]> asedeno.scripts.mit.edu Git - PuTTY.git/log
PuTTY.git
18 years agoPlatform-independent support for zooming around session list by typing
Owen Dunn [Wed, 21 Sep 2005 14:53:32 +0000 (14:53 +0000)]
Platform-independent support for zooming around session list by typing
in the session name box.  Thwarted on Unix by GTK not automatically
scrolling the listbox to the selected item, but we can fix that in
the platform-specific side later.

[originally from svn r6337]

18 years agoMention that sessions can be saved from "Change Settings" in using.but.
Jacob Nevins [Mon, 19 Sep 2005 14:34:01 +0000 (14:34 +0000)]
Mention that sessions can be saved from "Change Settings" in using.but.

[originally from svn r6332]

18 years agoAIX 5.1 has <utmpx.h> but no updwtmpx(). Treat the latter as a reason to
Ben Harris [Wed, 14 Sep 2005 11:00:34 +0000 (11:00 +0000)]
AIX 5.1 has <utmpx.h> but no updwtmpx().  Treat the latter as a reason to
OMIT_UTMP.  Reported by Mike Protts.

[originally from svn r6311]

18 years agoSome systems (HP-UX) don't yet have <sys/select.h>, putting select() in
Ben Harris [Wed, 14 Sep 2005 10:53:39 +0000 (10:53 +0000)]
Some systems (HP-UX) don't yet have <sys/select.h>, putting select() in
<sys/time.h>.  Cope with this.  Where <sys/select.h> _is_ available, though,
use it (since it's where POSIX puts select()).  Problem reported by Mike
Protts.

[originally from svn r6310]

18 years agoWhen asked to malloc zero bytes, malloc one byte instead. This ensures
Ben Harris [Tue, 13 Sep 2005 20:17:10 +0000 (20:17 +0000)]
When asked to malloc zero bytes, malloc one byte instead.  This ensures
that we get a unique pointer rather than NULL (which ANSI C otherwise permits).
Problem pointed out by Mike Protts.

[originally from svn r6308]

18 years agoThe ANSI-C constant FILENAME_MAX is ludicrously small on some systems.
Ben Harris [Tue, 13 Sep 2005 20:08:25 +0000 (20:08 +0000)]
The ANSI-C constant FILENAME_MAX is ludicrously small on some systems.
Use the POSIX PATH_MAX if it exists, and fall back to 1024 otherwise.
We should really allocate filenames dynamically if PATH_MAX isn't defined.

[originally from svn r6307]

18 years agoSome hosts don't have TIOCSCTTY. Don't try to use it on them.
Ben Harris [Tue, 13 Sep 2005 19:57:37 +0000 (19:57 +0000)]
Some hosts don't have TIOCSCTTY.  Don't try to use it on them.
Patch from Mike Protts.

[originally from svn r6306]

18 years agoA couple of places in sk_newlistener were using AF_INET6 even with NO_IPV6.
Ben Harris [Tue, 13 Sep 2005 19:54:01 +0000 (19:54 +0000)]
A couple of places in sk_newlistener were using AF_INET6 even with NO_IPV6.
Correct them.

[originally from svn r6305]

18 years agoNothing seems to use the "int64" type, and it apparently causes conflicts
Ben Harris [Tue, 13 Sep 2005 19:24:35 +0000 (19:24 +0000)]
Nothing seems to use the "int64" type, and it apparently causes conflicts
on some platform, so remove it.  Thanks to Mike Protts for spotting this.

[originally from svn r6304]

18 years agoApparently the OS X port is stalled.
Jacob Nevins [Tue, 13 Sep 2005 14:12:16 +0000 (14:12 +0000)]
Apparently the OS X port is stalled.

[originally from svn r6302]
[this svn revision also touched putty-wishlist]

18 years agoWhen the first element in a preference list was unrecognised, PuTTY would
Ben Harris [Mon, 12 Sep 2005 15:45:29 +0000 (15:45 +0000)]
When the first element in a preference list was unrecognised, PuTTY would
hang when reading it because strtok() kept getting the full list passed in.
Fix this, and add an assert() for an assumption documented in a comment while
I'm in the area.

[originally from svn r6294]

18 years agoMention relationship between terminal types, keyboard sequences, and
Ben Harris [Sat, 10 Sep 2005 17:36:52 +0000 (17:36 +0000)]
Mention relationship between terminal types, keyboard sequences, and
termcap/terminfo.  Suggested by Joachim Durchholz.

[originally from svn r6285]

18 years agoImplement hmac-sha1-96. It's RECOMMENDED in the current transport draft,
Ben Harris [Sat, 10 Sep 2005 16:19:53 +0000 (16:19 +0000)]
Implement hmac-sha1-96.  It's RECOMMENDED in the current transport draft,
and we don't have any strong reason not to implement it, for all that it's
rather pointless.

