twin-develop

[Twin-dev] ZMODEM between TWIN terminals

[Rick v. R. <ri...@op...>]

Hi,

I've been writing a proposal in the issues section before I knew about
this list.  Let's see if it is alive ;-)

https://github.com/cosmos72/twin/issues/48

What I'm proposing is to make a small extension to TWIN that recognises
ZMODEM sequences and connects a terminal ready to send to one ready to
receive.  Having `lrzsz` installed, enter `sz /etc/motd` in one and `rz`
in the other and the file would be passed.  SSH sessions would be
transparant to this.

ZMODEM is ancient, stems from the BBS modem days.  It is used to mix
downloads and uploads with teletype operations.  SSH does not do such
nifty things.  Also, I believe I would start using it to pass textual
information between terminals, where it can be much nicer than copy/paste.

Interesting use-case?

Did I get the architectural side right?  Recognising a control code in
`server/tty.c`, switch the tty to a ZMODEM transport mode and proxy
bytes through `server/builtin.c` until termination?


Cheers,
 -Rick

Re: [Twin-dev] Mac OS X?

[max <mas...@gm...>]

On 02/25/2012 04:29 AM, Scott Duensing wrote:
> Has anyone built Twin on OS X?  Lion, specifically?  When I run
> "MAKE=gmake ./configure --prefix=/opt/local" things go well until...
> [...]
> creating all the Makefiles...
> mkdir: /clients: Permission denied
> ./scripts/Makefiles.sh: line 66: /clients/Makefile: No such file or directory
> [...]
> Same if I use "make" instead of "gmake".  :-/  Thanks!
> 
> Scott

Hello Scott,

once upon a time, SourceForge had a compile farm with Max OS X servers
too and I fixed twin to compile and run on them too.

Nowadays, I don't have access to any Mac OS X machine, sorry.

Max

[Twin-dev] Mac OS X?

[Scott D. <sc...@ja...>]

Has anyone built Twin on OS X?  Lion, specifically?  When I run
"MAKE=gmake ./configure --prefix=/opt/local" things go well until...

<snip>

...
creating all the Makefiles...
mkdir: /clients: Permission denied
./scripts/Makefiles.sh: line 66: /clients/Makefile: No such file or directory
mkdir: /clients: Permission denied
./scripts/Makefiles.sh: line 66: /clients/mapscrn/Makefile: No such
file or directory
mkdir: /contrib: Permission denied
./scripts/Makefiles.sh: line 66: /contrib/Makefile: No such file or directory
mkdir: /contrib: Permission denied
./scripts/Makefiles.sh: line 66: /contrib/twcd/Makefile: No such file
or directory
mkdir: /docs: Permission denied
./scripts/Makefiles.sh: line 66: /docs/Makefile: No such file or directory
mkdir: /include: Permission denied
./scripts/Makefiles.sh: line 66: /include/Makefile: No such file or directory
mkdir: /include: Permission denied
./scripts/Makefiles.sh: line 66: /include/TT/Makefile: No such file or directory
mkdir: /include: Permission denied
./scripts/Makefiles.sh: line 66: /include/TT/pli/Makefile: No such
file or directory
mkdir: /include: Permission denied
./scripts/Makefiles.sh: line 66: /include/Tutf/Makefile: No such file
or directory
mkdir: /include: Permission denied
./scripts/Makefiles.sh: line 66: /include/Tw/Makefile: No such file or directory
mkdir: /libs: Permission denied
./scripts/Makefiles.sh: line 66: /libs/libTT/HW/Makefile: No such file
or directory
mkdir: /libs: Permission denied
./scripts/Makefiles.sh: line 66: /libs/libTT/Makefile: No such file or directory
mkdir: /libs: Permission denied
./scripts/Makefiles.sh: line 66: /libs/libTutf/Makefile: No such file
or directory
mkdir: /libs: Permission denied
./scripts/Makefiles.sh: line 66: /libs/libTw/Makefile: No such file or directory
mkdir: /libs: Permission denied
./scripts/Makefiles.sh: line 66: /libs/Makefile: No such file or directory
./scripts/Makefiles.sh: line 66: /Makefile: Permission denied
mkdir: /scripts: Permission denied
./scripts/Makefiles.sh: line 66: /scripts/Makefile: No such file or directory
mkdir: /server: Permission denied
./scripts/Makefiles.sh: line 66: /server/extensions/Makefile: No such
file or directory
mkdir: /server: Permission denied
./scripts/Makefiles.sh: line 66: /server/HW/hw_tty_common/Makefile: No
such file or directory
mkdir: /server: Permission denied
./scripts/Makefiles.sh: line 66: /server/HW/hw_tty_linux/Makefile: No
such file or directory
mkdir: /server: Permission denied
./scripts/Makefiles.sh: line 66: /server/HW/Makefile: No such file or directory
mkdir: /server: Permission denied
./scripts/Makefiles.sh: line 66: /server/Makefile: No such file or directory
mkdir: /themes: Permission denied
./scripts/Makefiles.sh: line 66: /themes/hw_gfx/Makefile: No such file
or directory
mkdir: /themes: Permission denied
./scripts/Makefiles.sh: line 66: /themes/Makefile: No such file or directory
configure: creating ./config.status
config.status: creating makeautoconf
config.status: creating conf/conf.auto
config.status: creating include/autoconf.h
config.status: include/autoconf.h is unchanged

Configuration:

  Items in parenthesis (...) will be built as shared libraries/modules

  Compiler:                   gcc
  Compiler flags:             -O2 -fomit-frame-pointer -fno-strict-aliasing
  Compiler features used:     <none>
  Shared libraries support:   Darwin-libtool
  Server features:            (socket) (wm) (wm_rc) (term)
  Server socket driver flags: gz pthreads alien
  Server HW drivers:          (twin) (display) (tty)
  Server HW tty driver flags: lrawkbd twterm termcap
  Libraries:                  (libTutf) (libTw) (libTT)
  TT library HW targets:      (twin) (twin_tunnel) (twin_detunnel) (xml)
  Install path:               /opt/local/bin, /opt/local/lib/twin,
/opt/local/share/twin

to show/change other configuration settings, you can run one of:
'./configure <options>', 'scripts/Configure.sh <options>',
'make menuconfig' or 'make config'

twin is now (hopefully) configured.

As next step, you probably want to run 'gmake'.
oneill:twin-0.6.2 scott$ gmake config
gmake: *** No rule to make target `config'.  Stop.
oneill:twin-0.6.2 scott$ gmake menuconfig
gmake: *** No rule to make target `menuconfig'.  Stop.
oneill:twin-0.6.2 scott$ gmake
You should run './configure [options]'
to configure twin for your system
gmake: *** [all] Error 1

<snip>

Same if I use "make" instead of "gmake".  :-/  Thanks!


Scott

[Twin-dev] Update on Slackware maintainer for Twin

[Bradley D. T. <Bradley@NorthTech.US>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

I've just been designated the new maintainer for Twin.

Sometimes it takes a couple of days, and sometimes it takes much longer
for the new SlackBuild scripts to be approved and released into the
repo, but as soon as that occurs the new installation scripts for
version 0.6.2 will appear at:

http://slackbuilds.org/repository/13.37/system/twin/

Right now it's still at version 0.6.1

It will compile on 32 or 64 bit architectures.

The precompiled binary packages I currently have via gopher (both 0.6.1
and 0.6.2) are for 64 bit architectures (or multi-lib enabled
installations) only.

Once again, they are available for download at:

gopher://gopher.northtech.us/1/software/packages

I'll make another announcement as soon as the the new SBo is available.

Kindest regards,


- -- 
Bradley D. Thornton
Manager Network Services
NorthTech Computer
TEL: +1.310.388.9469  (US)
TEL: +44.203.318.2755 (UK)
TEL: +41.43.508.05.10 (CH)
http://NorthTech.US

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Find this cert at x-hkp://pool.sks-keyservers.net

iQEcBAEBAwAGBQJPSEBjAAoJEE1wgkIhr9j3h/UIAIraheqwMygnW6NfiD6WMKmU
s+vIMjsmcnZMgIkvPzVp032+iDuL96p4WmPWkHlvQ1rS0QEnMe6n33ntsoUjEnW8
6/DsbopOwyvgAbAfJNkMxjO5gGDUo9RuVfK3NEeUh488OcrOAhsZ2Zzfv8T3zmub
xR2sq8DNBwX7wsqSwuupsfYa5byTClqrbqq9JehSrT2OZLr2WVVkUPA/2hTtyrLK
EnlOEzqmb2qFqNgZoI2EoPOsxYo+zse0gSk3QritJ3qn0GRUN9bb5T9DWdSAw192
IfCg9xbd/JhiJynPlyIEDkI0/6FNLVzD2jPbr3rWDvDQV9O43uLB71/bY423XKA=
=7qAt
-----END PGP SIGNATURE-----

Re: [Twin-dev] Twin popularity is on the rise, but I can't find the site...

