UUdecode and NTemacs

UUdecode and NTemacs

[Fabrice POPINEAU]

Hi,

The following code  in gnus-uu.el (5.4.36)  needs some modification to
work under NTemacs:

	  ;; Start a new uudecoding process.
	  (let ((cdir default-directory))
	    (unwind-protect
		(progn
		  (cd gnus-uu-work-dir)
		  (setq gnus-uu-uudecode-process
			(start-process
			 "*uudecode*"
			 (get-buffer-create gnus-uu-output-buffer-name)
;			 shell-file-name shell-command-switch
;			 (format "cd %s %s uudecode" gnus-uu-work-dir
;				 gnus-shell-command-separator))))
			 "uudecode")))
	      (cd cdir)))

There are 2 problems:

- gnus-uu-work-dir is "f:/tmp/gnusa00195" and if your shell is "cmd",
it can't 'cd' to some path with '/', it needs '\'
- anyway, I'm not sure NTemacs can accept "cd f:\\tmp\\gnusa00195 &&
uudecode" as a program to start.

As there was a (cd gnus-uu-work-dir) just before, is it useful to keep
the "cd tmpdir &&" into the starting process ?

Also, under NTemacs, shelex.exe would be a better viewer for most of
the files you want to decode.

Fabrice POPINEAU
------------------------
e-mail:         popineau@ese-metz.fr    |  "God is real ...              
	Fabrice.POPINEAU@supelec.fr     |
voice-mail:   +33 (0) 387764715         |          ... unless integer ?"
surface-mail: Supelec, 2 rue E. Belin,  |
	      F-57078 Metz Cedex 3      |

Re: UUdecode and NTemacs

[Lars Magne Ingebrigtsen]

Fabrice POPINEAU <popineau@esemetz.ese-metz.fr> writes:

> - gnus-uu-work-dir is "f:/tmp/gnusa00195" and if your shell is "cmd",
> it can't 'cd' to some path with '/', it needs '\'

I though GatesOS treated "/" as "\" just about anywhere?

> - anyway, I'm not sure NTemacs can accept "cd f:\\tmp\\gnusa00195 &&
> uudecode" as a program to start.

gnus-shell-command-separator => ";" by default...

> As there was a (cd gnus-uu-work-dir) just before, is it useful to keep
> the "cd tmpdir &&" into the starting process ?

I'm not altogether sure, actually...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: UUdecode and NTemacs

[Dewey M. Sasser]

>>>>> "Lars" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

    Lars> Fabrice POPINEAU <popineau@esemetz.ese-metz.fr> writes:
    >> - gnus-uu-work-dir is "f:/tmp/gnusa00195" and if your shell is
    >> "cmd", it can't 'cd' to some path with '/', it needs '\'

    Lars> I though GatesOS treated "/" as "\" just about anywhere?

Except the shell.  Anywhere through the API is OK, but not the shell.
Don't worry, I'm sure it's just an oversight and they'll restrict it
to only '\' in the next release.  (Pardon the cynicism.)

The problem is that the built-in shell commands (and most others)
treat '/' as the option character.

One work around is to get another shell.

    >> - anyway, I'm not sure NTemacs can accept "cd
    >> f:\\tmp\\gnusa00195 && uudecode" as a program to start.

    Lars> gnus-shell-command-separator => ";" by default...

I *think* you can separate commands for the NT shell with either "^"
or "&" (only one of them, but I don't remember which).


-- 
Dewey M. Sasser 			voice: (617) 494-6000 
dewey@newvision.com                     PGP Key from public servers  
                                        PGP mail preferred.
---
The heart has its reasons which reason knows nothing of
  			  -- Pascal

Re: UUdecode and NTemacs

[Stainless Steel Rat]

-----BEGIN PGP SIGNED MESSAGE-----

>>>>> "tGBGnB" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

tGBGnB> Fabrice POPINEAU <popineau@esemetz.ese-metz.fr> writes:
>> - gnus-uu-work-dir is "f:/tmp/gnusa00195" and if your shell is "cmd",
>> it can't 'cd' to some path with '/', it needs '\'
tGBGnB> I though GatesOS treated "/" as "\" just about anywhere?

No, it doesn't.  "/" is a command switch (analogous to "-" in Unix).  GNU
Emacs compiled for Win32 is supposed to translate "/" in paths to "\".  On
this, I think Fabrice goofed; GNU Emacs will pass the correct directory
separator to the OS.

>> - anyway, I'm not sure NTemacs can accept "cd f:\\tmp\\gnusa00195 &&
>> uudecode" as a program to start.
tGBGnB> gnus-shell-command-separator => ";" by default...

On this Fabrice is correct; neither COMMAND.COM nor CMD.EXE (the two
Windows command interpreters) support CD with a drive spec.  4DOS and 4NT
do with the CDD command, but they are comercial add-on shells.  Also,
COMMAND.COM does not have a command separator character at all; not sure
about CMD.EXE.  4DOS and 4NT do; it is "^".