[originally from svn r6284]

18 years agoAdd support for diffie-hellman-group-exchange-sha256. Tested against a
Ben Harris [Sun, 4 Sep 2005 14:53:39 +0000 (14:53 +0000)]
Add support for diffie-hellman-group-exchange-sha256.  Tested against a
patched OpenSSH server.  This is controlled by the same user settings
as diffie-hellman-group-exchange-sha1, which may not be optimal, especially
given that they're both referred to as dh-gex-sha1 in saved sessions.

[originally from svn r6272]

18 years agoMention xp-wont-run.
Ben Harris [Sat, 3 Sep 2005 17:29:28 +0000 (17:29 +0000)]
Mention xp-wont-run.

[originally from svn r6264]

18 years agoNow that my arcfour-fixes draft has been approved by the IESG, use the
Ben Harris [Sat, 3 Sep 2005 17:03:49 +0000 (17:03 +0000)]
Now that my arcfour-fixes draft has been approved by the IESG, use the
IANA-assigned names for its modes.

[originally from svn r6263]

18 years agoRestructure things so that a single entry in the KEX preference list can
Ben Harris [Sat, 3 Sep 2005 13:41:43 +0000 (13:41 +0000)]
Restructure things so that a single entry in the KEX preference list can
correspond to multiple SSH-2 KEX algorithms.  We already do the equivalent
for cipher algorithms.

[originally from svn r6262]

18 years agoCheck ssh->v2_session_id_len _after_ assigning to it.
Jacob Nevins [Fri, 2 Sep 2005 10:51:09 +0000 (10:51 +0000)]
Check ssh->v2_session_id_len _after_ assigning to it.

[originally from svn r6260]

18 years agoDon't bother compiling SHA-256 for now -- I need to think a bit before I use
Ben Harris [Wed, 31 Aug 2005 22:32:05 +0000 (22:32 +0000)]
Don't bother compiling SHA-256 for now -- I need to think a bit before I use
it.

[originally from svn r6253]

18 years agoSHA-256 implementation, for use in future KEX algorithms, in particular
Ben Harris [Wed, 31 Aug 2005 21:48:22 +0000 (21:48 +0000)]
SHA-256 implementation, for use in future KEX algorithms, in particular
diffie-hellman-group-exchange-sha256, which the last DHGEX draft defined.
Code lifted from Simon's "crypto" directory, with changes to make it look
more like sshsh512.c.

[originally from svn r6252]

18 years agoAdd infrastructure for supporting multiple hashes in key exchange.
Ben Harris [Wed, 31 Aug 2005 20:43:06 +0000 (20:43 +0000)]
Add infrastructure for supporting multiple hashes in key exchange.
Nothing very surprising here.

[originally from svn r6251]

18 years agoRename ssh_md5 and ssh_sha1 to ssh_hmac_md5 and ssh_hmac_sha1 respectively.
Ben Harris [Wed, 31 Aug 2005 19:11:19 +0000 (19:11 +0000)]
Rename ssh_md5 and ssh_sha1 to ssh_hmac_md5 and ssh_hmac_sha1 respectively.
This is to make room for a hash abstraction that's likely to want to use
ssh_sha1, at least.

[originally from svn r6249]

18 years agoExplicitly note that "remote command" semantics typically involve the server
Jacob Nevins [Wed, 31 Aug 2005 16:14:21 +0000 (16:14 +0000)]
Explicitly note that "remote command" semantics typically involve the server
closing the connection after the command has executed.

[originally from svn r6246]

18 years agoFurther progress in the direction of variable exchange hashes -- stash the
Ben Harris [Tue, 30 Aug 2005 22:39:35 +0000 (22:39 +0000)]
Further progress in the direction of variable exchange hashes -- stash the
outgoing KEXINIT, and only start generating the exchange hash once we know
which KEX method we're using.

[originally from svn r6242]

18 years agoBeginnings of support for multiple exchange hashes in SSH-2: rather than
Ben Harris [Tue, 30 Aug 2005 20:38:57 +0000 (20:38 +0000)]
Beginnings of support for multiple exchange hashes in SSH-2: rather than
storing a SHA-1 hash of the client and server version strings, store the
strings themselves so we can feed them through the appropriate hash when
we know what it is.

[originally from svn r6241]

18 years agoMore versions of WeOnlyDo have the rekey bug (but they've fixed it now).
Jacob Nevins [Mon, 29 Aug 2005 00:34:03 +0000 (00:34 +0000)]
More versions of WeOnlyDo have the rekey bug (but they've fixed it now).

[originally from svn r6233]

18 years agoI think this should fix various problems with queued incoming data not being
Jacob Nevins [Fri, 26 Aug 2005 21:17:49 +0000 (21:17 +0000)]
I think this should fix various problems with queued incoming data not being
processed and incoming data being processed out of order, which I suspect is
the cause of `ssh1-fwd-trouble' as noted by Gevan Dutton. I'm not able to
test the failure case, but it doesn't seem to have obviously broken anything
in the cases I have tested, anyway.