[Bradley D. T. <Bradley@NorthTech.US>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160



On 02/24/2012 01:34 PM, max wrote:

> I have a copy of all the contents that was hosted there, so moving the
> website somewhere else is easy :)
> 
> I appreciate your offer, and I thank you for your committment, anyway I
> think it's easier - and a more long term solution - to simply move the
> website to Sourceforge.
> 
> If you are willing to help moving in this direction, I will really
> appreciate your contribution. If you want, I can send you the website
> copy I have saved.

Yes I would be more than happy to help out :)

I'll contact you offlist in that regard, yet in the meantime, to share
with the list, here is the URL to the article I mentioned that I
published last year:
http://northtech.us/content/20110311/adding-convenient-layers-persistent-abstraction-simplify-remote-administration

Looks like I'll be writing up a few more articles soon too!

I've also wrote to the Slackware package maintainer of record asking if
he would like me to take over maintenance of it. I wrote to him last
year advising that version 0.6.2 was available at the now defunct
website but there was no update to the Sbo.

I also maintain binary packages of Twin which are available here:

gopher://gopher.northtech.us/9/software/packages/twin-0.6.2-x86_64-1_SBo.tgz

I get between 10 and 20 downloads a month of this package myself, which
is pretty good considering that I only make it available via the Gopher
protocol, and it's a 64bit only package for a single, particular Linux
distribution.

If I'm named the maintainer of the SBo for twin, I'll be maintaining in
both 32 and 64 bit versions, and I have plans for creating RPMs for
CentOS, Scientific Linux, and Mageia, as well as a more current Debian
package as part of my repo.

Well, thanks for such a wonderful product - one that came from thinking
outside of the box, and therefore has more utility than many quite
understand.

I was beginning to get worried whether something had happened to you
after everything was silent for so long.

Kindest regards,


- -- 
Bradley D. Thornton
Manager Network Services
NorthTech Computer
TEL: +1.310.388.9469  (US)
TEL: +44.203.318.2755 (UK)
TEL: +41.43.508.05.10 (CH)
http://NorthTech.US

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Find this cert at x-hkp://pool.sks-keyservers.net

iQEcBAEBAwAGBQJPSCbFAAoJEE1wgkIhr9j3NrwH/1+eAI0txVXt2YVg0YkQWRAc
nd4pT+/J/O8pVSYEi+TOj8pTXlZosPdQzijZi0l/lC4x/vYRf0FVQvSU9V1+DREs
2U4xsGJD5gJtDw6sXws94ngHY6nta/MJYL3xXBZITOPJOZiALyHJWqcO81JzTz/3
08ZuCivm25QnRt76Y0RybTQX9cpLMIp+nP20rqSLsgL1ZmuHERQsgKLejIpRizK7
m/OVXT5gb4ERQ6WmP5Fj16tA5qv10Eq4Yscah5Aq9z7sPh+SpYSIGCUR3S6eX1pK
KE0w0k1qxzAopE8xYUXuM/ZpIHKAb5E1Ym4HvK3X1KCX3uop4dPWhg0vOvMgCNE=
=ugKx
-----END PGP SIGNATURE-----

[Twin-dev] new version twin-0.6.2, fixes some more 64-bit bugs

[max <mas...@gm...>]

Hello everybody,

with *ONLY* 3 years delay,
and thanks to the kind soul that pointed out
http://linuz.sns.it/~max/twin/ is not accessible anymore,

I released twin-0.6.2 on SourceForge.

Enjoy.

P.S. there are still some 64-bit related bugs. I have fixed one related
to scrolling, and I will release a new twin version when I find and fix
more bugs, or when I get convinced that there are no more serious 64-bit
related bugs.

P.P.S. if the few readers of this mailing list did not realize yet,
I basically work on twin only when somebody prompts me...
it's a side effect on focusing my limited free time on other tasks
(including my new toy: http://sf.net/projects/fstransform/ )


On 09/26/2009 03:47 PM, Massimiliano Ghilardi wrote:
> On Saturday 26 September 2009, Robin Atwood wrote:
>> On Saturday 26 September 2009, Massimiliano Ghilardi wrote:
>>> I am glad you appreciate TWIN.
>>
>> Max -
>> Did you ever make any progress with twin-0.6 and 64 bit systems? I would
>> still like to upgrade someday from 0.4.6. Lack of utf-8 support is a bit of
>> a hassle.
>>
>> Regards
>> -Robin
> 
> Hello Robin,
> 
> you remind me that I forgot to release a new version of twin
> after fixing 64-bit support... it seems I am getting old ;-)
> 
> Anyway, I just released twin-0.6.2 now, with fully working 64-bit support
> (I use it regularly and I did not notice any problems).
> For the moment, it is available only from http://linuz.sns.it/~max/twin/
> 
> I should upload it also to sourceforge, I will do that as soon as I have a bit 
> more time.
> 
> Best regards,
> 
> Massimiliano
> 

Re: [Twin-dev] Twin popularity is on the rise, but I can't find the site...

[max <mas...@gm...>]

On 02/22/2012 03:04 AM, Bradley D. Thornton wrote:
> Hi Max,
> 
> I can't find the Twin website:
> 
> http://linuz.sns.it/~max/twin
> 
> http://linuz.sns.it/~max/twin/docs/README
> 
> etc...
> 
> That's the site where all of the docs, tutorial, README, etc was
> residing, as well as the current version (0.6.2).
> 
> The sourceforge site has vitually none of that, and the most current
> version there is 0.6.1
> 
> I'll be happy to host or mirror for you if you like, and I'm very
> interested in beefing things up a little to make it more accessable.

Hello Bradley,

you're right, twin website is no longer accessible.
It was hosted on a university server, and I don't think it's appropriate
to try to restore it on the same server (my university years are long gone).

I have a copy of all the contents that was hosted there, so moving the
website somewhere else is easy :)

I appreciate your offer, and I thank you for your committment, anyway I
think it's easier - and a more long term solution - to simply move the
website to Sourceforge.

If you are willing to help moving in this direction, I will really
appreciate your contribution. If you want, I can send you the website
copy I have saved.


> I wrote an article on my site and many visitors had been reporting 404
> errors.
> 
> I also maintain a repo for the 0.6.2 and 0.6.1 version packages of Twin
> for Slackware Linux which is very popular (about 10 downloads/day).
> 
> When I visited the sourceforge site I noticed that there were also 20
> downloads just this week too.

After you mentioned it, I checked, and I must tell I am surprised by the
quite high - and steady - number of downloads.

Especially if you consider I wrote twin in 1992 as a DOS program, and
the Linux version was a one-day hack in 1999, after many years the DOS
version was sitting unused on my PC...

> I would like to write an actual users manual to for twin, and would
> appreciate some input from you directly, as well as others on the list
> if they are so inclined.

I am glad to help, but please consider I don't have experience with
manpages, nroff and so on, and unluckily also no time to learn them.

What I can do is to write some plain-text explanations, or, if you
prefer, answer your questions. Do you think it can be enough?

> Twin is an invaluable tool - in spite of the lack of unicode support,

This is an interesting sentence, as twin actually *has* unicode support.
By current standards, anyway, configuring twin to use unicode is
definitely non-trivial and poorly documented.
And it also has several limitations: for example, it can only do
left-to-right text without ligatures.

> and I would really hate to see this most excellent window manager vanish
> - a concern that many others have voiced as well.
> 
> Looking forward to hearing back from you at your earliest convenience,
> 
> Kindest regards,

Be assured I have no intentions to let twin vanish. Even if I am not
doing active development anymore, I still do bug-fixing, and I would
surely appreciate contributions, especially for the most problematic
parts (Makefiles to say one).

In the meantime, since it seems I forgot to upload twin 0.6.2 on
Sourceforge, I will start by fixing that.

Thanks for your interest,

Massimiliano Ghilardi

[Twin-dev] Twin popularity is on the rise, but I can't find the site...

[Bradley D. T. <Bradley@NorthTech.US>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

Hi Max,

I can't find the Twin website:

http://linuz.sns.it/~max/twin

http://linuz.sns.it/~max/twin/docs/README

etc...

That's the site where all of the docs, tutorial, README, etc was
residing, as well as the current version (0.6.2).

The sourceforge site has vitually none of that, and the most current
version there is 0.6.1

I'll be happy to host or mirror for you if you like, and I'm very
interested in beefing things up a little to make it more accessable.

I wrote an article on my site and many visitors had been reporting 404
errors.

I also maintain a repo for the 0.6.2 and 0.6.1 version packages of Twin
for Slackware Linux which is very popular (about 10 downloads/day).

When I visited the sourceforge site I noticed that there were also 20
downloads just this week too.