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQCVAwUBM0WL6Z6VRH7BJMxHAQFyAQQAm7vGg6meW45uYTX8+m79Le3KSgZUxMw7
bk5Oz6Zw1BIhU6P5AE0xzoQgEp9X0FVUkkHP9WaGJpRRyq70kMxh83QKkJ4pwKFF
aptHNFbtaAw9ECa29JVcgRXwthrJybAOvulMc0DW7E7gS+Hn7bdoWJixAGB9IEZm
TeCUV6x7gr8=
=7nu6
-----END PGP SIGNATURE-----
-- 
Rat <ratinox@peorth.gweep.net>    \ Caution: Happy Fun Ball may suddenly
PGP Key: at a key server near you! \ accelerate to dangerous speeds.
                                    \ 

Re: UUdecode and NTemacs

[Lars Magne Ingebrigtsen]

Stainless Steel Rat <ratinox@peorth.gweep.net> writes:

> On this Fabrice is correct; neither COMMAND.COM nor CMD.EXE (the two
> Windows command interpreters) support CD with a drive spec.  4DOS and 4NT
> do with the CDD command, but they are comercial add-on shells.  Also,
> COMMAND.COM does not have a command separator character at all; not sure
> about CMD.EXE.  4DOS and 4NT do; it is "^".

Well, I don't use GatesOS.  Could someone mail me patches for these
things? 

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: UUdecode and NTemacs

[Fabrice POPINEAU]

>>>>> "Lars" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

    Lars> Stainless Steel Rat <ratinox@peorth.gweep.net> writes:
    >> On this Fabrice is correct; neither COMMAND.COM nor CMD.EXE
    >> (the two Windows command interpreters) support CD with a drive
    >> spec.  

That may be the explanation for what follows.

    >> 4DOS and 4NT do with the CDD command, but they are comercial
    >> add-on shells.  Also, COMMAND.COM does not have a command
    >> separator character at all; not sure about CMD.EXE.  4DOS and
    >> 4NT do; it is "^".

    Lars> Well, I don't use GatesOS.  

I wonder if I will ever be forgiven for running it :-)

    Lars> Could someone mail me patches for these things?

I tried to debug this some weeks ago and was puzzled. Last weekn, I
thought that I had the solution, and I'm still puzzled. In fact, the
standard code *almost* works, provided you have:

(setq gnus-shell-command-separator "&")

But I'm getting various errors after the first "X u" processing, like:

- can't remove gnus-uu-work-dir
- try to select destroyed buffer when decoding in another group,
- ...

What is really strange is that if I apply the following patch,
everything is OK:

--- gnus-uu.el~ Thu Mar 13 01:29:04 1997
+++ g:/local/emacs/site-lisp/gnus/lisp/gnus-uu.el       Mon Apr 07 09:50:16 1997

@@ -1416,9 +1416,10 @@
                        (start-process
                         "*uudecode*"
                         (get-buffer-create gnus-uu-output-buffer-name)
-                        shell-file-name shell-command-switch
-                        (format "cd %s %s uudecode" gnus-uu-work-dir
-                                gnus-shell-command-separator))))
+;                       shell-file-name shell-command-switch
+;                       (format "cd %s %s uudecode" gnus-uu-work-dir
+;                               gnus-shell-command-separator))))
+                        "uudecode")))
              (cd cdir)))
          (set-process-sentinel
           gnus-uu-uudecode-process 'gnus-uu-uudecode-sentinel)

May be this code should be conditionalized by the shell-file-name's
value:

(string-match "cmd\\.exe" shell-file-name)


-- 
Fabrice POPINEAU
------------------------
e-mail:         popineau@ese-metz.fr    |  "God is real ...              
	Fabrice.POPINEAU@supelec.fr     |
voice-mail:   +33 (0) 387764715         |          ... unless integer ?"
surface-mail: Supelec, 2 rue E. Belin,  |
	      F-57078 Metz Cedex 3      |

Re: UUdecode and NTemacs

[Stainless Steel Rat]

-----BEGIN PGP SIGNED MESSAGE-----

>>>>> "FP" == Fabrice POPINEAU <popineau@esemetz.ese-metz.fr> writes:

FP> I tried to debug this some weeks ago and was puzzled. Last weekn, I
FP> thought that I had the solution, and I'm still puzzled. In fact, the
FP> standard code *almost* works, provided you have:

FP> (setq gnus-shell-command-separator "&")

But unfortunately this will still break on Windows 95 machines because
there is no command separator in COMMAND.COM.

