strange char echo

[Bryan Christ]

does anyone know what might cause characters to be echoed to the screen
even though noecho() was called at startup?



_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Scott Furry]

Two schools of thought....
a) hardware
b) software

For a) assuming some kind of external comms - do you have some form of
loop back?
I'm going to say this is probably unlikely and would be special case.
However, I have to throw it out.

For b) when you did you initial build (assuming you built from source)
did you config using the flag "with_sp_funcs"?
 From the source see the configure file (line 9350 sets the flag
NCURSES_SP_FUNCS) and lib_echo.c (NCURSES_SP_FUNCS is used to enable the
build of the noecho function at line 77).

Regards,
Scott




Bryan Christ wrote:

> does anyone know what might cause characters to be echoed to the screen
> even though noecho() was called at startup?
>
>
>
> _______________________________________________
> Bug-ncurses mailing list
> Bug-ncurses@...
> http://lists.gnu.org/mailman/listinfo/bug-ncurses
>
>  

_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Bryan Christ]

a) is unlikely because the terminal is Gnome Terminal.
b) i do not see a configure option with_sp_funcs (or anything like it)
and when i look at the symbol table i do see echo and noecho.

On Thu, 2009-11-05 at 20:26 +0000, Scott Furry wrote:

> Two schools of thought....
> a) hardware
> b) software
>
> For a) assuming some kind of external comms - do you have some form of
> loop back?
> I'm going to say this is probably unlikely and would be special case.
> However, I have to throw it out.
>
> For b) when you did you initial build (assuming you built from source)
> did you config using the flag "with_sp_funcs"?
>  From the source see the configure file (line 9350 sets the flag
> NCURSES_SP_FUNCS) and lib_echo.c (NCURSES_SP_FUNCS is used to enable the
> build of the noecho function at line 77).
>
> Regards,
> Scott
>
>
>
>
> Bryan Christ wrote:
> > does anyone know what might cause characters to be echoed to the screen
> > even though noecho() was called at startup?
> >
> >
> >
> > _______________________________________________
> > Bug-ncurses mailing list
> > Bug-ncurses@...
> > http://lists.gnu.org/mailman/listinfo/bug-ncurses
> >
> >  
>
>
>
> _______________________________________________
> Bug-ncurses mailing list
> Bug-ncurses@...
> http://lists.gnu.org/mailman/listinfo/bug-ncurses

_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Scott Furry]

Apologies...wasn't sure how you were using ncurses. Had to throw out the
'corner case'

My bad...should have directed you up a few more lines.
If you look up at line 9341 in 'configure' (in the ncurses directory)
you will see a reference to..
--enable-sp-funcs or --disable-sp-funcs
This also will cause NCURSES_SP_FUNS to be enabled/disabled.

To see all the available options with a build, you can either type
./configure --help | more      <-- pipe the output
- or -
./configure --help >> somefile.txt      <-- my prefered option as then I
can look at all the available options in a text editor

If this isn't the case then I may have to defer to the 'gurus' on this one.

Good Luck.
Scott

Bryan Christ wrote:

> a) is unlikely because the terminal is Gnome Terminal.
> b) i do not see a configure option with_sp_funcs (or anything like it)
> and when i look at the symbol table i do see echo and noecho.
>
> On Thu, 2009-11-05 at 20:26 +0000, Scott Furry wrote:
>  
>> Two schools of thought....
>> a) hardware
>> b) software
>>
>> For a) assuming some kind of external comms - do you have some form of
>> loop back?
>> I'm going to say this is probably unlikely and would be special case.
>> However, I have to throw it out.
>>
>> For b) when you did you initial build (assuming you built from source)
>> did you config using the flag "with_sp_funcs"?
>>  From the source see the configure file (line 9350 sets the flag
>> NCURSES_SP_FUNCS) and lib_echo.c (NCURSES_SP_FUNCS is used to enable the
>> build of the noecho function at line 77).
>>
>> Regards,
>> Scott
>>
>>
>> Bryan Christ wrote:
>>    
>>> does anyone know what might cause characters to be echoed to the screen
>>> even though noecho() was called at startup?
>>>      

_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Bryan Christ]

Scott,

My reply on item "b" was basically that that the following yields
nothing:

./configure --help |grep sp_funcs
./configure --help |grep sp-funcs

opening configure in nano and searching for sp_funcs and sp-funcs finds
nothing either.

