faq.but

大名鼎鼎的远程登录软件putty的Symbian版源码

\S{faq-outofmem}{Question} After trying to establish an SSH 2
connection, PuTTY says \q{Out of memory} and dies.

If this happens just while the connection is starting up, this often
indicates that for some reason the client and server have failed to
establish a session encryption key. Somehow, they have performed
calculations that should have given each of them the same key, but
have ended up with different keys; so data encrypted by one and
decrypted by the other looks like random garbage.

This causes an \q{out of memory} error because the first encrypted
data PuTTY expects to see is the length of an SSH message. Normally
this will be something well under 100 bytes. If the decryption has
failed, PuTTY will see a completely random length in the region of
two \e{gigabytes}, and will try to allocate enough memory to store
this non-existent message. This will immediately lead to it thinking
it doesn't have enough memory, and panicking.

If this happens to you, it is quite likely to still be a PuTTY bug
and you should report it (although it might be a bug in your SSH
server instead); but it doesn't necessarily mean you've actually run
out of memory.

\S{faq-outofmem2}{Question} When attempting a file transfer, either
PSCP or PSFTP says \q{Out of memory} and dies.

This is almost always caused by your login scripts on the server
generating output. PSCP or PSFTP will receive that output when they
were expecting to see the start of a file transfer protocol, and
they will attempt to interpret the output as file-transfer protocol.
This will usually lead to an \q{out of memory} error for much the
same reasons as given in \k{faq-outofmem}.

This is a setup problem in your account on your server, \e{not} a
PSCP/PSFTP bug. Your login scripts should \e{never} generate output
during non-interactive sessions; secure file transfer is not the
only form of remote access that will break if they do.

On Unix, a simple fix is to ensure that all the parts of your login
script that might generate output are in \c{.profile} (if you use a
Bourne shell derivative) or \c{.login} (if you use a C shell).
Putting them in more general files such as \c{.bashrc} or \c{.cshrc}
is liable to lead to problems.

\S{faq-psftp-slow}{Question} PSFTP transfers files much slower than PSCP.

The throughput of PSFTP 0.54 should be much better than 0.53b and
prior; we've added code to the SFTP backend to queue several blocks
of data rather than waiting for an acknowledgement for each. (The
SCP backend did not suffer from this performance issue because SCP
is a much simpler protocol.)

\S{faq-bce}{Question} When I run full-colour applications, I see
areas of black space where colour ought to be.

You almost certainly need to enable the \q{Use background colour to
erase screen} setting in the Terminal panel. Note that if you do
this in mid-session, in versions before 0.54, it may not take effect
until you reset the terminal (see \k{faq-resetterm}).

\S{faq-resetterm}{Question} When I change some terminal settings,
nothing happens.

Some of the terminal options (notably Auto Wrap and
background-colour screen erase) actually represent the \e{default}
setting, rather than the currently active setting. The server can
send sequences that modify these options in mid-session, but when
the terminal is reset (by server action, or by you choosing \q{Reset
Terminal} from the System menu) the defaults are restored.

In versions 0.53b and prior, if you change one of these options in
the middle of a session, you will find that the change does not
immediately take effect. It will only take effect once you reset
the terminal.

In version 0.54, the behaviour has changed - changes to these
settings take effect immediately.

\S{faq-idleout}{Question} My PuTTY sessions unexpectedly close after
they are idle for a while.

Some types of firewall, and almost any router doing Network Address
Translation (NAT, also known as IP masquerading), will forget about
a connection through them if the connection does nothing for too
long. This will cause the connection to be rudely cut off when
contact is resumed.

You can try to combat this by telling PuTTY to send \e{keepalives}:
packets of data which have no effect on the actual session, but
which reassure the router or firewall that the network connection is
still active and worth remembering about.

Keepalives don't solve everything, unfortunately; although they
cause greater robustness against this sort of router, they can also
cause a \e{loss} of robustness against network dropouts. See
\k{config-keepalive} in the documentation for more discussion of
this.

\S{faq-timeout}{Question} PuTTY's network connections time out too
quickly when network connectivity is temporarily lost.

This is a Windows problem, not a PuTTY problem. The timeout value
can't be set on per application or per session basis. To increase
the TCP timeout globally, you need to tinker with the Registry.

On Windows 95, 98 or ME, the registry key you need to create or
change is

\c HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\VxD\
\c   MSTCP\MaxDataRetries

(it must be of type DWORD in Win95, or String in Win98/ME).
(See MS Knowledge Base article
\W{http://support.microsoft.com/default.aspx?scid=kb;en-us;158474}{158474}
for more information.)

On Windows NT or 2000, the registry key is

\c HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\
\c   Parameters\TcpMaxDataRetransmissions

and it must be of type DWORD.
(See MS Knowledge Base article
\W{http://support.microsoft.com/default.aspx?scid=kb;en-us;120642}{120642}
for more information.)

Set the key's value to something like 10. This will cause Windows to
try harder to keep connections alive instead of abandoning them.

\S{faq-puttyputty}{Question} When I \cw{cat} a binary file, I get
`PuTTYPuTTYPuTTY' on my command line.

Don't do that, then.

This is designed behaviour; when PuTTY receives the character
Control-E from the remote server, it interprets it as a request to
identify itself, and so it sends back the string \q{\cw{PuTTY}} as
if that string had been entered at the keyboard. Control-E should
only be sent by programs that are prepared to deal with the
response. Writing a binary file to your terminal is likely to output
many Control-E characters, and cause this behaviour. Don't do it.
It's a bad plan.

To mitigate the effects, you could configure the answerback string
to be empty (see \k{config-answerback}); but writing binary files to
your terminal is likely to cause various other unpleasant behaviour,
so this is only a small remedy.

\S{faq-wintitle}{Question} When I \cw{cat} a binary file, my window
title changes to a nonsense string.

Don't do that, then.

It is designed behaviour that PuTTY should have the ability to
adjust the window title on instructions from the server. Normally
the control sequence that does this should only be sent
deliberately, by programs that know what they are doing and intend
to put meaningful text in the window title. Writing a binary file to
your terminal runs the risk of sending the same control sequence by
accident, and cause unexpected changes in the window title. Don't do
it.

\S{faq-password-fails}{Question} My keyboard stops working once
PuTTY displays the password prompt.

No, it doesn't. PuTTY just doesn't display the password you type, so
that someone looking at your screen can't see what it is.

Unlike the Windows login prompts, PuTTY doesn't display the password
as a row of asterisks either. This is so that someone looking at
your screen can't even tell how \e{long} your password is, which
might be valuable information.

\S{faq-keyboard}{Question} One or more function keys don't do what I
expected in a server-side application.

If you've already tried all the relevant options in the PuTTY
Keyboard panel, you may need to mail the PuTTY maintainers and ask.

It is \e{not} usually helpful just to tell us which application,
which server operating system, and which key isn't working; in order
to replicate the problem we would need to have a copy of every
operating system, and every application, that anyone has ever
complained about.

PuTTY responds to function key presses by sending a sequence of
control characters to the server. If a function key isn't doing what
you expect, it's likely that the character sequence your application
is expecting to receive is not the same as the one PuTTY is sending.
Therefore what we really need to know is \e{what} sequence the
application is expecting.

The simplest way to investigate this is to find some other terminal
environment, in which that function key \e{does} work; and then
investigate what sequence the function key is sending in that
situation. One reasonably easy way to do this on a Unix system is to
type the command \c{cat}, and then press the function key. This is
likely to produce output of the form \c{^[[11~}. You can also do
this in PuTTY, to find out what sequence the function key is
producing in that. Then you can mail the PuTTY maintainers and tell
us \q{I wanted the F1 key to send \c{^[[11~}, but instead it's
sending \c{^[OP}, can this be done?}, or something similar.

You should still read the
\W{http://www.chiark.greenend.org.uk/~sgtatham/putty/feedback.html}{Feedback
page} on the PuTTY website (also provided as \k{feedback} in the
manual), and follow the guidelines contained in that.

\S{faq-openssh-bad-openssl}{Question} Since my SSH server was upgraded
to OpenSSH 3.1p1/3.4p1, I can no longer connect with PuTTY.

There is a known problem when OpenSSH has been built against an
incorrect version of OpenSSL; the quick workaround is to configure
PuTTY to use SSH protocol 2 and the Blowfish cipher.

For more details and OpenSSH patches, see
\W{http://bugzilla.mindrot.org/show_bug.cgi?id=138}{bug 138} in the
OpenSSH BTS.

This is not a PuTTY-specific problem; if you try to connect with
another client you'll likely have similar problems. (Although PuTTY's
default cipher differs from many other clients.)

\e{OpenSSH 3.1p1:} configurations known to be broken (and symptoms):

\b SSH 2 with AES cipher (PuTTY says "Assertion failed! Expression:
(len & 15) == 0" in sshaes.c, or "Out of memory", or crashes)

\b SSH 2 with 3DES (PuTTY says "Incorrect MAC received on packet")

\b SSH 1 with Blowfish (PuTTY says "Incorrect CRC received on
packet")

\b SSH 1 with 3DES

\e{OpenSSH 3.4p1:} as of 3.4p1, only the problem with SSH 1 and
Blowfish remains. Rebuild your server, apply the patch linked to from
bug 138 above, or use another cipher (e.g., 3DES) instead.

\e{Other versions:} we occasionally get reports of the same symptom
and workarounds with older versions of OpenSSH, although it's not
clear the underlying cause is the same.

\S{faq-ssh2key-ssh1conn}{Question} Why do I see "Couldn't load private
key from ..."? Why can PuTTYgen load my key but not PuTTY?

It's likely that you've generated an SSH protocol 2 key with PuTTYgen,
but you're trying to use it in an SSH 1 connection. SSH1 and SSH2 keys
have different formats, and (at least in 0.52) PuTTY's reporting of a
key in the wrong format isn't optimal.

To connect using SSH 2 to a server that supports both versions, you
need to change the configuration from the default (see \k{faq-ssh2}).

\S{faq-rh8-utf8}{Question} When I'm connected to a Red Hat Linux 8.0
system, some characters don't display properly.

A common complaint is that hyphens in man pages show up as a-acute.

With release 8.0, Red Hat appear to have made UTF-8 the default
character set. There appears to be no way for terminal emulators such
as PuTTY to know this (as far as we know, the appropriate escape
sequence to switch into UTF-8 mode isn't sent).

A fix is to configure sessions to RH8 systems to use UTF-8
translation - see \k{config-charset} in the documentation. (Note that
if you use \q{Change Settings}, changes may not take place immediately
- see \k{faq-resetterm}.)

If you really want to change the character set used by the server, the
right place is \c{/etc/sysconfig/i18n}, but this shouldn't be
necessary.

\S{faq-screen}{Question} Since I upgraded to PuTTY 0.54, the
scrollback has stopped working when I run \c{screen}.

PuTTY's terminal emulator has always had the policy that when the
\q{alternate screen} is in use, nothing is added to the scrollback.
This is because the usual sorts of programs which use the alternate
screen are things like text editors, which tend to scroll back and
forth in the same document a lot; so (a) they would fill up the
scrollback with a large amount of unhelpfully disordered text, and
(b) they contain their \e{own} method for the user to scroll back to
the bit they were interested in. We have generally found this policy
to do the Right Thing in almost all situations.