[originally from svn r6221]

18 years agoAdd support for generating project files for use with Dev-C++, contributed
Ben Harris [Mon, 22 Aug 2005 20:37:13 +0000 (20:37 +0000)]
Add support for generating project files for use with Dev-C++, contributed
by Florian Gaab.

[originally from svn r6201]

18 years agoFlorian Gaab reports that freeSSHd 1.0.7, which claims a "softwareversion"
Ben Harris [Mon, 22 Aug 2005 20:07:46 +0000 (20:07 +0000)]
Florian Gaab reports that freeSSHd 1.0.7, which claims a "softwareversion"
of "WeOnlyDo-1.2.6", mishandles repeat key exchange.  Add it to the list.

[originally from svn r6200]

18 years agoBen Rudiak-Gould points out that we should be using WM_APP as the base for
Jacob Nevins [Wed, 10 Aug 2005 18:31:24 +0000 (18:31 +0000)]
Ben Rudiak-Gould points out that we should be using WM_APP as the base for
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]

18 years agodraft-ietf-secsh-transport-24 says that only "SSH-" at the start of a line
Ben Harris [Sun, 24 Jul 2005 13:46:14 +0000 (13:46 +0000)]
draft-ietf-secsh-transport-24 says that only "SSH-" at the start of a line
marks a version string.  It's a bit vague about the definition of a line,
but I think it's reasonable to assume that they'll end with LF.  Change
do_ssh_init() to ignore "SSH-" anywhere else.  This makes the existing state
machine overkill, so replace it with something a little more readable.

[originally from svn r6138]

18 years agoThis is getting silly; nearly 25% of our mirrors are now in the US.
Jacob Nevins [Sun, 17 Jul 2005 13:37:58 +0000 (13:37 +0000)]
This is getting silly; nearly 25% of our mirrors are now in the US.
Discourage more strongly mirrors in well-served areas in the Feedback section.
Also, duplicate that text on the Mirrors page, along with a request to tell us
the country (since lots of people still don't).

[originally from svn r6109]
[this svn revision also touched putty-website]

18 years agoPatch from Colin Watson: we were sometimes passing stack storage to putenv(),
Jacob Nevins [Fri, 15 Jul 2005 11:47:28 +0000 (11:47 +0000)]
Patch from Colin Watson: we were sometimes passing stack storage to putenv(),
which is Bad (in his case, it caused TERM to end up unset). Use malloc()'d
storage instead.

[originally from svn r6095]

18 years agoMissing argument in MALLOC_LOG version of snrealloc() macro.
Jacob Nevins [Wed, 6 Jul 2005 18:26:41 +0000 (18:26 +0000)]
Missing argument in MALLOC_LOG version of snrealloc() macro.

[originally from svn r6075]

18 years agoPiers Finlayson reports that "DigiSSH_2.0" chokes (in a new and exciting
Ben Harris [Tue, 5 Jul 2005 21:15:23 +0000 (21:15 +0000)]
Piers Finlayson reports that "DigiSSH_2.0" chokes (in a new and exciting
manner) on rekeys.  Add it to the list.

[originally from svn r6067]

19 years agoTry to make it slightly clearer that TIS/CryptoCard are generic, since
Jacob Nevins [Wed, 29 Jun 2005 18:20:17 +0000 (18:20 +0000)]
Try to make it slightly clearer that TIS/CryptoCard are generic, since
SSH:TDGv2 seems to think that TIS is only relevant to its original
application (whose name eludes me at the moment).

[originally from svn r6038]

19 years agoIn local-to-remote copies, abort if fxp_init() fails, rather than ploughing
Ben Harris [Sat, 25 Jun 2005 21:43:09 +0000 (21:43 +0000)]
In local-to-remote copies, abort if fxp_init() fails, rather than ploughing
on and trying to transfer files, which caused a null-pointer deference.

Thanks to Fernando Najera for reporting the bug.

[originally from svn r6021]

19 years agoQuote session name in command line example for robustness.
Ben Harris [Sat, 25 Jun 2005 19:22:13 +0000 (19:22 +0000)]
Quote session name in command line example for robustness.

[originally from svn r6020]

19 years agoImprove IPA representation of "PuTTY", as suggested by John Lunney. My
Ben Harris [Wed, 22 Jun 2005 10:00:09 +0000 (10:00 +0000)]
Improve IPA representation of "PuTTY", as suggested by John Lunney.  My
rationale (as mailed to him):

I think you're right.  I got the pronunciation there from the second edition
of the OED and my Collins dictionary at home, both of which believe that
"pretty" is pronounced /'prItI/, but, at least to me, those two vowels are
different.  Both of them think that /i/ doesn't occur in English words, the
vowel in "beat" being /i:/.  The third edition of the OED, though, adds /i/
as an English vowel in its pronunciation guide, with "happy" as an example
of its use.  I'll update the FAQ following your suggestion.

[originally from svn r5989]

19 years agoFix an apparently-harmless error spotted by Ben Rudiak-Gould:
Jacob Nevins [Tue, 21 Jun 2005 20:13:48 +0000 (20:13 +0000)]
Fix an apparently-harmless error spotted by Ben Rudiak-Gould:
do_ssh2_transport() was returning the wrong value for rekeys after the first.
This apparent error was introduced in r4901, but we can't see any reason for
the change to have been made. If it turns out to be a mistake to revert it,
I'm sure we'll find out.

Here for posterity is Simon's analysis:

| A lot of the return values from do_ssh2_transport appear to be vestigial: it
| used to be that a zero return from do_ssh2_transport meant it had handled the
| packet internally, and a 1 return meant the packet wasn't a transport-layer
| one and needed to pass on to do_ssh2_authconn. Since r4901, however, the
| layer discrimination is done based on the message type ranges, and the only
| remaining dependency on the return value from do_ssh2_transport is a special
| case in ssh2_protocol which detects the first 1 return and makes the
| initialisation call to do_ssh2_authconn.
|
| Therefore, the gratuitous 1 return on every key exchange as a result of the
| confusing if statement is simply ignored in ssh2_protocol (because
| ssh->protocol_initial_phase_done is already TRUE). So the remaining question
| was, why does the _lack_ of that 1 return not cause a problem, if the if's
| sense is indeed reversed?
|
| The answer is that 1 is still returned, just not by the crReturn inside the
| if statement. It's returned by the next crReturn, just after
| wait_for_rekey(). Which suggests that in fact, the if statement has the
| correct sense, but the crReturn inside it has the wrong value - it should be
| returning _zero_, to indicate that every NEWKEYS after the first one is
| uninteresting to the authconn code, and on the very first run through that
| doesn't happen and the NEWKEYS gets all the way to the crReturn(1) later on.

[originally from svn r5986]
[r4901 == a4ba0268389027f2985da7a45ddbf7b84104266d]

19 years agoMake the sanity-checks on the size of incoming packets much stricter. We now
Ben Harris [Mon, 20 Jun 2005 13:56:30 +0000 (13:56 +0000)]
Make the sanity-checks on the size of incoming packets much stricter.  We now
enforce the following:

* Packet must have at least one byte of payload and four bytes of padding.
* Total packet length must not exceed 35000 bytes compressed.
* Total packet length including length field must be a multiple of cipher
  block size (or eight bytes).

The feebleness of our old checks was noticed by Ben Rudiak-Gould.

[originally from svn r5981]

19 years agoA major purpose of PuTTY's memory-allocation functions is to succeed or die
Ben Harris [Sun, 19 Jun 2005 14:17:24 +0000 (14:17 +0000)]
A major purpose of PuTTY's memory-allocation functions is to succeed or die
trying, so there's no need to check their return values for NULL.  Spotted
by Ben Rudiak-Gould.

[originally from svn r5978]

19 years agoMove comment about ECHO and LINE input modes to a more sensible position.
Ben Harris [Sun, 19 Jun 2005 13:57:50 +0000 (13:57 +0000)]
Move comment about ECHO and LINE input modes to a more sensible position.
Spotted by Ben Rudiak-Gould.

[originally from svn r5976]

19 years agoDetection of "auth-agent@openssh.com" was too liberal. Spotted by
Jacob Nevins [Tue, 14 Jun 2005 23:20:42 +0000 (23:20 +0000)]
Detection of "auth-agent@openssh.com" was too liberal. Spotted by
Ben Rudiak-Gould.

[originally from svn r5962]

19 years agoWe should wait until the Rlogin server indicates that it's happy to receive
Jacob Nevins [Tue, 14 Jun 2005 14:48:17 +0000 (14:48 +0000)]
We should wait until the Rlogin server indicates that it's happy to receive
window-size notifications before we send them.
This clears up a problem where the first password entry always failed.

[originally from svn r5960]

19 years ago"SanskritFritz" points out that digits at the start of RTF pastes were being
Jacob Nevins [Thu, 9 Jun 2005 10:05:29 +0000 (10:05 +0000)]
"SanskritFritz" points out that digits at the start of RTF pastes were being
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]

19 years agoDouble-free on mkdir error, spotted by Brian Hartsock.
Jacob Nevins [Wed, 8 Jun 2005 15:14:47 +0000 (15:14 +0000)]
Double-free on mkdir error, spotted by Brian Hartsock.

[originally from svn r5929]

19 years agoRemove a couple of mistaken references to Telnet in comments.
Jacob Nevins [Sat, 28 May 2005 13:46:04 +0000 (13:46 +0000)]
Remove a couple of mistaken references to Telnet in comments.

