touched putty-website,putty-wishlist]

15 years agoAnother, subtly different version string for ssh2-bug-maxpkt, from Tim Kosse.
Jacob Nevins [Fri, 18 Jul 2008 21:39:49 +0000 (21:39 +0000)]
Another, subtly different version string for ssh2-bug-maxpkt, from Tim Kosse.

[originally from svn r8120]
[this svn revision also touched putty-wishlist]

15 years agoAt least one version of gcc won't figure out that "clear" is never
Simon Tatham [Thu, 17 Jul 2008 17:01:27 +0000 (17:01 +0000)]
At least one version of gcc won't figure out that "clear" is never
read uninitialised (because the only circumstance under which it
isn't initialised is when "update" is FALSE, in which case it isn't
read either). Placate it.

[originally from svn r8119]

15 years agoReplicate r8110 from PSFTP to PSCP, and various other minor patchery from
Jacob Nevins [Fri, 11 Jul 2008 19:24:56 +0000 (19:24 +0000)]
Replicate r8110 from PSFTP to PSCP, and various other minor patchery from
Roger Picton.

[originally from svn r8117]
[r8110 == e01df1da9f41aa7554e99708c117241c59fcef3d]

15 years agoPatch from Tim Kosse: check back->exitcode() in both loops on
Simon Tatham [Sun, 6 Jul 2008 12:24:56 +0000 (12:24 +0000)]
Patch from Tim Kosse: check back->exitcode() in both loops on
ssh_sftp_loop_iteration(), not just one. Fixes exiting on a negative
response to the host key confirmation prompt on Windows (because
winsftp.c doesn't have the equivalent of uxsftp.c's no_fds_ok); on
Unix it worked already but gave a suboptimal error message, which is
fixed too by this patch.

[originally from svn r8110]

15 years agoPatch from Tim Kosse to make PSFTP remember to close files on
Simon Tatham [Sun, 6 Jul 2008 12:11:34 +0000 (12:11 +0000)]
Patch from Tim Kosse to make PSFTP remember to close files on
various error conditions.

[originally from svn r8109]

16 years agoPatch from Aidan Kehoe to extend the Arabic shaping code to Urdu and
Simon Tatham [Sun, 29 Jun 2008 09:14:26 +0000 (09:14 +0000)]
Patch from Aidan Kehoe to extend the Arabic shaping code to Urdu and
Persian, by adding some additional Unicode code points to the
shapetypes[] array.

[originally from svn r8097]

16 years agoCosmetic: use `appname' in more places, so that Unix PuTTYtel announces itself
Jacob Nevins [Sun, 15 Jun 2008 13:26:08 +0000 (13:26 +0000)]
Cosmetic: use `appname' in more places, so that Unix PuTTYtel announces itself
thus.

[originally from svn r8075]

16 years agoExorcise beeps from the Colours pane in Gtk.
Jacob Nevins [Sun, 15 Jun 2008 12:39:09 +0000 (12:39 +0000)]
Exorcise beeps from the Colours pane in Gtk.

The colour list box beeped at the user whenever it found that
something other than exactly one colour was selected. This seems to
happen implicitly in Gtk when the pane is changed. In Gtk1, this gave
you a beep whenever you left the Colours dialog after having selected
a colour from the list; in Gtk2, you additionally got a beep _every_
time you subsequently re-entered the Colours dialog (for reasons I
haven't investigated). Windows was unaffected.

Also, in Gtk (unlike Windows), it's possible for the user to go back
to the state where no items in the list box are selected at all.

For these reasons, stop beeping at the user, and instead blank the RGB
edit boxes as a hint that edits to them would be futile. (Really we
should be disabling them entirely, but the cross-platform edit
controls aren't up to that yet.)

[originally from svn r8074]

16 years agoWork around a bug in early versions of GTK (which I'm still forced
Simon Tatham [Wed, 11 Jun 2008 18:03:35 +0000 (18:03 +0000)]
Work around a bug in early versions of GTK (which I'm still forced
to use, gah) in which the "model" argument to
gtk_tree_selection_get_selected_rows() couldn't be NULL.

[originally from svn r8069]

16 years agoManfred Schwarb points out that scroll wheel support stopped working
Simon Tatham [Tue, 10 Jun 2008 20:18:23 +0000 (20:18 +0000)]
Manfred Schwarb points out that scroll wheel support stopped working
with the switch to GTK2. This turns out to be because, where GTK1
represented the scroll wheel as mouse buttons 4 and 5 and generated
GdkEventButton when it was moved, GTK2 has moved wheel actions out
into a new event type GdkEventScroll which we were not handling. Now
we do, so scroll wheel support should be back in place.