On Thu, 2009-11-05 at 21:15 +0000, Scott Furry wrote:

> Apologies...wasn't sure how you were using ncurses. Had to throw out the
> 'corner case'
>
> My bad...should have directed you up a few more lines.
> If you look up at line 9341 in 'configure' (in the ncurses directory)
> you will see a reference to..
> --enable-sp-funcs or --disable-sp-funcs
> This also will cause NCURSES_SP_FUNS to be enabled/disabled.
>
> To see all the available options with a build, you can either type
> ./configure --help | more      <-- pipe the output
> - or -
> ./configure --help >> somefile.txt      <-- my prefered option as then I
> can look at all the available options in a text editor
>
> If this isn't the case then I may have to defer to the 'gurus' on this one.
>
> Good Luck.
> Scott
>
> Bryan Christ wrote:
> > a) is unlikely because the terminal is Gnome Terminal.
> > b) i do not see a configure option with_sp_funcs (or anything like it)
> > and when i look at the symbol table i do see echo and noecho.
> >
> > On Thu, 2009-11-05 at 20:26 +0000, Scott Furry wrote:
> >  
> >> Two schools of thought....
> >> a) hardware
> >> b) software
> >>
> >> For a) assuming some kind of external comms - do you have some form of
> >> loop back?
> >> I'm going to say this is probably unlikely and would be special case.
> >> However, I have to throw it out.
> >>
> >> For b) when you did you initial build (assuming you built from source)
> >> did you config using the flag "with_sp_funcs"?
> >>  From the source see the configure file (line 9350 sets the flag
> >> NCURSES_SP_FUNCS) and lib_echo.c (NCURSES_SP_FUNCS is used to enable the
> >> build of the noecho function at line 77).
> >>
> >> Regards,
> >> Scott
> >>
> >>
> >> Bryan Christ wrote:
> >>    
> >>> does anyone know what might cause characters to be echoed to the screen
> >>> even though noecho() was called at startup?
> >>>      
>
>
> _______________________________________________
> Bug-ncurses mailing list
> Bug-ncurses@...
> http://lists.gnu.org/mailman/listinfo/bug-ncurses

_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Scott Furry]

Brian,

Try Ctrl+Shift+underscore to use the "Go To Line" command.
In Nano, this will allow you to enter line number and column number to
advance the cursor.

'configure' line 9341: "# Check whether --enable-sp-funcs or
--disable-sp-funcs was given."

There is also the config.log file which should show the configure
settings used to build ncurses.
It should show lines like:

configure:9338: checking if you want to build with experimental SCREEN
extensions
configure:9348: result: no

...which covers the test for --enable-sp-funcs or --disable-sp-funcs.

BTW, I'm working with the latest ncurses that includes all patches upto
the end of October.

Regards,
Scott

Bryan Christ wrote:

> Scott,
>
> My reply on item "b" was basically that that the following yields
> nothing:
>
> ./configure --help |grep sp_funcs
> ./configure --help |grep sp-funcs
>
> opening configure in nano and searching for sp_funcs and sp-funcs finds
> nothing either.
>
> On Thu, 2009-11-05 at 21:15 +0000, Scott Furry wrote:
>  
>> Apologies...wasn't sure how you were using ncurses. Had to throw out the
>> 'corner case'
>>
>> My bad...should have directed you up a few more lines.
>> If you look up at line 9341 in 'configure' (in the ncurses directory)
>> you will see a reference to..
>> --enable-sp-funcs or --disable-sp-funcs
>> This also will cause NCURSES_SP_FUNS to be enabled/disabled.
>>
>> To see all the available options with a build, you can either type
>> ./configure --help | more      <-- pipe the output
>> - or -
>> ./configure --help >> somefile.txt      <-- my prefered option as then I
>> can look at all the available options in a text editor
>>
>> If this isn't the case then I may have to defer to the 'gurus' on this one.
>>
>> Good Luck.
>> Scott
>>
>> Bryan Christ wrote:
>>    
>>> a) is unlikely because the terminal is Gnome Terminal.
>>> b) i do not see a configure option with_sp_funcs (or anything like it)
>>> and when i look at the symbol table i do see echo and noecho.
>>>
>>> On Thu, 2009-11-05 at 20:26 +0000, Scott Furry wrote:
>>>  
>>>      
>>>> Two schools of thought....
>>>> a) hardware
>>>> b) software
>>>>
>>>> For a) assuming some kind of external comms - do you have some form of
>>>> loop back?
>>>> I'm going to say this is probably unlikely and would be special case.
>>>> However, I have to throw it out.
>>>>
>>>> For b) when you did you initial build (assuming you built from source)
>>>> did you config using the flag "with_sp_funcs"?
>>>>  From the source see the configure file (line 9350 sets the flag
>>>> NCURSES_SP_FUNCS) and lib_echo.c (NCURSES_SP_FUNCS is used to enable the
>>>> build of the noecho function at line 77).
>>>>
>>>> Regards,
>>>> Scott
>>>>
>>>>
>>>> Bryan Christ wrote:
>>>>    
>>>>        
>>>>> does anyone know what might cause characters to be echoed to the screen
>>>>> even though noecho() was called at startup?