[originally from svn r5853]

19 years agoMiscellaneous updates:
Jacob Nevins [Mon, 23 May 2005 12:41:21 +0000 (12:41 +0000)]
Miscellaneous updates:
 - note Visual Foo version requirements
 - note which things are done for you in our source snapshots
 - other tweaks

[originally from svn r5839]

19 years agoFactor out all local SSH disconnections into a new function ssh_disconnect(),
Jacob Nevins [Sat, 21 May 2005 16:49:27 +0000 (16:49 +0000)]
Factor out all local SSH disconnections into a new function ssh_disconnect(),
and add the ability to treat a local disconnection as "unclean" -- notably, if
we can't agree any authentication methods to even try; someone was complaining
that the PuTTY window by default just disappears for no apparent reason in this
circumstance.

Also, use appropriate disconnect codes for those SSH2_MSG_DISCONNECT messages
that we do send.

I don't think I've seriously broken any user-visible behaviour, but the way
that connection-close distinctions are transmitted to the front-end is shaky
(or so it seems to me), so there may be non-ideal changes on some platforms.

[originally from svn r5824]

19 years agoFix documentation of NO_MANIFESTS (oops).
Jacob Nevins [Sat, 21 May 2005 15:09:36 +0000 (15:09 +0000)]
Fix documentation of NO_MANIFESTS (oops).

[originally from svn r5821]

19 years agoAdd NO_MANIFESTS option to Windows build, as the manifests apparently cause
Jacob Nevins [Sat, 21 May 2005 14:35:21 +0000 (14:35 +0000)]
Add NO_MANIFESTS option to Windows build, as the manifests apparently cause
trouble for 64-bit Windows builds.
Also flag the build flags that only apply to Windows.

[originally from svn r5820]

19 years agoUse {Get,Set}WindowLongPtr() instead of {Get,Set}WindowLong() for compatibility
Jacob Nevins [Sat, 21 May 2005 14:16:43 +0000 (14:16 +0000)]
Use {Get,Set}WindowLongPtr() instead of {Get,Set}WindowLong() for compatibility
with 64-bit Windows. Untested on 64-bit, but it doesn't appear to have broken
anything on 32-bit.

[originally from svn r5819]

19 years agoMinor memory leak spotted by Mikhail Kruk.
Jacob Nevins [Fri, 20 May 2005 21:52:07 +0000 (21:52 +0000)]
Minor memory leak spotted by Mikhail Kruk.

[originally from svn r5815]

19 years agoconnection_fatal() should be called after ssh_closing() and other accesses to
Jacob Nevins [Mon, 16 May 2005 14:41:08 +0000 (14:41 +0000)]
connection_fatal() should be called after ssh_closing() and other accesses to
`ssh', since it potentially frees it and potentially doesn't return.
Only affected a couple of rare circumstances.

[originally from svn r5787]

19 years agoDon't try and set up reconfigured port-forwardings if the connection isn't yet
Jacob Nevins [Mon, 16 May 2005 08:31:05 +0000 (08:31 +0000)]
Don't try and set up reconfigured port-forwardings if the connection isn't yet
ready for them. Spotted by Martin Dushkov.

[originally from svn r5786]

19 years agoAdd experimental support for detecting BREAK on input and propagating it as
Ben Harris [Sat, 14 May 2005 22:01:10 +0000 (22:01 +0000)]
Add experimental support for detecting BREAK on input and propagating it as
TS_BRK on output.  This is tested to the extent that other data survive the
escaping performed by PARMRK, at least on my system.  Actual passing on
of BREAK is as-yet untested.

[originally from svn r5779]

19 years agoUse the packet dispatch table to handle USERAUTH_BANNER messages, which should
Jacob Nevins [Thu, 12 May 2005 15:09:35 +0000 (15:09 +0000)]
Use the packet dispatch table to handle USERAUTH_BANNER messages, which should
hopefully solve `drop-banner'. I haven't been able to test the failure case,
but the behaviour with OpenSSH appears no worse.

[originally from svn r5772]
[this svn revision also touched putty-wishlist]

19 years agos/public/private/ spotted by Walter Cleverly.
Jacob Nevins [Mon, 9 May 2005 21:01:11 +0000 (21:01 +0000)]
s/public/private/ spotted by Walter Cleverly.

[originally from svn r5765]

19 years agoMake Makefile.gtk build again on Linux (assume <utmpx.h>).
Jacob Nevins [Mon, 9 May 2005 13:27:51 +0000 (13:27 +0000)]
Make Makefile.gtk build again on Linux (assume <utmpx.h>).

[originally from svn r5764]

19 years agoFix what looks like a cut-and-paste error which was stopping Unix
Simon Tatham [Sun, 8 May 2005 11:47:59 +0000 (11:47 +0000)]
Fix what looks like a cut-and-paste error which was stopping Unix
Plink building on MacOS X.

[originally from svn r5762]

19 years agoCorrect apparent misspelling of `SIOCATMARK'.
Jacob Nevins [Thu, 5 May 2005 22:47:30 +0000 (22:47 +0000)]
Correct apparent misspelling of `SIOCATMARK'.

[originally from svn r5749]

19 years agoImplement `bypass-ssh2-userauth', since from correspondence it sounds like
Jacob Nevins [Thu, 5 May 2005 22:37:54 +0000 (22:37 +0000)]
Implement `bypass-ssh2-userauth', since from correspondence it sounds like
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]

19 years agoAlso blank X authentication data when using SSH-1, and add a comment explaining
Ben Harris [Sat, 30 Apr 2005 17:08:48 +0000 (17:08 +0000)]
Also blank X authentication data when using SSH-1, and add a comment explaining
why this may not be sufficient to hide it.

[originally from svn r5720]

19 years agoX11 authentication data are generally sensitive, so blank them from packet
Ben Harris [Sat, 30 Apr 2005 16:13:11 +0000 (16:13 +0000)]
X11 authentication data are generally sensitive, so blank them from packet
logs by default.

[originally from svn r5719]

19 years agoUnlike the AES and Blowfish code, our implementations of the various DES
Ben Harris [Sat, 30 Apr 2005 14:30:07 +0000 (14:30 +0000)]
Unlike the AES and Blowfish code, our implementations of the various DES
modes of operation all took separate source and destination pointers.  They
were never called with those pointers different, though, so reduce them to
a single pointer like everything else uses.

[originally from svn r5716]

19 years agoFix stupid typo that probably utterly broke SSH-1 support, and caused compiles
Ben Harris [Sat, 30 Apr 2005 14:26:46 +0000 (14:26 +0000)]
Fix stupid typo that probably utterly broke SSH-1 support, and caused compiles
with GCC to fail.  Not sure how it survived long enough to test, really.

[originally from svn r5715]

19 years agoMinimally document 3des-ctr and blowfish-ctr.
Ben Harris [Thu, 28 Apr 2005 09:03:16 +0000 (09:03 +0000)]
Minimally document 3des-ctr and blowfish-ctr.

[originally from svn r5703]

19 years agoRemove comment explaining why 3des-ctr is disabled, since it isn't.
Ben Harris [Thu, 28 Apr 2005 09:00:50 +0000 (09:00 +0000)]
Remove comment explaining why 3des-ctr is disabled, since it isn't.

[originally from svn r5702]

19 years agoEnable blowfish-ctr by default. It's been tested and found working.
Ben Harris [Thu, 28 Apr 2005 08:56:03 +0000 (08:56 +0000)]
Enable blowfish-ctr by default.  It's been tested and found working.
Thanks to der Mouse for help with the testing.

[originally from svn r5701]

19 years agoIn blowfish-ctr mode, increment the counter in the correct order. Thanks to
Ben Harris [Thu, 28 Apr 2005 08:24:46 +0000 (08:24 +0000)]
In blowfish-ctr mode, increment the counter in the correct order.  Thanks to
der Mouse for spotting the same error in my 3des-ctr implementation.

[originally from svn r5700]

19 years agoFix two more stupid bugs in 3des-ctr:
Ben Harris [Thu, 28 Apr 2005 08:21:04 +0000 (08:21 +0000)]
Fix two more stupid bugs in 3des-ctr:
 - We were using the first word of each block of keystream block twice and the
   second not at all.
 - We were incrementing the high-order word of the counter after every block
   rather than the low-order one.

With those fixed, our 3des-ctr implementation interoperates with the one in
Moussh.  Thanks to der Mouse for his help with the testing.

3des-ctr is now enabled by default.

[originally from svn r5699]

19 years agoOn monochrome displays, display the cursor in reverse video so that it's
Ben Harris [Wed, 27 Apr 2005 21:42:51 +0000 (21:42 +0000)]
On monochrome displays, display the cursor in reverse video so that it's
visible on reversed out text.  This only applies to active block cursors for
now.

[originally from svn r5698]

19 years agoTiny manual CSE of previous commit.
Ben Harris [Wed, 27 Apr 2005 21:22:40 +0000 (21:22 +0000)]
Tiny manual CSE of previous commit.

[originally from svn r5697]

19 years agoOn 1bpp displays, ignore colour attributes. This makes pterm minimally useful
Ben Harris [Wed, 27 Apr 2005 21:09:45 +0000 (21:09 +0000)]
On 1bpp displays, ignore colour attributes.  This makes pterm minimally useful
there, though (e.g.) switching to using reverse video for the cursor would
probably also help.  Displays with other silly depths (e.g. 2bpp) aren't
catered for, but I suspect they're rare in the X world.