[originally from svn r8063]

16 years agoManfred Schwarb also mentions that Alt+drag is captured by at least
Simon Tatham [Sat, 7 Jun 2008 16:30:45 +0000 (16:30 +0000)]
Manfred Schwarb also mentions that Alt+drag is captured by at least
one well known window manager (KDE's); document that Shift+Alt+drag
is worth trying as a workaround.

[originally from svn r8059]

16 years agoManfred Schwarb points out that the README contains an outdated
Simon Tatham [Sat, 7 Jun 2008 16:25:46 +0000 (16:25 +0000)]
Manfred Schwarb points out that the README contains an outdated
comment about GTK 2 not being supported.

[originally from svn r8058]

16 years agoAdd an include statement to Makefile.gtk that imports if present a
Simon Tatham [Thu, 5 Jun 2008 17:06:39 +0000 (17:06 +0000)]
Add an include statement to Makefile.gtk that imports if present a
file called Makefile.local. This means that if you're compiling on a
platform that needs COMPAT definitions, you can put them in a local
file and not have to type them on the command line every time.

[originally from svn r8045]

16 years agoAt last, merge the putty-gtk2 branch back into the trunk!
Simon Tatham [Wed, 4 Jun 2008 23:05:48 +0000 (23:05 +0000)]
At last, merge the putty-gtk2 branch back into the trunk!

[originally from svn r8037]

16 years agoMerged from trunk up to r8033.
Simon Tatham [Sun, 1 Jun 2008 11:19:06 +0000 (11:19 +0000)]
Merged from trunk up to r8033.

[originally from svn r8034]
[r8033 == e81a8cf795825bc133678d33c4e2ca4585edf811]

16 years agoBeen meaning to do this for years: introduce a configuration option
Simon Tatham [Sun, 1 Jun 2008 11:16:32 +0000 (11:16 +0000)]
Been meaning to do this for years: introduce a configuration option
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]

16 years agoRe-jig the combo box handling ifdefs so that we can compile with GTK
Simon Tatham [Sat, 31 May 2008 19:23:45 +0000 (19:23 +0000)]
Re-jig the combo box handling ifdefs so that we can compile with GTK
versions >= 2.0 (when the new list boxes came in) but < 2.4 (when
the new combo boxes came in). Since some combo boxes are handled
using the old list-box code, this means that the two lots of code
can both be compiled in at once in some situations!

[originally from svn r8031]

16 years agoApparently Vista's printf-like functions don't support %n by default.
Jacob Nevins [Sat, 31 May 2008 17:22:29 +0000 (17:22 +0000)]
Apparently Vista's printf-like functions don't support %n by default.
We could explicitly re-enable %n, but we only use it in one place, so take
the path of least resistance and remove that single instance. This stops
dupvprintf() getting stuck in a loop (a behaviour that's caused by a workaround
for a broken libc).

<http://msdn.microsoft.com/en-us/library/ms175782(VS.80).aspx>

[originally from svn r8030]

16 years agoUpdate for `win-randseed-location' fix.
Jacob Nevins [Sat, 31 May 2008 17:11:16 +0000 (17:11 +0000)]
Update for `win-randseed-location' fix.

[originally from svn r8029]

16 years agoOn some systems, strncpy is a macro, and putting preprocessor
Simon Tatham [Sat, 31 May 2008 13:29:32 +0000 (13:29 +0000)]
On some systems, strncpy is a macro, and putting preprocessor
directives in the middle of a macro invocation appears to be frowned
on. Irritating, but there we go.

[originally from svn r8026]

16 years agoMerge from trunk up to r8020.
Simon Tatham [Wed, 28 May 2008 19:28:17 +0000 (19:28 +0000)]
Merge from trunk up to r8020.

[originally from svn r8021]
[r8020 == 14d825d42fb4d36f671ea4d05c6d254f7796c7e5]

16 years agoOS X Leopard, it turns out, has a new and exciting strategy for
Simon Tatham [Wed, 28 May 2008 19:23:57 +0000 (19:23 +0000)]
OS X Leopard, it turns out, has a new and exciting strategy for
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]

16 years agoUpdate the FAQ to state more explicitly that PuTTY contains no
Simon Tatham [Wed, 21 May 2008 09:48:40 +0000 (09:48 +0000)]
Update the FAQ to state more explicitly that PuTTY contains no
OpenSSL code.

[originally from svn r8004]

16 years agoMention Vista.
Jacob Nevins [Sat, 26 Apr 2008 18:45:06 +0000 (18:45 +0000)]
Mention Vista.