I would like to write an actual users manual to for twin, and would
appreciate some input from you directly, as well as others on the list
if they are so inclined.

Twin is an invaluable tool - in spite of the lack of unicode support,
and I would really hate to see this most excellent window manager vanish
- - a concern that many others have voiced as well.

Looking forward to hearing back from you at your earliest convenience,

Kindest regards,


- -- 
Bradley D. Thornton
Manager Network Services
NorthTech Computer
TEL: +1.310.388.9469  (US)
TEL: +44.203.318.2755 (UK)
TEL: +41.43.508.05.10 (CH)
http://NorthTech.US

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Find this cert at x-hkp://pool.sks-keyservers.net

iQEcBAEBAwAGBQJPRE0+AAoJEE1wgkIhr9j3/CEH/2C4Mm1WbRkwZWIi4+s0xq79
EzUwWNWYuJr1wOgijmsgqXuZe7tLCbF8guv+2liatAFJS+ZvNs/7Ahi484ztnF/G
xEkaLM6mlahtPRnDpMftS2l0JqkQjo3Q6oGT3VPDwxYBLwNjAYVdCrRonDq56oW1
Kico/nfaofd39qK+jIzDbBqKUBPb9NIt/G2sM4oK572tDgg8Ek2UOpvGab/nbzqP
lPxKmj2DOMjOIySD7rYPcPVitsKQBqwMfP9eyKtR+nvgEmOzAdpNagjRl32Afehc
SsOHZk2T02tOL8taCEMpLvoXhX6p6lpVI6tiu+SilzZlc4tU9JZ8Pp9cTn5H2rA=
=K5p2
-----END PGP SIGNATURE-----

Re: [Twin-dev] Building on Ubuntu 10.10

[max <mas...@gm...>]

On 06/16/2011 05:57 PM, Scott Duensing wrote:
> I am trying to build Twin on Ubuntu 10.10 (yea, I know I need to get
> with the times).  It's putting a lot of extra "-e"s in the .flag and
> .module files.  Any ideas?
> 
> Thanks!
> 
> 
> Scott
> 

Hello Scott,

it's a quite common problem. When it happens, edit the file 'makerules'
and replace the line
  ECHO_E:=echo -e
with
  ECHO_E:=/bin/echo -e
then remove all corrupted files (.modules and .*.flags) with
  make clean
  find -name .modules | xargs rm
finally launch again 'make <your options>'

Please let me know if this solves your problem, or if you get further
errors.

Regards,

Massimiliano

[Twin-dev] Building on Ubuntu 10.10

[Scott D. <sc...@ja...>]

I am trying to build Twin on Ubuntu 10.10 (yea, I know I need to get
with the times).  It's putting a lot of extra "-e"s in the .flag and
.module files.  Any ideas?

Thanks!


Scott

Re: [Twin-dev] Points of clarification: twattach / twdisplay

[max <mas...@gm...>]

On 06/07/2011 09:47 AM, Eric Gillespie wrote:
>
> [...]
>
> I wasn't actually running this entirely on the console. I also have
> exactly the same issues if I run using a purely gfx/X11 server. Unicode
> doesn't seem to work.
> 
>     Risking to tell the obvious, for Unicode support you need, among other
>     things, an Unicode font. I know several Unicode fonts for X11 - some
>     are free and should be easy
>     to find, you can start 'xfontsel' and choose 'iso10646' from the
>     'rgstry' pull-down menu, maybe you have some already installed as well.
> 
> 
> Yup. Got several of those, even have a couple of monospace ones that
> have iso10646-1 encoding.
>  
> [...] I'm trying to get to a place where I can get irssi to render
> characters from multiple regions - whether they be European, Asian or
> whatever. Irssi can do the job, it just requires a fully UTF-8
> environment to do so. I've had irssi do UTF-8 fine inside
> gnome-terminal, though I think gnome-terminal's not particularly great.

Hello Eric,

Sorry for the delay,

I understand what you're saying, and you're right that things are easier
on X11 than on Linux console.

On the other hand, in order to be able to help you in your specific
case, I need a sufficiently detailed problem report ... I am afraid that
anything resembling "it doesn't work" will not let me help you much.

So, for the moment, the best I can do is to explain what you can do to
get (hopefully) the detailed problem report I need:



1) start a program you want to use in UTF-8 mode (irssi) inside a
terminal where it works in UTF-8 (gnome-terminal) and take a snapshot
while it shows several non-ASCII characters - please do NOT take a
snapshot of personal information :)

2) open a twterm inside twin, which must be inside X11 and using a
Unicode font - if possible, you should use the same font used by the
other terminal (gnome-terminal).

3) enable utf-8 mode in twterm: inside twterm, run the usual command
   echo -e '\033%G'

4) start the same program as before (irssi) with the same options, but
now inside the twterm you just prepared in 2) and 3)

5) check that the program (irssi) is actually using UTF-8 (how? depends
on the program... I cannot help here)

6) take a snapshot while it (irssi) shows several non-ASCII characters -
possibly they should be the same data as the previous snapshot - even
they will be probably garbled.

7) send me both snapshots

> [...]
> 
> I recently struck incidences where twin wasn't receiving keys properly
> from the keyboard, even though those keys worked fine in console mode
> ordinarily. I.e. the space key. I'd hit the space key, and nothing would
> react... Any clues of where to look for this?
>
> Thanks,
> Eric.


Hmm... there were some bugs in raw keyboard driver:
one was fixed in twin 0.6.0, another in twin 0.6.2.

Can you try to upgrade to 0.6.2 and see if the problem is still there?
If you still experience it after upgrading, I will need you to run
strace(8) on twin while it shows this problem, and send me strace(8) output.

Waiting for your feedback,

Massimiliano

Re: [Twin-dev] Points of clarification: twattach / twdisplay

[max <mas...@gm...>]

On 04/06/2011 10:35 AM, Eric Gillespie wrote:
> [...]
>
> whew. Glad to hear it. I've been "evangelising" Twin for quite a
> while, (several years now) and I even have a .twinrc that works very
> very well for me, but I had to beef up the amount of memory used to read
> the .twinrc.  I've got twstart/twdisplay down pat. However, I have
> another problem with it.
>
> Unicode. Or more properly, lack of proper UTF-8 support.  I'll give an
> example.
>
> I'm on Linux (what else would I be on?), running a twin with a tty
> display, and a gfx display. I find this is often the best way to keep my
> irssi IRC client up and running even if Xorg drops its biscuits and
> kills itself.  Irssi supports utf-8 fine - I've tested it in
> gnome-terminal. But irssi doesn't manage to do the utf-8 thing inside a
> twterm at all. I get lots of weird looking characters that are basically
> the twterm puking.
>
> What do I look at, and how do I set things up so that UTF-8 works for me
> seamlessly?
>
> Great set of programs, by the way.
>
> Cheers, Max
> Eric.
>

Hello Eric,

Unicode and UTF-8 support is a bit tricky in Linux console, and adding 
more software layers (twin in this case) does not help it.

Risking to tell the obvious, for Unicode support you need, among other 
things, an Unicode font.

I know several Unicode fonts for X11 - some are free and should be easy 
to find, you can start 'xfontsel' and choose 'iso10646' from the 
'rgstry' pull-down menu, maybe you have some already installed as well.

As you probably know, text-mode Linux console does not support full 
Unicode: it is a VGA limitation that console fonts can have at most 256 
glyphs (512 with some tricks). Unluckily, I do not know enough about 
Linux kernel framebuffer to tell if it's different in this respect.

So, with only 256 (or 512) glyphs available in hardware, the most 
critical point is that you must choose a font containing the subset of 
Unicode _YOU_ need. For Italy, I would choose an iso8859-1 (latin1) or 
iso8859-15 (latin15) font. I cannot help for your country and language, 
as you did not tell them :)

The next steps are to find, install and finally load the font into VGA 
hardware - for the latter you can use setfont(8) or a similar command.

Then you should enable UTF-8 mode in Linux console (WITHOUT starting 
twin): to enable UTF-8 mode you can use
   echo -e '\033%G'
and to disable it:
   echo -e '\033%@'
this step *may* be unnecessary, as some programs could perform it 
automatically.

Then run your favourite program and check if it can correctly show 
non-ascii characters in Linux console.

If it cannot, maybe it did not autodetect that it is running
inside an UTF-8 capable terminal, and you need to find
some option or configuration inside the program itself
to manually enable UTF-8 output...