[originally from svn r5696]

19 years agoMake palette changes use "best match" colours too.
Ben Harris [Wed, 27 Apr 2005 20:30:47 +0000 (20:30 +0000)]
Make palette changes use "best match" colours too.

[originally from svn r5695]

19 years agoRather than checking for <sys/sockio.h>, just include it if our other
Ben Harris [Wed, 27 Apr 2005 15:42:10 +0000 (15:42 +0000)]
Rather than checking for <sys/sockio.h>, just include it if our other
attempts at finding SIOCATMARK have failed.  This removes one of our
Autoconf tests, which is always nice.

[originally from svn r5690]

19 years agoUse the correct key order for 3des-ctr.
Ben Harris [Wed, 27 Apr 2005 12:39:52 +0000 (12:39 +0000)]
Use the correct key order for 3des-ctr.

[originally from svn r5687]

19 years agoIn SSH-2 mode, PuTTY wasn't sending KEXINIT until it had received one from the
Ben Harris [Wed, 27 Apr 2005 11:10:54 +0000 (11:10 +0000)]
In SSH-2 mode, PuTTY wasn't sending KEXINIT until it had received one from the
server, which led to stalemate if the server did the same.  PuTTY now sends
KEXINIT proactively as soon as it's worked out that it's talking SSH-2.

[originally from svn r5685]

19 years agomkunxarc.sh is supposed to be silent, so I'm having it discard the
Simon Tatham [Wed, 27 Apr 2005 08:09:32 +0000 (08:09 +0000)]
mkunxarc.sh is supposed to be silent, so I'm having it discard the
stderr output from mkauto.sh. (I debated discarding it within
mkauto.sh itself, but decided against it.)

[originally from svn r5684]

19 years agoAccording to the termio(7I) on Solaris, OLCUC is overridden by OPOST, so we
Ben Harris [Tue, 26 Apr 2005 00:03:50 +0000 (00:03 +0000)]
According to the termio(7I) on Solaris, OLCUC is overridden by OPOST, so we
should send it with the SSH terminal modes too.

[originally from svn r5680]

19 years agoSquash OPOST locally when we're not in line-editing mode, and propagate
Ben Harris [Mon, 25 Apr 2005 23:57:45 +0000 (23:57 +0000)]
Squash OPOST locally when we're not in line-editing mode, and propagate
the other output flags with SSH.  This means that when I log into a remote
system using Plink and then run "stty -onlcr" I get the expected
stair-stepping.

[originally from svn r5679]

19 years agoUse pututxline() in place of pututline(), since the former is standardised by
Ben Harris [Mon, 25 Apr 2005 23:28:25 +0000 (23:28 +0000)]
Use pututxline() in place of pututline(), since the former is standardised by
X/Open and actually seems to be more common (NetBSD has it).  Also use
updwtmpx() rather than directly writing to the wtmpx file, though more for
reasons of aesthetics than anything practical.

[originally from svn r5678]

19 years agoAsk GDK to give us reasonable approximations if it can't get precisely the
Ben Harris [Mon, 25 Apr 2005 22:46:08 +0000 (22:46 +0000)]
Ask GDK to give us reasonable approximations if it can't get precisely the
colours we asked it for.  This means that I can run pterm on an 8-bit
PseudoColor display even if I have another program running.

[originally from svn r5677]

19 years agoCheck for <sys/sockio.h> and include it in uxnet.c if we find it. It's
Ben Harris [Mon, 25 Apr 2005 18:51:15 +0000 (18:51 +0000)]
Check for <sys/sockio.h> and include it in uxnet.c if we find it.  It's
necessary on Solaris if we want to use SIOCATMARK.  Using sockatmark() might
be preferable, but despite being notionally standard it's missing on
Solaris 9 and Mac OS X 10.3.9, whereas everyone seems to have SIOCATMARK
somewhere.

[originally from svn r5676]

19 years agouxpty.c uses non-X/Open facilities (notably strsignal()), so don't define
Ben Harris [Mon, 25 Apr 2005 17:21:08 +0000 (17:21 +0000)]
uxpty.c uses non-X/Open facilities (notably strsignal()), so don't define
_XOPEN_SOURCE.  We do still need _GNU_SOURCE in order to get grantpt()
on GNU systems.  This allows uxpty.c to compile on NetBSD.

[originally from svn r5675]

19 years agoMove generation of the "configure" script into its own script so that
Ben Harris [Mon, 25 Apr 2005 16:36:43 +0000 (16:36 +0000)]
Move generation of the "configure" script into its own script so that
people who check the code out of Subversion can get it to go.

[originally from svn r5674]

19 years agoAdd a mechanism for using autoconf to detect the quirks of Unix systems
Ben Harris [Mon, 25 Apr 2005 15:55:06 +0000 (15:55 +0000)]
Add a mechanism for using autoconf to detect the quirks of Unix systems
rather than relying on the user to edit the Makefile.  Makefile.gtk
still works as well as it ever did, but now we get a Makefile.in alongside
it.  mkunxarc.sh now relies on autoconf and friends to build the configure
script for the Unix source distribution.

[originally from svn r5673]

19 years agoOn some systems (NetBSD 1.6 and Solaris 9, at least), GCC doesn't understand
Ben Harris [Sun, 24 Apr 2005 14:43:00 +0000 (14:43 +0000)]
On some systems (NetBSD 1.6 and Solaris 9, at least), GCC doesn't understand
the semantics of assert(0) and believes it can return.  Add a gratuitous
exit(1) to convince it that this won't happen, and hence quell a couple of
warnings about variables' being used uninitialised.

[originally from svn r5669]

19 years agoNow that we've got at least some SDCTR modes working (and aes256-ctr is our
Ben Harris [Sat, 23 Apr 2005 16:22:51 +0000 (16:22 +0000)]
Now that we've got at least some SDCTR modes working (and aes256-ctr is our
default preferred cipher), add code to inject SSH_MSG_IGNOREs to randomise
the IV when using CBC-mode ciphers.  Each cipher has a flag to indicate
whether it needs this workaround, and the SSH packet output maze has gained
some extra complexity to implement it.

[originally from svn r5659]

19 years agoStir in process ID at start.
Jacob Nevins [Fri, 22 Apr 2005 15:47:28 +0000 (15:47 +0000)]
Stir in process ID at start.

[originally from svn r5655]

19 years agoChange proxy-dns `Auto' default for SOCKS5 from local DNS to remote DNS.
Jacob Nevins [Fri, 22 Apr 2005 00:00:49 +0000 (00:00 +0000)]
Change proxy-dns `Auto' default for SOCKS5 from local DNS to remote DNS.
SOCKS5 should always be able to do this, and I suspect our not doing so
dates from when the SOCKS proxy types were under a single configuration
option (pre-r3168).

[originally from svn r5654]

19 years agoFirst crack at `terminal-modes' in SSH. PuTTY now sends ERASE by default,
Jacob Nevins [Thu, 21 Apr 2005 13:57:08 +0000 (13:57 +0000)]
First crack at `terminal-modes' in SSH. PuTTY now sends ERASE by default,
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]

19 years agoIfdef out the actual code supporting 3des-ctr and blowfish-ctr, since GCC
Ben Harris [Wed, 20 Apr 2005 22:52:54 +0000 (22:52 +0000)]
Ifdef out the actual code supporting 3des-ctr and blowfish-ctr, since GCC
now notices that it isn't used.

[originally from svn r5652]

19 years agoSimon (accidentally, I think) enabled 3des-ctr and blowfish-ctr. Turn them
Ben Harris [Wed, 20 Apr 2005 22:47:29 +0000 (22:47 +0000)]
Simon (accidentally, I think) enabled 3des-ctr and blowfish-ctr.  Turn them
back off again since they're still untested.

[originally from svn r5651]

19 years agoMinimally document our support of aes*-ctr.
Ben Harris [Wed, 20 Apr 2005 21:29:57 +0000 (21:29 +0000)]
Minimally document our support of aes*-ctr.

[originally from svn r5650]

19 years agoEnable arcfour256-draft-00@putty.projects.tartarus.org, since I've now tested
Ben Harris [Wed, 20 Apr 2005 20:57:00 +0000 (20:57 +0000)]
Enable arcfour256-draft-00@putty.projects.tartarus.org, since I've now tested
it against an independent implementation.

[originally from svn r5649]

19 years agoInvent a way of specifying control characters numerically in ctrlparse():
Jacob Nevins [Tue, 19 Apr 2005 19:18:14 +0000 (19:18 +0000)]
Invent a way of specifying control characters numerically in ctrlparse():
^<27>, ^<0x1B>, ^<033>. (This doesn't tread on any syntax that already had a
non-null behaviour.)

[originally from svn r5647]

19 years agoPull out parsing of ^C style strings from the terminal answerback code to
Jacob Nevins [Tue, 19 Apr 2005 18:58:29 +0000 (18:58 +0000)]
Pull out parsing of ^C style strings from the terminal answerback code to
its own function, since I'll be wanting it for `terminal-modes'.

[originally from svn r5646]

19 years agoRemove "none" from the MACs we offer to support in SSH-2. (It was at the
Jacob Nevins [Tue, 19 Apr 2005 13:27:21 +0000 (13:27 +0000)]
Remove "none" from the MACs we offer to support in SSH-2. (It was at the
end, after the REQUIRED "hmac-sha1".) This has been present since SSH-2
support was introduced (r569).

[originally from svn r5643]
[r569 == 35205e5cb764dca5afaf1e90dc6d485fcc5ed283]