[originally from svn r7993]

16 years agoOops; prevent further segfault during setup, which apparently only
Simon Tatham [Mon, 14 Apr 2008 18:00:57 +0000 (18:00 +0000)]
Oops; prevent further segfault during setup, which apparently only
show up when building without debugging...

[originally from svn r7991]

16 years agoPrevent assertion failure in the case where the user manipulates the
Simon Tatham [Mon, 14 Apr 2008 17:57:45 +0000 (17:57 +0000)]
Prevent assertion failure in the case where the user manipulates the
filter checkboxes to filter the currently selected font out of the
family list and then does something in one of the other list boxes
or the size edit box.

[originally from svn r7990]

16 years agoJust noticed that selecting "client:Bitstream Vera Sans Mono 10" in
Simon Tatham [Sun, 13 Apr 2008 07:48:10 +0000 (07:48 +0000)]
Just noticed that selecting "client:Bitstream Vera Sans Mono 10" in
the font config box and then invoking the unifontsel causes the box
to come up empty rather than populated with that font. Turns out
that I completely forgot to have pangofont_canonify_fontname()
return the flags word, ahem.

[originally from svn r7988]

16 years agoUtterly hideous new approach to extracting install-sh from the
Simon Tatham [Fri, 11 Apr 2008 13:28:38 +0000 (13:28 +0000)]
Utterly hideous new approach to extracting install-sh from the
autoconf/automake edifice, since my previous approach of guessing
its pathname turns out not to work on at least one kind of system.

[originally from svn r7984]

16 years agoIn the new unified font handling, my strategy so far for combining
Simon Tatham [Sat, 5 Apr 2008 13:37:20 +0000 (13:37 +0000)]
In the new unified font handling, my strategy so far for combining
client- and server-side fonts into a single namespace was mainly to
hope there would naturally be no collisions, and to provide
disambiguating "client:" and "server:" prefixes for manual use in
emergencies.

Jacob points out, however, that his system not only has a namespace
clash but worse still the clash is at the name "fixed", which is our
default font! So, modify my namespace policy to use the
disambiguating prefixes everywhere by default, and use _unprefixed_
names only if the user types one in by hand.

In particular, I've changed the keys used to store font names in
Unix saved session files. Font names read from the new keys will be
passed straight to the new unifont framework; font names read from
the old keys will have "server:" prepended. So any existing
configuration file for GTK1 PuTTY should now work reliably in GTK2
PuTTY and select the same font, even if that font is one on which
your system (rather, your client+server combination) has a font
namespace clash.

[originally from svn r7973]

16 years agoEnsure the new `depth' and `treepath' structure fields in struct
Simon Tatham [Sat, 5 Apr 2008 12:53:32 +0000 (12:53 +0000)]
Ensure the new `depth' and `treepath' structure fields in struct
selparam are always properly initialised.

[originally from svn r7972]

16 years agoFix misleading parameter name.
Jacob Nevins [Sat, 5 Apr 2008 12:26:49 +0000 (12:26 +0000)]
Fix misleading parameter name.

[originally from svn r7970]

16 years agogtk_combo_box_get_active_text didn't appear until GTK 2.6; bodge
Simon Tatham [Fri, 4 Apr 2008 12:23:29 +0000 (12:23 +0000)]
gtk_combo_box_get_active_text didn't appear until GTK 2.6; bodge
around it for earlier versions of GTK 2.

[originally from svn r7969]

16 years agoFix the jarring change of window size on expanding the SSH branch of
Simon Tatham [Fri, 4 Apr 2008 11:37:06 +0000 (11:37 +0000)]
Fix the jarring change of window size on expanding the SSH branch of
the configuration tree.

[originally from svn r7968]

16 years agoI apparently missed out a piece of code when doing the new GTK2 list
Simon Tatham [Fri, 4 Apr 2008 11:02:26 +0000 (11:02 +0000)]
I apparently missed out a piece of code when doing the new GTK2 list
box: shortcut activations for list boxes are missing.

That's the last thing on the to-do list. We're now ready to merge
back to the trunk, given only some final testing!

[originally from svn r7967]

16 years agoReinstate all the GTK1-specific code under ifdefs, and verify that
Simon Tatham [Fri, 4 Apr 2008 10:56:26 +0000 (10:56 +0000)]
Reinstate all the GTK1-specific code under ifdefs, and verify that
we can now build and run successfully using both GTK1 and GTK2 by
giving appropriate options to make. (Specifically, to override the
default of GTK2 in favour of GTK1, "make GTK_CONFIG=gtk-config".)

[originally from svn r7966]