Once everything is well, you just need insert twin in the picture:
1) start twin and tell it the encoding of the loaded VGA font:
    you will need to run something like
      twin --hw=tty,charset=iso8859-15
    (obviously the exact charset name depends on which font you actually
     have chosen and loaded)
2) enable utf-8 mode in twterm: inside twterm, run the usual command
      echo -e '\033%G'
3) start your favourite program, and check again if it can show
    non-ascii characters.
4) again, you may need to (somehow) enable UTF-8 output
    in your favourite program


Good luck, and please tell me if it works - or which problems you found

Max

Re: [Twin-dev] Points of clarification: twattach / twdisplay

[max <mas...@gm...>]

Hello Bradley,

if you don't mind, I will try to be more concise than you in my answer.

On 04/01/2011 03:17 PM, Bradley D. Thornton wrote:
> [...]
> In all of the official documentation I've come across so far, I've
> noticed no mention of *twstart*.
>
> Now, this is kind of odd, I started Twin by invoking it as '$ twin', or
> '$ twin&' for a while until I found *twstart*.
 >
 > [... skipped strange behaviour of twstart ...]
 >
 > Can I get a little background on twstart?

twstart is a tiny script that does the following:
1) looks for already-running instances of twin
2) if found, tries to start "twdisplay" connected to it
3) otherwise, starts a new twin.

If you got strange behaviour, maybe you can investigate it more in 
detail to understand if it's a bug or not. But since you seem to have 
doubts related to twdisplay/twattach, you should _for_sure_ clarify them 
before proceeding, as they are a fundamental part of twin.

> Okay that just leads into the next questions. I had a really hard time
> following the tutorial wrt attach/detach functionality, and I'd like to
> maybe have firm grasp, since I'm in the planning stages of evangelizing
> Twin.

Evangelization? *laugh*
I am definitely amused, but I don't know if I should also be scared ;-)

> back when I was runing twin, I tried to follow along, but *twattach* and
> twdisplay kept throwing errors like, "display not set" or "twdisplay not
> set" or something like that.

To run twdisplay or twattach, you must tell them which instance of twin 
to connect to:
twattach --twin@<TWDISPLAY> --hw=<display>
twdisplay --twin@<TWDISPLAY> [--hw=<display>]

you can omit the option --twin@<TWDISPLAY> only if TWDISPLAY is already 
set, which is not the usual situation when you run them.

The fundamental difference between twattach and twdisplay is:

if you compare twin to a VNC server, twdisplay acts like a VNC client:
it receives information about what to draw, and actually draws it

instead twattach simply tells twin to start a new display: the display
will be directly managed by twin, not by twattach.

> First, I logged in at run-level 3 on tty3 and did:
>
> $ twstart
>
> I got twin in full screen and opened a couple of twterms, ssh'd out to
> some hosts.
>
> Then, I detached from the menu. Then ran twstart again. no problem.

good.

> This session is ":0" (meaning display #0 ???), as indicated in the upper
> right of twin's title bar.

yes, it means you started instance 0, and its TWDISPLAY is :0
(same notation as DISPLAY used by X11)

> Then I logged into tty2, ran "startx", opened an xterm, and ran/got:
>
> $ twstart
> twin: autoprobing `--hw=gfx' display driver.
> twin: starting display driver module `HW/hw_gfx'...
> twin: ...module `HW/hw_gfx' successfully started.
>
> Twin starts up, very nicely. On the title bar I have:
>
> "Twin :1 (display #1, right?)

this is strange, twstart should have detected there was already instance 
:0 running.

> With a ^+Alt+Fn I can switch from X Windows to a Linux console.
>
> I hit ^+Alt+F3 and I am back in ":0"
>
> *** I have no keyboard control inside my *twterms*. I have gpm installed
> on the base system, so I have a mouse and the curser moves.
>
> I click outside the focus of the twterms and the menu changes. I have
> control of Twin, and can even open another twterm, but even in this new
> twterm I have no keyboard control over my CLI.
>
> I have keyboard control over Twin though. I can hit the pause button to
> bring up my menus.
>
> So from the menu I detach, and reattach with:
>
> $twstart
>
> All is well again. all three twterms are acceping keyboard input.

this is strange as well, maybe you hit a bug.
Which version of twin are you using?

if the behaviour is reproducible, it's worth investigating it more.:
can you strace(1) twin before doing the X-to-console switch, then press 
some keys (and the twterms don't react) ?

> - ------------------------------------------------------------
> Switch back to X and back to Linux Console a couple of times
> - ------------------------------------------------------------
>
> This is a recurring problem. Fortunately attaching and reattaching with
> twstart takes care of it, but it's a real hassle to do everytime.
>
> There's no problem w/Twin or the twinterms under X - it works whenever I
> come back by *Alt+F7* just fine.
>
> - ------------
> Gets weirder
> - ------------
>
> At some point, things get weird, and "$ twstart" would only open display :0

that's actually the *expected* behaviour

> So I began to specify "$ twstart :0" and "$ twstart :1" - I opened a
> third twin session in Xfce by doing "$ twstart :2"

> - ----------------------
> Observations - Initial
> - ----------------------
>
> 1.) Sometimes Twin starts with Left button control and sometimes it
> starts with Right button mouse control for the menus. Odd.
>
> Maybe it had something to do with a *.twinrc* I found in the docs and
> cp'd to my ~ ?????

Yes, it's probably due to the .twinrc you copied in your home folder.

> I moved that file to ".twinrcc" to use the system
> default instead for the time being, we can get back to that at anytime,
> and I need to know how twin is going to react with and without one If
> I'm going to generate operational docs for it.
>
> 	Q: Any suggestions on hunting this issue down?

rename/move the .twinrc you copied in your home folder and check if the 
issue goes away. If it does, it means that your ~/.twinrc sometimes gets 
loaded and sometimes not. Check the messages in Menu -> = -> Messages 
when you get the first behaviour, check again when you get the second 
behaviour, and search for errors and for differences.

You can also send me the two sets of messages.

> 2.) A few days ago, I blew out my twin and couldn't reattach to it at
> all, whenever I would type "$ twstart" it would say it started and then
> just sit there - but no Twin. Same with "$ twstart&" - Nothing. the job
> would start and bg, but no display would popup.
>
> This happened, I think, after trying to figure out *twattach* and
> *twidisplay, unsuccessfully.

One way to screw is using twattach to tell twin to attach to itself.

> 3.) Twin consistently breaks wrt twterm CLI access within Twin, on Linux
> console whenever I "^+Alt+Fn". i.e., If I: ^+Alt+F3", then I have no
> keyboard control in any twterm - unless.....
> [...]
> 	Q: Can you explain what might be going on here?

I answered above about how you can extract more details of what's happening.

> +============================================================+
> | twattach / twdisplay / twin / twstart - the differences??? |
> +============================================================+
>
> I get the following from *twattach --help* :
>
> <snip>
>
> ~$ twattach --help
> Usage: twattach [OPTIONS] --hw=<display>  [...]
> Currently known options:
>   -d                       detach display
>   -s, --share              allow multiple simultaneous displays (default)
>   -x, --excl               request exclusive display - detach all others
>   -f, --force              force running even with wrong protocol version
> [...]
>
> - -----------------------------
> - -s is supposed to be the default, but when if, for example, I do a:
>
> $ twstart :3
>
> in a linux console when I already have display ":3" running in X, it
> kills the instance of twin running under the X Session.

Then you found a bug. Again, which version are you using?
I start thinking it's related to the 64-bit bug fixed in version 0.6.2.

> - -----------------------------
> - -d
>
> Is this so I can detach a parent instance of a twin session from within
> one of its twterm's? And if so, is the following syntax the correct syntax?:
>
> $ twattach -d
>
> or do I need more parameters?

you need
twattach [--twin@<TWDISPLAY>] -d --hw=<display>
and it is used to detach twin from a display. nothing to do with twterm

>[...]
> If, on a Linux console, I do:
>
> $ twstart :3
>
> When I already have a :3 running in another Linux console, I get a brand
> new shiny :4 instead.
>
> If I do a:
>
> $ twattach -s :3

the correct syntax for what you're twying to do is:
twattach --twin@:3 --hw=tty

> - -----------------------------
> - -f
>
> I have no idea what that means, or in what kind of situation I would use it.

to bypass version checks. twattach and twdisplay check that they are the 
same version as the twin they connect to and, without -f, they refuse to 
run if the versions don't match

> I'm lost here when it somes to $TWDISPLAY. When (on a tty Linux console)
> I check my global environment variables with "$ set" I see no TWDISPLAY
> variable.

TWDISPLAY is set by twin, and is visible only in shells and commands 
started inside twin.
It's exactly analogous to DISPLAY: it's not set in the console, but only 
in shells and commands you start inside X11.

 > [...]