_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Thomas Dickey-2]

On Thu, 5 Nov 2009, Scott Furry wrote:

> Two schools of thought....
> a) hardware
> b) software
>
> For a) assuming some kind of external comms - do you have some form of loop
> back?
> I'm going to say this is probably unlikely and would be special case.
> However, I have to throw it out.
>
> For b) when you did you initial build (assuming you built from source) did
> you config using the flag "with_sp_funcs"?
> From the source see the configure file (line 9350 sets the flag
> NCURSES_SP_FUNCS) and lib_echo.c (NCURSES_SP_FUNCS is used to enable the
> build of the noecho function at line 77).

You're being misled by a trick with the ifdef's and define's: there's
always a "noecho", but the ifdef's provide a different name for the
function which accepts a SCREEN* parameter.  Since the bulk of the logic
in most cases is in the latter, the plain "noecho" looks like it's just
an add-on.

> Scott
>
>
>
>
> Bryan Christ wrote:
>> does anyone know what might cause characters to be echoed to the screen
>> even though noecho() was called at startup?
>>
>>
>>
>> _______________________________________________
>> Bug-ncurses mailing list
>> Bug-ncurses@...
>> http://lists.gnu.org/mailman/listinfo/bug-ncurses
>>
>>
>
>
>
> _______________________________________________
> Bug-ncurses mailing list
> Bug-ncurses@...
> http://lists.gnu.org/mailman/listinfo/bug-ncurses
>

--
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net


_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Thomas Dickey-2]

On Thu, 5 Nov 2009, Bryan Christ wrote:

> does anyone know what might cause characters to be echoed to the screen
> even though noecho() was called at startup?

aside from a new bug report?

It's possible that the ioctl to disable echoing is failing, e.g., if
stdout is redirected.

(n)curses disables echo at initialization, and simulates the echo/noecho
behavior.

--
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net


_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Bryan Christ]

ya.  i considered that they may just be stubs.

On Thu, 2009-11-05 at 23:53 +0000, Thomas Dickey wrote:

> On Thu, 5 Nov 2009, Scott Furry wrote:
>
> > Two schools of thought....
> > a) hardware
> > b) software
> >
> > For a) assuming some kind of external comms - do you have some form of loop
> > back?
> > I'm going to say this is probably unlikely and would be special case.
> > However, I have to throw it out.
> >
> > For b) when you did you initial build (assuming you built from source) did
> > you config using the flag "with_sp_funcs"?
> > From the source see the configure file (line 9350 sets the flag
> > NCURSES_SP_FUNCS) and lib_echo.c (NCURSES_SP_FUNCS is used to enable the
> > build of the noecho function at line 77).
>
> You're being misled by a trick with the ifdef's and define's: there's
> always a "noecho", but the ifdef's provide a different name for the
> function which accepts a SCREEN* parameter.  Since the bulk of the logic
> in most cases is in the latter, the plain "noecho" looks like it's just
> an add-on.
>
> > Scott
> >
> >
> >
> >
> > Bryan Christ wrote:
> >> does anyone know what might cause characters to be echoed to the screen
> >> even though noecho() was called at startup?
> >>
> >>
> >>
> >> _______________________________________________
> >> Bug-ncurses mailing list
> >> Bug-ncurses@...
> >> http://lists.gnu.org/mailman/listinfo/bug-ncurses
> >>
> >>
> >
> >
> >
> > _______________________________________________
> > Bug-ncurses mailing list
> > Bug-ncurses@...
> > http://lists.gnu.org/mailman/listinfo/bug-ncurses
> >
>

_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Bryan Christ]

Now that I think of it, I've only seen this problem when I try to run my
software on a remote systems via SSH.  I just tried to reproduce it
locally and couldn't.  The problem manifests 9 out of 10 times on remote
connections.  When I tried it 10 times locally, 100% pass.

Thoughts?

On Thu, 2009-11-05 at 23:56 +0000, Thomas Dickey wrote:

> On Thu, 5 Nov 2009, Bryan Christ wrote:
>
> > does anyone know what might cause characters to be echoed to the screen
> > even though noecho() was called at startup?
>
> aside from a new bug report?
>
> It's possible that the ioctl to disable echoing is failing, e.g., if
> stdout is redirected.
>
> (n)curses disables echo at initialization, and simulates the echo/noecho
> behavior.
>

_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Thomas Dickey-2]

On Fri, 6 Nov 2009, Bryan Christ wrote:

> Now that I think of it, I've only seen this problem when I try to run my
> software on a remote systems via SSH.  I just tried to reproduce it
> locally and couldn't.  The problem manifests 9 out of 10 times on remote
> connections.  When I tried it 10 times locally, 100% pass.
>
> Thoughts?

just guessing - ssh may not be passing the ioctl information along
fast enough to make it look as if it succeeded.