Anyway, the drive letter problem can probably be fixed by using UNC
pathnames.  That is, instead of 'C:\FOO' you use '\\C\FOO'.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQCVAwUBM0kj7p6VRH7BJMxHAQEyHQP9HO526zWtXalqk5vEV+uaL3NpkMAp9ur3
UO72lcNlEWIejD96yxVrMDqcLhoAWchiky2X6YWxURop2umZR017dlIb4AYogTM6
dHH36pjdszb6zJz3h18S3bT74n+TTyd7rUXPUIn5NqVrvkxLkYwhPWTdWm1AWLgr
875ZpAqy3go=
=zCZT
-----END PGP SIGNATURE-----
-- 
Rat <ratinox@peorth.gweep.net>    \ If Happy Fun Ball begins to smoke, get
PGP Key: at a key server near you! \ away immediately. Seek shelter and cover
                                    \ head.

Re: UUdecode and NTemacs

[Andy Eskilsson]

It just struck me.. there exists some dos/win95/nt[1] ports of for
example bash.. this would maybe be an alternative?

For more info on bash throw your favourite browser at:
	http://www.neongenesis.com/~jack/djgpp-work/beta/index.html

Look at the bottom of the page for the latest.. 

	/Andy

[1] There are some problems with djgpp and NT4, I am not 100% sure
    where it lies, but look through the djgpp newsgroup for more.

Re: UUdecode and NTemacs

[Stainless Steel Rat]

-----BEGIN PGP SIGNED MESSAGE-----

>>>>> "AE" == Andy Eskilsson <flognat@fukt.hk-r.se> writes:

AE> For more info on bash throw your favourite browser at:
AE> 	http://www.neongenesis.com/~jack/djgpp-work/beta/index.html

DJGPP builds DOS programs, not Win32 console programs.  This is a problem
because GNU Emacs on Win32 platforms can have exactly one DOS program
running underneath it.  Start bash.exe and try to invoke your DOS uudecode
program and Emacs will lock up, and you will not be able to break out.

There are several Win32 console mode shells out there (the NTEmacs FAQ
mentions at least two), but like 4DOS and 4NT they are not standard shells.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQCVAwUBM0kkop6VRH7BJMxHAQFmwQP/W1K7y7wO8XsPa3IUo8feHFwpTO0nPST5
OK6c2JJ7LXzoOUXzsgZIAX3e89DB1Yw2Pbpx3CH33RUaFQQfhoFkGFJ7/Wz65n+h
IhQUhQoqGsMzaTXWnYLf8Up1vYYxpwpv17iLgTqOrskbMvLbU3TUUTWvOqFKdne/
JyhmY+acgkI=
=NOeY
-----END PGP SIGNATURE-----
-- 
Rat <ratinox@peorth.gweep.net>    \ Warning: pregnant women, the elderly, and
PGP Key: at a key server near you! \ children under 10 should avoid prolonged
                                    \ exposure to Happy Fun Ball.

Re: UUdecode and NTemacs

[Scott Blachowicz]

Stainless Steel Rat <ratinox@peorth.gweep.net> wrote:

> DJGPP builds DOS programs, not Win32 console programs.  This is a problem
> because GNU Emacs on Win32 platforms can have exactly one DOS program
> running underneath it.  Start bash.exe and try to invoke your DOS uudecode
> program and Emacs will lock up, and you will not be able to break out.

It also means that you can't see long filenames from NT, but that you can from 
Win95.  The Win95 "DOS" portion identifies itself as DOS 7.0 and handles the 
DPMI INT mumble that gets LFN information.  The NT "DOS" stuff IDs as DOS 5.5 
(I think) and doesn't handle that DPMI interrupt.  I've run into this problem 
trying to migrate a GNUmake-based build environment from Win95 to NT (3.51 and 
4.0).  If anyone knows otherwise, I'd like to hear about it...

Scott Blachowicz  Ph: 206/283-8802x240   Mathsoft (Data Analysis Products Div)
                                         1700 Westlake Ave N #500
scott@statsci.com                        Seattle, WA USA   98109
Scott.Blachowicz@seaslug.org

Re: UUdecode and NTemacs

[Stainless Steel Rat]

-----BEGIN PGP SIGNED MESSAGE-----

>>>>> "SB" == Scott Blachowicz <scott@bloke.statsci.com> writes:

SB> It also means that you can't see long filenames from NT, but that you
SB> can from Win95.  The Win95 "DOS" portion identifies itself as DOS 7.0
SB> and handles the DPMI INT mumble that gets LFN information.

I think DJGPP-compiled programs on NT 4.0 should get the LFN information,
but you may need to specifically tell the extender to get them.  Do that by
changing the line in DJGPP.ENV that reads "+LFN=n" to "+LFN=y".

Still, it does no good under Emacs, as I said.  If you are going to use GNU
Emacs on a Win32 box and you are going to use inferior processes, you need
a Win32 console mode shell.  There is really no way around it.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQCUAwUBM0lELp6VRH7BJMxHAQHzqgP1Hccp/BL9gwBDPFDLSRE9QONacPp5Qe5g
4qoMeoZZnCHTxkIOCXxlDy+lQ/jEO4BWnfEdHQHIrIC0UtzMDyXghFbox1TyI84+
qG919JH/pLKRuTdzQPSRH2jZVzhmlfG9QR4QSLW+dL8xuT4JYcZZvIo2ZwFI6X+1
bhgsvZzuOA==
=pSG6
-----END PGP SIGNATURE-----
-- 
Rat <ratinox@peorth.gweep.net>    \ Happy Fun Ball contains a liquid core,
PGP Key: at a key server near you! \ which, if exposed due to rupture, should
                                    \ not be touched, inhaled, or looked at.

Re: UUdecode and NTemacs

[Scott Blachowicz]

Stainless Steel Rat <ratinox@peorth.gweep.net> wrote:

> I think DJGPP-compiled programs on NT 4.0 should get the LFN information,
> but you may need to specifically tell the extender to get them.  Do that by
> changing the line in DJGPP.ENV that reads "+LFN=n" to "+LFN=y".

Nope...doesn't work.  I was specifically doing all that stuff - same system, 
same source files, same DJGPP files, dual booting between Win95 & NT4 - it 
didn't work.  It does a DPMI INT 21h (in a _use_lfn.c file) to check for 
availability of LFN handling (and I believe that test fails for NT4).

> Still, it does no good under Emacs, as I said.

True, but I could still use a GNU make that works the same under Win95 & NT4 
and understands my long filenames...I haven't tried my MSVC-build of the 
GNU-distributed make-3.75.tar.gz yet, but will eventually.  Next step after 
that might be the cygnus gnu-win32 environment, but that involves more changes 
to my makefiles, I think, to get their forward-slashed paths...but this is 
getting a bit off topic for this list...

Scott Blachowicz  Ph: 206/283-8802x240   Mathsoft (Data Analysis Products Div)
                                         1700 Westlake Ave N #500
scott@statsci.com                        Seattle, WA USA   98109
Scott.Blachowicz@seaslug.org

Re: UUdecode and NTemacs

[Richard Hoskins]

Stainless Steel Rat <ratinox@peorth.gweep.net> writes:

> DJGPP builds DOS programs, not Win32 console programs.  This is a problem
> because GNU Emacs on Win32 platforms can have exactly one DOS program
> running underneath it.  Start bash.exe and try to invoke your DOS uudecode
> program and Emacs will lock up, and you will not be able to break out.
> 
> There are several Win32 console mode shells out there (the NTEmacs FAQ
> mentions at least two), but like 4DOS and 4NT they are not standard shells.
> 

Cygnus has a Win32 bash.  I have used it with NTEmacs on Win 95 with
few significant problems.

 http://www.cygnus.com/misc/gnu-win32/

Re: UUdecode and NTemacs

[Andrew Innes]

On Fri, 4 Apr 1997 10:59:10 +0200 (MET DST), Fabrice POPINEAU <popineau@esemetz.ese-metz.fr> said:
>Hi,
>
>The following code  in gnus-uu.el (5.4.36)  needs some modification to
>work under NTemacs:
>
>	  ;; Start a new uudecoding process.
>	  (let ((cdir default-directory))
>	    (unwind-protect
>		(progn
>		  (cd gnus-uu-work-dir)
>		  (setq gnus-uu-uudecode-process
>			(start-process
>			 "*uudecode*"
>			 (get-buffer-create gnus-uu-output-buffer-name)
>;			 shell-file-name shell-command-switch
>;			 (format "cd %s %s uudecode" gnus-uu-work-dir
>;				 gnus-shell-command-separator))))
>			 "uudecode")))
>	      (cd cdir)))
>
>There are 2 problems:
>
>- gnus-uu-work-dir is "f:/tmp/gnusa00195" and if your shell is "cmd",
>it can't 'cd' to some path with '/', it needs '\'

You can solve this by binding directory-sep-char to '\' while calling
expand-file-name; this will give the correct format on Windows 95/NT.

>- anyway, I'm not sure NTemacs can accept "cd f:\\tmp\\gnusa00195 &&
>uudecode" as a program to start.

If you are "on" drive F at the time, then this will work on Windows NT.
It won't work on Windows 95, using the standard shell.

>As there was a (cd gnus-uu-work-dir) just before, is it useful to keep
>the "cd tmpdir &&" into the starting process ?

It should be a no-op, I think.  If the original code above works without
that, then you have your solution.

AndrewI

-- 
Andrew Innes                    Software Developer, Electronic Publishing
Harlequin Limited, Longstanton House, Longstanton, Cambridge CB4 5BU, U.K.
Tel: +44 1954 785468  Fax: +44 1954 785444  WWW: http://www.harlequin.com