> Let's say the name of localhost is 'hammer' (it actually is lol). I want
> to have a simultaneous display, and I've already got a Twin session
> displaying in tty3, and I'm on tty4.
>
> Am I close to being correct on some of the following?:
>
> $ twattach --hw=tty3:0 (if Twin display :0 is running on tty3?)
> $ twattach --hw=hammer@tty3:0 -s (-s is assumed to be the default)
> $ twstart (with the two variations above.)
>
> $ twattach --hw=twin@:0 (I'm not sure what this means really.

if you have twin running in tty3, you are currently in tty4
and want to open (actually, "attach") the existing twin also to tty4,
the command is:

twstart

or, if you like doing things manually, you need to know the instance 
number of the twin running on tty3 (its TWDISPLAY, it's written on the 
menu bar at the extreme right). Let's suppose it is :0 - then you can:

twattach --twin@:0 --hw=tty

> And I have no idea what --hw=ggi is for.

To use GGI library as display, but support is pretty incomplete.
See http://www.ggi-project.org/

> #--------Under X Windows (Xfce)----------------
>
> $ twattach --hw=gfx:0 (if I want two simultaneous sessions in different
> 			pager windows under Xfce)

again, the syntax is
   twattach --twin@<TWDISPLAY> --hw=gfx
but you could simply run
   twstart --hw=gfx
or, since --hw=gfx is the first display method tried, a bare
   twstart
will suffice


> $ twattach --hw=X:0
> or
> $ twattach --hw=X:0 -x (to kill the other instance and display in my new
> 			pager windows under Xfce?)

you should not put ':' after --hw=X. I think you mean
   twattach --twin@:0 --hw=X
   twattach --twin@:0 --hw=X -x


> #--------Remote Management of Twin displays-----
 > [...]
 > - ------------------------------------
 >
 > Finally (I promise LOL), Should I be able to ssh into a box, start
 > either *screen* or *tmux* and be able to run twin inside of those?
 > [...]

I will leave this for later, after we are both confident that
1) twin, twattach and twdisplay are working correctly on your box
2) you have learned how to use twattach and twdisplay

Only a last point: yes, you can run twin inside "screen" or "tmux",
but you'd probably get much better results doing the opposite,
i.e. running "screen" or "tmux" inside twin.


> - ------------------------------------
>
> The Tutorial seems to indicate that any options and arguments would be
> invoked pretty much the same whethor *twdisplay* or *twattach* is used.
>
> Am I correct in this assumption? For the most part?
>
>
> Locally, I'm typically running many SSH sessions in twterm's inside of
> twin, and running tmux or screen on those remote hosts.
>
> When, under Xfce on my localhost, I invoke the following from within a
> twterm
>
> $ twstart :7 # (as an example)
>
> I get:
>
> tallship@hammer:~$ twstart :9
>
> twin: autoprobing `--hw=gfx' display driver.
>
> twin: starting display driver module `HW/hw_gfx'...
>
> gfx_InitHW() failed: DISPLAY is not set
> twin: ...module `HW/hw_gfx' failed to start.

this is not normal. you started twin (and twterm) inside Xfce,
so DISPLAY _should_ be set.

>
> twin: autoprobing `--hw=X' display driver.
>
> twin: starting display driver module `HW/hw_X'...
>
> X11_InitHW() failed: DISPLAY is not set
> twin: ...module `HW/hw_X' failed to start.
>
> twin: autoprobing `--hw=twin' display driver.
>
> twin: starting display driver module `HW/hw_twin'...
>
> twin: ...module `HW/hw_twin' successfully started.
>
> I don't quite get it. First it tries *gfx*, then *X*, and then *twin*,
> which opens up (as you proudly state) another session of twin, inside a
> session of twin. neat :) But what is going on above and what would be a
> good command line to avoid the autoprobing and just do it?

twstart --hw=gfx
twstart --hw=X

will not try any display autodetection and just open a gfx or X display 
(provided that DISPLAY is set).

If for some (strange) reason, DISPLAY is not set, you can use

twstart --hw=gfx@:0
twstart --hw=X@:0

which are equivalent to

DISPLAY=:0 twstart --hw=gfx
DISPLAY=:0 twstart --hw=X

> +============+
> | Conclusion |
> +============+
>
> I suppose this is enough for me to get a really good grasp to start off
> with, and then experiment, and then begin writing up a step by step
> manual with explicit examples.
>
> I already have got the interest of several people who like Twin (Do you
> prefer *Twin* *TWin* *TWIN* or *twin* when it is being referred to?),

Twin is ok :)
And in the middle of a sentence you can also write twin lowercase.

> and have asked me for some docs to guide them along.
>
> Your tutorial, although comprehensive, I found confusing, and so did
> some others who are using twin only very basically.

I understand. For sure, one positive thing that will result from your 
efforts is a lot of ideas about how to improve the tutorial :)

> Once again, I apologize for this extremely long post, but I want to
> devote an appropriate amount of attention to Twin in order to utilize it
> to the fullest :)
>
> I sincerely hope that I am not presenting too much of a burden by asking.

Well, after a certain point I cheated and left some parts for later,
after we will have clarified how to use twdisplay and twattach.

>
> Kindest regards,
 >
> - --
> Bradley D. Thornton
> Manager Network Services
> NorthTech Computer
> TEL: +1.760.666.2703  (US)
> TEL: +44.702.405.1909 (UK)
> http://NorthTech.US
>

[Twin-dev] Points of clarification: twattach / twdisplay

[Bradley D. T. <Bradley@NorthTech.US>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

+==========+
| Preamble |
+==========+

Alright. You know you're in for some voluminous trouble when someone
says, "Preamble" before anything else, at the beginning of a document.

Let me start off first by apologizing in advance for such a long and
perhaps even seemingly endless post.

I am in the beginning of putting together a walk-through manual, with
explicit demonstrations of how one would actually invoke specific
features and use Twin in particular situations - complete with examples
and I'm also planning for some videos too :)

#-----I recommend getting a cup of coffee before starting to read-----


Okay, for me at least, the docs, although rather exhaustive and
comprehensive, really aren't clicking with me in some places.

I'm sure I'll have a lot of questions moving forward, with more
intricate particulars, but for now I just need a solid grounding in some
of the basics.

Although literature regarding Twin out there is rather sparse, I found
some seemingly undocumented  stuff at one site covering the installation
of Twin.

In all of the official documentation I've come across so far, I've
noticed no mention of *twstart*.

Now, this is kind of odd, I started Twin by invoking it as '$ twin', or
'$ twin &' for a while until I found *twstart*.

When I ran twstart for the first time, Twin came up full screen (in one
of my pager windows under Xfce) and my left mouse button was suddenly
the one that activated the menus, with some other inconsistencies in the
manipulation of Twin from when I invoked it as *twin*.

Can I get a little background on twstart?

Okay that just leads into the next questions. I had a really hard time
following the tutorial wrt attach/detach functionality, and I'd like to
maybe have firm grasp, since I'm in the planning stages of evangelizing
Twin.

back when I was runing twin, I tried to follow along, but *twattach* and
twdisplay kept throwing errors like, "display not set" or "twdisplay not
set" or something like that.

+===============+
|  Particulars  |
+===============+

I'll try to follow the chronology as best I recall...

- ------------------
On a Linux Console
- ------------------

First, I logged in at run-level 3 on tty3 and did:

$ twstart

I got twin in full screen and opened a couple of twterms, ssh'd out to
some hosts.

Then, I detached from the menu. Then ran twstart again. no problem.

This session is ":0" (meaning display #0 ???), as indicated in the upper
right of twin's title bar.

- ----------------
Running X w/Xfce
- ----------------

Then I logged into tty2, ran "startx", opened an xterm, and ran/got:

$ twstart
twin: autoprobing `--hw=gfx' display driver.
twin: starting display driver module `HW/hw_gfx'...
twin: ...module `HW/hw_gfx' successfully started.

Twin starts up, very nicely. On the title bar I have:

"Twin :1 (display #1, right?)

With a ^+Alt+Fn I can switch from X Windows to a Linux console.

I hit ^+Alt+F3 and I am back in ":0"

*** I have no keyboard control inside my *twterms*. I have gpm installed
on the base system, so I have a mouse and the curser moves.

I click outside the focus of the twterms and the menu changes. I have
control of Twin, and can even open another twterm, but even in this new
twterm I have no keyboard control over my CLI.

I have keyboard control over Twin though. I can hit the pause button to
bring up my menus.

So from the menu I detach, and reattach with:

$twstart

All is well again. all three twterms are acceping keyboard input.

- ------------------------------------------------------------
Switch back to X and back to Linux Console a couple of times
- ------------------------------------------------------------

This is a recurring problem. Fortunately attaching and reattaching with
twstart takes care of it, but it's a real hassle to do everytime.

There's no problem w/Twin or the twinterms under X - it works whenever I
come back by *Alt+F7* just fine.

- ------------
Gets weirder
- ------------

At some point, things get weird, and "$ twstart" would only open display :0

So I began to specify "$ twstart :0" and "$ twstart :1" - I opened a
third twin session in Xfce by doing "$ twstart :2"

- ----------------------
Observations - Initial
- ----------------------

1.) Sometimes Twin starts with Left button control and sometimes it
starts with Right button mouse control for the menus. Odd.

Maybe it had something to do with a *.twinrc* I found in the docs and
cp'd to my ~ ????? I moved that file to ".twinrcc" to use the system
default instead for the time being, we can get back to that at anytime,
and I need to know how twin is going to react with and without one If
I'm going to generate operational docs for it.

	Q: Any suggestions on hunting this issue down?

2.) A few days ago, I blew out my twin and couldn't reattach to it at
all, whenever I would type "$ twstart" it would say it started and then
just sit there - but no Twin. Same with "$ twstart &" - Nothing. the job
would start and bg, but no display would popup.

This happened, I think, after trying to figure out *twattach* and
*twidisplay, unsuccessfully.

	Q: Nothing yet, I have questions below, later...


3.) Twin consistently breaks wrt twterm CLI access within Twin, on Linux
console whenever I "^+Alt+Fn". i.e., If I: ^+Alt+F3", then I have no
keyboard control in any twterm - unless.....

	a.) dettach and re-attach
or
	b.) If I "Alt+Fn" to any tty console and then "Alt+F3" back, I
		regain control of my keboard within twterms again.

	Q: Can you explain what might be going on here?

+============================================================+
| twattach / twdisplay / twin / twstart - the differences??? |
+============================================================+

I get the following from *twattach --help* :

<snip>

~$ twattach --help
Usage: twattach [OPTIONS] --hw=<display> [...]
Currently known options:
 -d                       detach display
 -s, --share              allow multiple simultaneous displays (default)
 -x, --excl               request exclusive display - detach all others
 -f, --force              force running even with wrong protocol version
 --twin@<TWDISPLAY>    specify server to contact (default is $TWDISPLAY)
 --hw=<display>[,options] start the given display driver
Currently known display drivers:
	X[@<XDISPLAY>]
	twin[@<TWDISPLAY>]
	tty[@<tty device>]
	ggi[@<ggi display>]


</snip>

- -----------------------------
- -s is supposed to be the default, but when if, for example, I do a:

$ twstart :3

in a linux console when I already have display ":3" running in X, it
kills the instance of twin running under the X Session.
- -----------------------------
- -d

Is this so I can detach a parent instance of a twin session from within
one of its twterm's? And if so, is the following syntax the correct syntax?:

$ twattach -d

or do I need more parameters?
- -----------------------------
- -x

This is the behaviour I *think* I'm getting when I'm using twstart
instead of twattach, thinking perhaps erroneously, that the -s switch is
implied, even though "twstart --help" doesn't show a -s option.

If, on a Linux console, I do:

$ twstart :3

When I already have a :3 running in another Linux console, I get a brand
new shiny :4 instead.

If I do a:

$ twattach -s :3

I get:

twattach: libTw error: TWDISPLAY is not set, while a:

$ twattach :3

yields an argument error due to usage

	Q: What would be a correct syntax? I've read the tutorial, but
		it leaves me a little confused, along with the different
		man pages and --help options.
	A couple of good hypothetical examples would be best for me to
	digest instead of the notation in the docs.
- -----------------------------
- -f

I have no idea what that means, or in what kind of situation I would use it.
- ----------------------------
 --twin@<TWDISPLAY>    specify server to contact (default is $TWDISPLAY)
 --hw=<display>[,options] start the given display driver
Currently known display drivers:
	X[@<XDISPLAY>]
	twin[@<TWDISPLAY>]
	tty[@<tty device>]
	ggi[@<ggi display>]

I'm lost here when it somes to $TWDISPLAY. When (on a tty Linux console)
I check my global environment variables with "$ set" I see no TWDISPLAY
variable.

I've tried (some of this is redundant and addressed above) some of the
following things:

$ twstart :n (where n is the number of a twin display?)

$ twattach :n

$ twattach --hw=:n

$ twstart --hw=:n

Let's say the name of localhost is 'hammer' (it actually is lol). I want
to have a simultaneous display, and I've already got a Twin session
displaying in tty3, and I'm on tty4.

Am I close to being correct on some of the following?:

$ twattach --hw=tty3:0 (if Twin display :0 is running on tty3?)
$ twattach --hw=hammer@tty3:0 -s (-s is assumed to be the default)
$ twstart (with the two variations above.)

$ twattach --hw=twin@:0 (I'm not sure what this means really.

And I have no idea what --hw=ggi is for.


#--------Under X Windows (Xfce)----------------

$ twattach --hw=gfx:0 (if I want two simultaneous sessions in different
			pager windows under Xfce)
$ twattach --hw=X:0
or
$ twattach --hw=X:0 -x (to kill the other instance and display in my new
			pager windows under Xfce?)

If I could get a couple of example of how I would do this on the
localhost, by 'localhost', 'hostname', FQDN, or IP Address?

#--------Remote Management of Twin displays-----

Let's say I have installed twin on a remote host.

Scenario I:	There's is no X-Server running on the remote host

I'm on hammer. I ssh into thor. From my Bash shell on thor I run
twstart. I detach from my running instance of twin and log out.

- From hammer, I want to display the running instance of twin on thor on
my tty3 that I've logged into.

1.) What would be the correct command line?

2.) What would be the correct command line if I wanted to display the
running instance of twin on thor If I was running Xfce locally on hammer
(X runs on Slackware, by default on tty7).

Scenario II:	There is an X-Server running on the remote host

1.) While in my X-Windows session on my localhost, hammer, (my window
manager is Xfce) what would be the correct command line to, I guess the
correct way to say it would be, "Display the instance of twin running on
thor as an *X-Client*, displaying on my localhost (The X-Server), hammer?

For example, if I'm my window manager is Xfce on hammer, and I wanted to
run rxvt as an X-Client on thor, and display it on my X-Server on my
localhost hammer, I would typically:

$ ssh -Y thor rxvt

or

$ ssh -Y thor.sld.tld rxvt

or

$ ssh -Y use...@th...d rxvt

- ------------------------------------

The Tutorial seems to indicate that any options and arguments would be
invoked pretty much the same whethor *twdisplay* or *twattach* is used.

Am I correct in this assumption? For the most part?

- ------------------------------------

Finally (I promise LOL), Should I be able to ssh into a box, start
either *screen* or *tmux* and be able to run twin inside of those?

Locally, I'm typically running many SSH sessions in twterm's inside of
twin, and running tmux or screen on those remote hosts.

When, under Xfce on my localhost, I invoke the following from within a
twterm

$ twstart :7 # (as an example)

I get:

tallship@hammer:~$ twstart :9

twin: autoprobing `--hw=gfx' display driver.

twin: starting display driver module `HW/hw_gfx'...

gfx_InitHW() failed: DISPLAY is not set
twin: ...module `HW/hw_gfx' failed to start.

twin: autoprobing `--hw=X' display driver.

twin: starting display driver module `HW/hw_X'...

X11_InitHW() failed: DISPLAY is not set
twin: ...module `HW/hw_X' failed to start.

twin: autoprobing `--hw=twin' display driver.

twin: starting display driver module `HW/hw_twin'...

twin: ...module `HW/hw_twin' successfully started.

I don't quite get it. First it tries *gfx*, then *X*, and then *twin*,
which opens up (as you proudly state) another session of twin, inside a
session of twin. neat :) But what is going on above and what would be a
good command line to avoid the autoprobing and just do it?

+============+
| Conclusion |
+============+

I suppose this is enough for me to get a really good grasp to start off
with, and then experiment, and then begin writing up a step by step
manual with explicit examples.

I already have got the interest of several people who like Twin (Do you
prefer *Twin* *TWin* *TWIN* or *twin* when it is being referred to?),
and have asked me for some docs to guide them along.

Your tutorial, although comprehensive, I found confusing, and so did
some others who are using twin only very basically.

Once again, I apologize for this extremely long post, but I want to
devote an appropriate amount of attention to Twin in order to utilize it
to the fullest :)

I sincerely hope that I am not presenting too much of a burden by asking.

Kindest regards,


- -- 
Bradley D. Thornton
Manager Network Services
NorthTech Computer
TEL: +1.760.666.2703  (US)
TEL: +44.702.405.1909 (UK)
http://NorthTech.US

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Find this cert at x-hkp://pool.sks-keyservers.net

iQEcBAEBAwAGBQJNldBOAAoJEE1wgkIhr9j3DdEIAKlU9/kOIuJD/kwzAoXB61Vl
2vGr3xZa5JRX7hkbJaRqu+ZZs7KDmUnMaCW8QX0Ff8LfCq1B1faEZh6CCaJUGGIv
DbGxFCrEJVoRt/T+TIq4X1Tk7/XktdAL6OBqcFa951wd/K6WPnUdB1v8fPnvrvbX
GPRsDoIZ7QeFAem8+WeDiAUELysnKmZcjlKLAFjCeKMtbBtU38Q5stddoyxVRpDK
gC3pguPU0TwoPCg4K3EvtxndabHemJYKengMFJ4dWwGoMl6p/pzf9rucdZI/M0FO
OvhqlMYb/SHUbQDxLiRr/h5Ph3HiGMZaE6JmrKhuGtfV1hL5vvaxmu0k/XehOcQ=
=frBI
-----END PGP SIGNATURE-----

Re: [Twin-dev] Any reason not to install the latest stable?

[max <mas...@gm...>]

On 02/13/2011 02:04 AM, Bradley D. Thornton wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: RIPEMD160
>
> It sounds like a dumb question. And maybe it is, but at this particular
> time, 0.6.2 is listed as the current release at
> http://linuz.sns.it/~max/twin/
>
> Although that looks great, here at sourceforge it shows 0.6.1 as the
> current release.

Hello Bradley,

I simply forgot to update Sourceforge.

> I had no problem installing twin-0.6.1 with a package management tool
> native to my particular Linux distro. A simple: "# sbopkg -s twin" and
> it's done.

Good.

> I've created a twin-0.6.2 package but haven't installed it yet. I'd like
> to get some clarification as to the apparent inconsistency above before
> proceeding, if possible :)

As you may have realized from the release dates and the 
almost-unexistent mailing list activity, I am still maintaining twin but 
I am not currently introducing new features.

In particular, 0.6.2 only contains bug fixes with respect to 0.6.1.
For completeness, here is the relevant change log:

 From 0.6.1 to 0.6.2: (Massimiliano Ghilardi, 24 Mar 2009)
	* Fixed buggy RAW keyboard support on Linux console
           (numeric keys did not work with some keymaps)
         * Fixed numeric overflow in twsysmon CPU and disk monitors.
	* Fixed broken libTw on 64 bit archs.

The last fix is actually quite important, as on 64 bit architectures 
most of the clients (including twterm) were severely broken in 0.6.1.

>
> I like Twin. I'm just getting started with it.
>
> Kindest regards,

Thanks :)

I hope this helps,

Massimiliano Ghilardi

>
>
> - --
> Bradley D. Thornton
> Manager Network Services
> NorthTech Computer
> TEL: +1.760.666.2703  (US)
> TEL: +44.702.405.1909 (UK)
> http://NorthTech.US

[Twin-dev] Any reason not to install the latest stable?

[Bradley D. T. <Bradley@NorthTech.US>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

It sounds like a dumb question. And maybe it is, but at this particular
time, 0.6.2 is listed as the current release at
http://linuz.sns.it/~max/twin/

Although that looks great, here at sourceforge it shows 0.6.1 as the
current release.

I had no problem installing twin-0.6.1 with a package management tool
native to my particular Linux distro. A simple: "# sbopkg -s twin" and
it's done.

I've created a twin-0.6.2 package but haven't installed it yet. I'd like
to get some clarification as to the apparent inconsistency above before
proceeding, if possible :)

I like Twin. I'm just getting started with it.

Kindest regards,


- -- 
Bradley D. Thornton
Manager Network Services
NorthTech Computer
TEL: +1.760.666.2703  (US)
TEL: +44.702.405.1909 (UK)
http://NorthTech.US

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Find this cert at x-hkp://pool.sks-keyservers.net

iQEcBAEBAwAGBQJNVy4VAAoJEE1wgkIhr9j3oHAH/RYUX9klqIdwAv7hMX8lcBIf
shHEpQ5d2mS1m0VKWSHxu++P5EDzEfnOY9de7u0TXdAEGdlfRf/OJ52cL9+pR3xR
v2i851mWQy1PRDigkBAKG+QGDOtapZxXBlJwfKZsdGrBoAjBsEOuxdvC2G+Imiq3
Ej6PMq6lHSZ5Dn173ZIKJUKA5TrSsFC58Niz3gze0JJQVvw/xXWgGfcs4+AQ14Ke
qTkLAsXJsQr8P4PFt+Uq5mJ0Ch4WfaRlAC7dLkBkZ5vSkn0EbU+4OlNhZRrwVd0d
6WjA2qEk9DztN780o1Bvv+Jk0bqnIDvhx9inXavbe/kSsCfukZgsf0jkEFJNUU8=
=RoVH
-----END PGP SIGNATURE-----

Re: [Twin-dev] twin-0.6.2 :: System freeze after load Socket Server module

[K_Wolf - G. <ad...@gm...>]

On 05/03/10 16:36, Massimiliano Ghilardi wrote:
> On Monday 03 May 2010, K_Wolf <ad...@gm...> wrote:
>   
>> Thanks Massimiliano,
>>
>> My apologies to all, because I forgot to send some useful data like these:
>>    
>> # uname -aa
>> Linux karma 2.6.32-gentoo-r8 #4 SMP Wed Apr 21 21:22:49 BRT 2010 x86_64
>> AMD Turion(tm) X2 Dual-Core Mobile RM-72 AuthenticAMD GNU/Linux
>>
>> And strace output when call the socket serve module:
>> .
>> .
>> .
>> [...]
>> open("/home/wolf/.TwinAuth", O_RDONLY)  = 5
>> read(5, "", 256)                        = 0
>> read(5, "", 256)                        = 0
>> read(5, "", 256)                        = 0
>> read(5, "", 256)                        = 0
>> read(5, "", 256)                        = 0
>> .
>> .
>> .
>> Until break signal.
>> I've atached the full log.
>>     
> Hello K_Wolf,
>
> thanks for the detailed information. It was crucial to understand and 
> reproduce your issue.
>
> Quick answer:
> rm /home/wolf/.TwinAuth
>
>
> Long answer: it's a bug in twin server.
> >From your strace.log it turns out that twin does not correctly handle
> corrupted ${HOME}/.TwinAuth files: in your case it is empty,
> while it should be 256 bytes long.x'
> Removing the file ${HOME}/.TwinAuth is an effective workaround,
> since it will be re-created next time you execute twin.
>
> I will fix this problem, and it will go in next version, but I am not sure
> it is worth to do a new release only for this bug.
> If you want I can send you a diff with the fix once I have it.
>
>
> I hope this helps,
>
> Massimiliano
>
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> Twin-develop mailing list
> Twi...@li...
> https://lists.sourceforge.net/lists/listinfo/twin-develop
>   
Thanks a lot, Massimiliano! It worked gracefully.
My powerful basic desktop is awesome now. :)

LONG LIFE TO TWIN!

Re: [Twin-dev] twin-0.6.2 :: System freeze after load Socket Server module

[Massimiliano G. <mas...@gm...>]

On Monday 03 May 2010, K_Wolf <ad...@gm...> wrote:
> Thanks Massimiliano,
>
> My apologies to all, because I forgot to send some useful data like these:
>    
> # uname -aa
> Linux karma 2.6.32-gentoo-r8 #4 SMP Wed Apr 21 21:22:49 BRT 2010 x86_64
> AMD Turion(tm) X2 Dual-Core Mobile RM-72 AuthenticAMD GNU/Linux
>
> And strace output when call the socket serve module:
> .
> .
> .
> [...]
> open("/home/wolf/.TwinAuth", O_RDONLY)  = 5
> read(5, "", 256)                        = 0
> read(5, "", 256)                        = 0
> read(5, "", 256)                        = 0
> read(5, "", 256)                        = 0
> read(5, "", 256)                        = 0
> .
> .
> .
> Until break signal.
> I've atached the full log.

Hello K_Wolf,

thanks for the detailed information. It was crucial to understand and 
reproduce your issue.

Quick answer:
rm /home/wolf/.TwinAuth


Long answer: it's a bug in twin server.
From your strace.log it turns out that twin does not correctly handle
corrupted ${HOME}/.TwinAuth files: in your case it is empty,
while it should be 256 bytes long.
Removing the file ${HOME}/.TwinAuth is an effective workaround,
since it will be re-created next time you execute twin.

I will fix this problem, and it will go in next version, but I am not sure
it is worth to do a new release only for this bug.
If you want I can send you a diff with the fix once I have it.


I hope this helps,

Massimiliano

Re: [Twin-dev] twin-0.6.2 :: System freeze after load Socket Server module

[Massimiliano G. <mas...@gm...>]

On Friday 30 April 2010, K_Wolf - Gmail wrote:
> Hi people, I'm new there and I really like text mode world. I'm serious.
> After download , extract, compile (default features) and install twin,
> I've experienced a weird problem:
>
> Running as root, twin is graceful. I can do everything with no strange
> behavior. Anyway, using my common user (non-admin) I can just run twin
> without "Socket Server module". If I try to load it, my system freeze
> (but don't display any ) and the CPU usage goes to 100%!!!
>

Hello K_Wolf,

can you provide the output produced by attaching 'strace' to twin when the 
problem happens?

That would definitely help.

Regards,

Massimiliano

[Twin-dev] twin-0.6.2 :: System freeze after load Socket Server module

[K_Wolf - G. <ad...@gm...>]

Hi people, I'm new there and I really like text mode world. I'm serious.
After download , extract, compile (default features) and install twin,
I've experienced a weird problem:

Running as root, twin is graceful. I can do everything with no strange
behavior. Anyway, using my common user (non-admin) I can just run twin
without "Socket Server module". If I try to load it, my system freeze
(but don't display any ) and the CPU usage goes to 100%!!!

Re: [Twin-dev] TWIN - missing user's environment variables when using twdm

[Robin A. <ro...@bi...>]

On Tuesday 29 September 2009, Massimiliano Ghilardi wrote:
> On Sunday 27 September 2009, Robin Atwood wrote:
> did you try to start twdm with the '--envrc' option?
> 
> It should do (almost) exactly what you need, i.e. tell twin to set it's own
> environment variables from the user's normal environment variables as first
> thing after a login through twdm.
> 
> I wrote 'almost' because what this option actually does is to instruct twin
> server to run the shell script system.twenvrc.sh (or, if it exists,
> $HOME/.twenvrc) at user login to gather user's environment variables, and
>  the default script only tries to read /etc/profile, $HOME/.profile and
>  $HOME/.bashrc, but not $HOME/.bash_profile.
> 
> In any case, all this stuff should matter ONLY if you start programs
>  directly from twin: starting them from a shell inside a terminal inside
>  twin SHOULD NOT give problems, as the shell is supposed to setup its own
>  environment when it starts.
> Please tell me if you have evidence that this is not the case.

Here is what I see in a twterm:

opal ~ # cat .twenvrc
source ~/.bash_profile
opal ~ # ps -ef | grep twdm
root     18718     1  0 23:07 ?        00:00:00 /usr/sbin/twdm --quiet --envrc 
--hw=tty@/dev/tty2,TERM=linux
root     18719     1  0 23:07 tty1     00:00:00 /usr/sbin/twdm --quiet --envrc 
--hw=tty@/dev/tty1,TERM=linux
root     32323 32083  0 23:50 pts/7    00:00:00 grep --colour=auto twdm
opal ~ # cat .bashrc | grep alias
# alias to save time!
alias ll="ls -lA"
alias rm="rm -i"
alias df="df -Th"
alias hist="history | grep"
opal ~ #

but if I type, say, "ll" I get "command not found". It looks like no scripts 
are being executed.

-Robin
-- 
----------------------------------------------------------------------
Robin Atwood.

"Ship me somewheres east of Suez, where the best is like the worst,
 Where there ain't no Ten Commandments an' a man can raise a thirst"
         from "Mandalay" by Rudyard Kipling
----------------------------------------------------------------------

Re: [Twin-dev] TWIN - Text Mode Windowing

[Massimiliano G. <mas...@gm...>]

On Sunday 27 September 2009, Eric Gillespie wrote:
>
> Darn. That reminds me. I'm still striking that gfx bug with the 0.6.1
> tree. It even shows up with the -hw=X driver too. I did find a way
> around it, sort of.
> * Run twdm on a spare tty (I normally use tty6)
> * log into that twdm as my normal user.
> * in addition, start up Xorg, logging into my favourite DE, and
> starting up another twdisplay pointing to the now-logged-in Xorg
>
> This means I have two twdisplays running, so that if Xorg goes down, I
> still have the text version running on tty6.

Hello Eric,

maybe I have missed some previous e-mail from you, because to my ears the 
above sounds like "this is how I worked around that gfx bug".
But unluckily I do not know (or remember) about the bug you're talking 
about... the only bug related to gfx that comes to my mind is this:

[Changelog.txt]
From 0.6.0 to 0.6.1: (Massimiliano Ghilardi, Milos Jakubicek, 17 Feb 2009)
        * Fixed uninitialized variable bug in --hw=gfx introduced in 0.6.0,
          sometimes causing twin/twdisplay not to draw anything when
          using --hw=gfx, especially on 64 bit archs.

So, if you can provide more details about the bug you mention I will surely 
appreciate - and I will be able to try fixing it.

> The side effect of this is:  if I have a drawing error like you have
> described while on X, I switch back to tty6, and then hit a key -
> doesn't seem to matter which key, and then I switch back to X. The
> problem doesn't always happen, but it's often enough that it bugs me.
> It seems to only happen when I try running curses-based programs -
> midnight commander and cdialog seem to be the main things that trip
> drawing to a twdisplay up.

If I understand correctly, you are describing an unwanted (or simply buggy) 
side-effect of your workaround. I guess that your sentence
> if I have a drawing error like you have described while on X

refers to this report from Robin Atwood:
> > all the box characters get garbled.
> > I must enter "LANG=en_GB" to get the correct 
> > translation. It's a minor point, but a bit strange.

Again, if you can provide more details, including the version (or versions) of 
twin where this bug happens, I will try to analyze and fix it.

> I haven't yet tried to see if the same bug is present in the new 0.6.2
> tree, though I only have a 32-bit machine. Can't wait for proper
> UTF-8.
>
> Cheers, Eric Gillespie,
> The Viking

Regards,

Massimiliano

Re: [Twin-dev] TWIN - missing user's environment variables when using twdm

[Massimiliano G. <mas...@gm...>]

On Sunday 27 September 2009, Robin Atwood wrote:
> On Saturday 26 September 2009, you wrote:
> > I have just installed it and it seems to work fine. The only peculiarity
> > is with the character encoding and twdm. My locale is set so
> > "LANG=en_GB.utf8". If I login to a console and type twin and then execute
> > an ncurses app (like "mc") in a terminal, the encoding looks fine; I get
> > all the boxes, etc. If I login via twdm and execute mc, all the box
> > characters get garbled. I must enter "LANG=en_GB " to get the correct
> > translation. It's a minor point, but a bit strange.
>
> OK, I have more info. My /root/.bash_profile contains export LANG="en_GB-
> ISO8859", which I must have set up a long time ago and forgotten about. The
> problem is that when I start twin via twdm, the script is not executed.
> Starting twin from a prompt gives the expected behaviour. None of the
> aliases in ~/.bashrc get set either, when I use twdm.
>
> HTH
> -Robin

Hello Robin,

did you try to start twdm with the '--envrc' option?

It should do (almost) exactly what you need, i.e. tell twin to set it's own 
environment variables from the user's normal environment variables as first 
thing after a login through twdm.

I wrote 'almost' because what this option actually does is to instruct twin 
server to run the shell script system.twenvrc.sh (or, if it exists, 
$HOME/.twenvrc) at user login to gather user's environment variables, and the 
default script only tries to read /etc/profile, $HOME/.profile and 
$HOME/.bashrc, but not $HOME/.bash_profile.

In any case, all this stuff should matter ONLY if you start programs directly 
from twin: starting them from a shell inside a terminal inside twin
SHOULD NOT give problems, as the shell is supposed to setup its own 
environment when it starts.
Please tell me if you have evidence that this is not the case.

Best regards,

Massimiliano

Re: [Twin-dev] TWIN - Text Mode Windowing

[Massimiliano G. <mas...@gm...>]

On Saturday 26 September 2009, Robin Atwood wrote:
> On Saturday 26 September 2009, Massimiliano Ghilardi wrote:
> > I am glad you appreciate TWIN.
>
> Max -
> Did you ever make any progress with twin-0.6 and 64 bit systems? I would
> still like to upgrade someday from 0.4.6. Lack of utf-8 support is a bit of
> a hassle.
>
> Regards
> -Robin

Hello Robin,

you remind me that I forgot to release a new version of twin
after fixing 64-bit support... it seems I am getting old ;-)

Anyway, I just released twin-0.6.2 now, with fully working 64-bit support
(I use it regularly and I did not notice any problems).
For the moment, it is available only from http://linuz.sns.it/~max/twin/

I should upload it also to sourceforge, I will do that as soon as I have a bit 
more time.

Best regards,

Massimiliano

Re: [Twin-dev] TWIN - Text Mode Windowing

[Robin A. <ro...@bi...>]

On Saturday 26 September 2009, Massimiliano Ghilardi wrote:
> I am glad you appreciate TWIN.

Max -
Did you ever make any progress with twin-0.6 and 64 bit systems? I would still 
like to upgrade someday from 0.4.6. Lack of utf-8 support is a bit of a 
hassle.

Regards
-Robin
-- 
----------------------------------------------------------------------
Robin Atwood.

"Ship me somewheres east of Suez, where the best is like the worst,
 Where there ain't no Ten Commandments an' a man can raise a thirst"
         from "Mandalay" by Rudyard Kipling
----------------------------------------------------------------------