I'd investigate it by building the application with ncurses' tracing
turned on (normally a feature of the debug-library...).

>
> On Thu, 2009-11-05 at 23:56 +0000, Thomas Dickey wrote:
>> On Thu, 5 Nov 2009, Bryan Christ wrote:
>>
>>> does anyone know what might cause characters to be echoed to the screen
>>> even though noecho() was called at startup?
>>
>> aside from a new bug report?
>>
>> It's possible that the ioctl to disable echoing is failing, e.g., if
>> stdout is redirected.
>>
>> (n)curses disables echo at initialization, and simulates the echo/noecho
>> behavior.
>>
>
>

--
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net


_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Bryan Christ]

I added these lines to my application (after initscr) and the problem
went away...

tcgetattr(STDIN_FILENO,&term_desc);
term_desc.c_lflag &= ~(ECHO);
tcsetattr(STDIN_FILENO,TCSADRAIN,&term_desc);

The version of ncurses I am using is:
ncurses = 5.7 (20081102)

Any known bugs of this nature with this build?

On Fri, 2009-11-06 at 21:36 +0000, Thomas Dickey wrote:

> On Fri, 6 Nov 2009, Bryan Christ wrote:
>
> > Now that I think of it, I've only seen this problem when I try to run my
> > software on a remote systems via SSH.  I just tried to reproduce it
> > locally and couldn't.  The problem manifests 9 out of 10 times on remote
> > connections.  When I tried it 10 times locally, 100% pass.
> >
> > Thoughts?
>
> just guessing - ssh may not be passing the ioctl information along
> fast enough to make it look as if it succeeded.
>
> I'd investigate it by building the application with ncurses' tracing
> turned on (normally a feature of the debug-library...).
>
> >
> > On Thu, 2009-11-05 at 23:56 +0000, Thomas Dickey wrote:
> >> On Thu, 5 Nov 2009, Bryan Christ wrote:
> >>
> >>> does anyone know what might cause characters to be echoed to the screen
> >>> even though noecho() was called at startup?
> >>
> >> aside from a new bug report?
> >>
> >> It's possible that the ioctl to disable echoing is failing, e.g., if
> >> stdout is redirected.
> >>
> >> (n)curses disables echo at initialization, and simulates the echo/noecho
> >> behavior.
> >>
> >
> >
>

_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Thomas Dickey-2]

On Tue, 10 Nov 2009, Bryan Christ wrote:

> I added these lines to my application (after initscr) and the problem
> went away...
>
> tcgetattr(STDIN_FILENO,&term_desc);
> term_desc.c_lflag &= ~(ECHO);
> tcsetattr(STDIN_FILENO,TCSADRAIN,&term_desc);
>
> The version of ncurses I am using is:
> ncurses = 5.7 (20081102)
>
> Any known bugs of this nature with this build?

not that I recall - low-level I/O was not touched much for a while.

(this year, I've made a number of changes, and broken initialization 2-3
times)

>
> On Fri, 2009-11-06 at 21:36 +0000, Thomas Dickey wrote:
>> On Fri, 6 Nov 2009, Bryan Christ wrote:
>>
>>> Now that I think of it, I've only seen this problem when I try to run my
>>> software on a remote systems via SSH.  I just tried to reproduce it
>>> locally and couldn't.  The problem manifests 9 out of 10 times on remote
>>> connections.  When I tried it 10 times locally, 100% pass.
>>>
>>> Thoughts?
>>
>> just guessing - ssh may not be passing the ioctl information along
>> fast enough to make it look as if it succeeded.
>>
>> I'd investigate it by building the application with ncurses' tracing
>> turned on (normally a feature of the debug-library...).
>>
>>>
>>> On Thu, 2009-11-05 at 23:56 +0000, Thomas Dickey wrote:
>>>> On Thu, 5 Nov 2009, Bryan Christ wrote:
>>>>
>>>>> does anyone know what might cause characters to be echoed to the screen
>>>>> even though noecho() was called at startup?
>>>>
>>>> aside from a new bug report?
>>>>
>>>> It's possible that the ioctl to disable echoing is failing, e.g., if
>>>> stdout is redirected.
>>>>
>>>> (n)curses disables echo at initialization, and simulates the echo/noecho
>>>> behavior.
>>>>
>>>
>>>
>>
>
>

--
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net


_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses