From caixianchao at cn.fujitsu.com  Wed Mar  5 06:32:23 2008
From: caixianchao at cn.fujitsu.com (Cai Xianchao)
Date: Wed, 05 Mar 2008 12:32:23 +0800
Subject: [PATCH]It should not be allowed if environment variable begins with
	a digit
Message-ID: <47CE2257.2050606@cn.fujitsu.com>

hi, all

'environment variable begins with a digit' should not be allowed in tcsh,
it is not allowed in both bash and ksh.  Shell variable cann't begin with
a digit too.

version?tcsh-6.15.00

Signed-off-by: Cai Xianchao <caixianchao at cn.fujitsu.com>

diff --git a/sh.func.c b/sh.func.c
index 92e2447..9f08dc4 100644
--- a/sh.func.c
+++ b/sh.func.c
@@ -1239,6 +1239,8 @@ dosetenv(Char **v, struct command *c)
    }

    vp = *v++;
+    if (!letter(*vp))
+        stderror(ERR_NAME | ERR_VARBEGIN);

    lp = vp;


From lizf at cn.fujitsu.com  Wed Mar  5 06:50:14 2008
From: lizf at cn.fujitsu.com (Li Zefan)
Date: Wed, 05 Mar 2008 13:50:14 +0900
Subject: [Spam-fortigate] [PATCH]It should not be allowed if environment
	variable begins with	a digit
In-Reply-To: <47CE2257.2050606@cn.fujitsu.com>
References: <47CE2257.2050606@cn.fujitsu.com>
Message-ID: <47CE2686.7080804@cn.fujitsu.com>

Cai Xianchao wrote:
> hi, all
>
> 'environment variable begins with a digit' should not be allowed in tcsh,
> it is not allowed in both bash and ksh.  Shell variable cann't begin with
> a digit too.
>
>   

The patch is obviously wrong. You are disallowing variables that don't 
begin with
a letter.

And why can't an env variable begins with a digit? It's not convincing Just
because bash and ksh disallow it. Does it cause any problems in the real
world?

> version?tcsh-6.15.00
>
> Signed-off-by: Cai Xianchao <caixianchao at cn.fujitsu.com>
>
> diff --git a/sh.func.c b/sh.func.c
> index 92e2447..9f08dc4 100644
> --- a/sh.func.c
> +++ b/sh.func.c
> @@ -1239,6 +1239,8 @@ dosetenv(Char **v, struct command *c)
>     }
>
>     vp = *v++;
> +    if (!letter(*vp))
> +        stderror(ERR_NAME | ERR_VARBEGIN);
>
>     lp = vp;
>
> _______________________________________________
> Tcsh mailing list
> Tcsh at mx.gw.com
> http://mx.gw.com/mailman/listinfo/tcsh
>   


From lizf at cn.fujitsu.com  Wed Mar  5 07:06:36 2008
From: lizf at cn.fujitsu.com (Li Zefan)
Date: Wed, 05 Mar 2008 14:06:36 +0900
Subject: [Spam-fortigate] Re: [Spam-fortigate] [PATCH]It should not be
	allowed if environment	variable begins with	a digit
In-Reply-To: <47CE2686.7080804@cn.fujitsu.com>
References: <47CE2257.2050606@cn.fujitsu.com> <47CE2686.7080804@cn.fujitsu.com>
Message-ID: <47CE2A5C.9050902@cn.fujitsu.com>

Li Zefan ??:
> Cai Xianchao wrote:
>   
>> hi, all
>>
>> 'environment variable begins with a digit' should not be allowed in tcsh,
>> it is not allowed in both bash and ksh.  Shell variable cann't begin with
>> a digit too.
>>
>>   
>>     
>
> The patch is obviously wrong. You are disallowing variables that don't 
> begin with
> a letter.
>
> And why can't an env variable begins with a digit? It's not convincing Just
> because bash and ksh disallow it. Does it cause any problems in the real
> world?
>
>   

Is this a compatibility break to add this restriction to env variable?

>> version?tcsh-6.15.00
>>
>> Signed-off-by: Cai Xianchao <caixianchao at cn.fujitsu.com>
>>
>> diff --git a/sh.func.c b/sh.func.c
>> index 92e2447..9f08dc4 100644
>> --- a/sh.func.c
>> +++ b/sh.func.c
>> @@ -1239,6 +1239,8 @@ dosetenv(Char **v, struct command *c)
>>     }
>>
>>     vp = *v++;
>> +    if (!letter(*vp))
>> +        stderror(ERR_NAME | ERR_VARBEGIN);
>>
>>     lp = vp;
>>
>> _______________________________________________
>> Tcsh mailing list
>> Tcsh at mx.gw.com
>> http://mx.gw.com/mailman/listinfo/tcsh
>>   
>>     
>
> _______________________________________________
> Tcsh mailing list
> Tcsh at mx.gw.com
> http://mx.gw.com/mailman/listinfo/tcsh
>   


From caixianchao at cn.fujitsu.com  Wed Mar  5 07:32:24 2008
From: caixianchao at cn.fujitsu.com (Cai Xianchao)
Date: Wed, 05 Mar 2008 13:32:24 +0800
Subject: [Spam-fortigate] [PATCH]It should not be allowed if environment
	variable begins with	a digit
In-Reply-To: <47CE2686.7080804@cn.fujitsu.com>
References: <47CE2257.2050606@cn.fujitsu.com> <47CE2686.7080804@cn.fujitsu.com>
Message-ID: <47CE3068.4080002@cn.fujitsu.com>

Li Zefan ??:
> Cai Xianchao wrote:
>> hi, all
>>
>> 'environment variable begins with a digit' should not be allowed in 
>> tcsh,
>> it is not allowed in both bash and ksh.  Shell variable cann't begin 
>> with
>> a digit too.
>>
>>   
>
> The patch is obviously wrong. You are disallowing variables that don't 
> begin with
> a letter.
>
> And why can't an env variable begins with a digit? It's not convincing 
> Just
> because bash and ksh disallow it. Does it cause any problems in the real
> world?

The problem is that env variable cann't be echoed correctly if it begins
with a digit.

For example:

[test at RHEL-5 ~]$ setenv 1t a
[test at RHEL-5 ~]$ echo $1t
t

The result is not 'a' as I expected. The '$1' is parsed as $argv[1] while
not the front part of the env variable '1t'.



>
>> version?tcsh-6.15.00
>>
>> Signed-off-by: Cai Xianchao <caixianchao at cn.fujitsu.com>
>>
>> diff --git a/sh.func.c b/sh.func.c
>> index 92e2447..9f08dc4 100644
>> --- a/sh.func.c
>> +++ b/sh.func.c
>> @@ -1239,6 +1239,8 @@ dosetenv(Char **v, struct command *c)
>>     }
>>
>>     vp = *v++;
>> +    if (!letter(*vp))
>> +        stderror(ERR_NAME | ERR_VARBEGIN);
>>
>>     lp = vp;
>>
>> _______________________________________________
>> Tcsh mailing list
>> Tcsh at mx.gw.com
>> http://mx.gw.com/mailman/listinfo/tcsh
>>   
>
>
>


From vcrhonek at redhat.com  Wed Mar  5 13:22:25 2008
From: vcrhonek at redhat.com (Vitezslav Crhonek)
Date: Wed, 05 Mar 2008 12:22:25 +0100
Subject: tcsh man page fixes
Message-ID: <47CE8271.40907@redhat.com>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

there are several Red Hat Bugzilla entries opened by Cai Xianchao
<caixianchao at cn.fujitsu.com> against tcsh man pages. All issues
are in latest upstream tcsh too, so I would like to discuss it
here.

1) In the description of builtin "set", "The second form sets name
to the null string" appears twice and the second is redundant.
(https://bugzilla.redhat.com/show_bug.cgi?id=430459)

diff --git a/tcsh.1 b/tcsh.1
- --- a/tcsh.1
+++ b/tcsh.1
@@ -3199,9 +3199,7 @@ specified, set only unique words keeping their order.
~ \-f prefers the first occurrence of a word, and \-l the last.
~ The fifth form sets the \fIindex\fR'th component of name to \fIword\fR;
~ this component must already exist.
- -The sixth form lists only the names of all shell variables that are read-only.
- -The seventh form makes \fIname\fR read-only, whether or not it has a value.
- -The second form sets \fIname\fR to the null string.
+The sixth form lists only the names of all shell variables that are read-only. The seventh form makes \fIname\fR read-only, whether or not it has a value.
~ The eighth form is the same as the third form, but
~ make \fIname\fR read-only at the same time.

It's true and should be fixed.

2) There is no influence to run a set-user ID script whether or not
option "-b" is used.
(https://bugzilla.redhat.com/show_bug.cgi?id=433908)

diff --git a/tcsh.man b/tcsh.man
index 0c9c3b6..75fb6d1 100644
- --- a/tcsh.man
+++ b/tcsh.man
@@ -108,8 +108,7 @@ The rest of the flag arguments are interpreted as follows:
~ Forces a ``break'' from option processing, causing any
~ further shell arguments to be treated as non-option arguments.  The remaining
~ arguments will not be interpreted as shell options.  This may be used to pass
- -options to a shell script without confusion or possible subterfuge.  The shell
- -will not run a set-user ID script without this option.
+options to a shell script without confusion or possible subterfuge.
~ .TP 4
~ .B \-c
~ Commands are read from the following argument (which must be present, and

There's no influence of "-b" I can see, so again, I think it should be fixed.

3) ", and must be a single argument" should be deleted
(https://bugzilla.redhat.com/show_bug.cgi?id=433914)

diff --git a/tcsh.man b/tcsh.man
index 0c9c3b6..b171aad 100644
- --- a/tcsh.man
+++ b/tcsh.man
@@ -112,8 +112,7 @@ options to a shell script without confusion or possible subterfuge.  The shell
~ will not run a set-user ID script without this option.
~ .TP 4
~ .B \-c
- -Commands are read from the following argument (which must be present, and
- -must be a single argument),
+Commands are read from the following argument (which must be present),
~ stored in the \fBcommand\fR shell variable for reference, and executed.
~ Any remaining arguments are placed in the \fBargv\fR shell variable.
~ .TP 4

I know there was a discuss about this here, but not about man pages.
I disagree with Cai and I think the statement in the man page is correct.
You can see out arguments on RH Bugzilla.

4) '\' can not be used to quote all delimiters

diff --git a/tcsh.man b/tcsh.man
index 0c9c3b6..3e389f4 100644
- --- a/tcsh.man
+++ b/tcsh.man
@@ -1035,7 +1035,7 @@ Substitute \fIl\fR for \fIr\fR.
~ \fIl\fR is simply a string like \fIr\fR, not a regular expression as in
~ the eponymous \fIed\fR(1) command.
~ Any character may be used as the delimiter in place of `/';
- -a `\\' can be used to quote the delimiter inside \fIl\fR and \fIr\fR.
+a `\\' can be used to quote the delimiter expect `(', `)', `|' and `>' inside \fIl\fR and \fIr\fR.
~ The character `&' in the \fIr\fR is replaced by \fIl\fR; `\\' also quotes `&'.
~ If \fIl\fR is empty (``''), the \fIl\fR from a previous substitution or the
~ \fIs\fR from a previous search or event number in event specification is used.

I'm not sure how this exactly works...

Sorry for longer post, but I won't fix it in Fedora/RHEL before it's commited
in upstream. All patches come from Cai.

Best regards,
Vitezslav Crhonek
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iD8DBQFHzoJws1VJEvNweZcRAniqAJ9X67KaUJO0SfzU4w6LbotlDzQcPwCdEoYv
8/5re0Pq6fLR0Fg9nqLrD1g=
=NkJk
-----END PGP SIGNATURE-----


From caixianchao at cn.fujitsu.com  Tue Mar 11 09:40:36 2008
From: caixianchao at cn.fujitsu.com (Cai Xianchao)
Date: Tue, 11 Mar 2008 15:40:36 +0800
Subject: [PATCH] '$number' should be equivalent to '$argv[number]' when number
	is out of range
Message-ID: <47D63774.5030702@cn.fujitsu.com>

hi,all

The manpage says that "$number Equivalent to ?$argv[number]?". They are
indeed the same if "num" is not larger than the number of argument, but
not if it is out of range, An error message should output in '$number'
as '$argv[number]'.

(package: tcsh-6.15.00)

I did it as follows:

[test at localhost ~]$ cat argv_num.tcsh
#!/bin/tcsh

echo $argv[1]

[test at localhost ~]$ cat num.tcsh
#!/bin/tcsh

echo $1

[test at localhost ~]$ tcsh argv_num.tcsh
argv: Subscript out of range.

[test at localhost ~]$ tcsh num.tcsh


Signed-off-by: Cai Xianchao <caixianchao at cn.fujitsu.com>

diff --git a/sh.dol.c b/sh.dol.c
index 73ac3b2..40c228a 100644
--- a/sh.dol.c
+++ b/sh.dol.c
@@ -507,9 +507,11 @@ Dgetdol(void)
         subscr = subscr * 10 + c - '0';
         c = DgetC(0);
         } while (c != DEOF && Isdigit(c));
-        unDredc(c);
-        if (subscr < 0)
+        vp = adrof(STRargv);
+        pb = blklen(vp->vec);
+        if (subscr < 0 || subscr > upb)
         stderror(ERR_RANGE);
+        unDredc(c);
         if (subscr == 0) {
         if (bitset) {
             dolp = dolzero ? STR1 : STR0;
@@ -535,7 +537,6 @@ Dgetdol(void)
         if (length)
         stderror(ERR_NOTALLOWED, "$%<num>");
 #endif
-        vp = adrof(STRargv);
         if (vp == 0) {
         vp = &nulargv;
         cleanup_until(&name);




From karim.hamidou at reseau.eseo.fr  Tue Mar 18 18:34:02 2008
From: karim.hamidou at reseau.eseo.fr (karim.hamidou at reseau.eseo.fr)
Date: Tue, 18 Mar 2008 17:34:02 +0100 (CET)
Subject: A TOUR file for tcsh
Message-ID: <45459.194.254.136.2.1205858042.squirrel@webmail.reseau.eseo.fr>

Hi,

I remember that there was in the ash distribution a file named TOUR which
introduced the various data structures of the shell.

Do you think it would be a good idea to write a TOUR file fot tcsh ?


Regards,
Karim Hamidou



From christos at zoulas.com  Tue Mar 18 19:32:49 2008
From: christos at zoulas.com (Christos Zoulas)
Date: Tue, 18 Mar 2008 13:32:49 -0400
Subject: A TOUR file for tcsh
In-Reply-To: <45459.194.254.136.2.1205858042.squirrel@webmail.reseau.eseo.fr>
	from karim.hamidou@reseau.eseo.fr (Mar 18,  5:34pm)
Message-ID: <20080318173249.5621256539@rebar.astron.com>

On Mar 18,  5:34pm, karim.hamidou at reseau.eseo.fr (karim.hamidou at reseau.eseo.fr) wrote:
-- Subject: A TOUR file for tcsh

| Hi,
| 
| I remember that there was in the ash distribution a file named TOUR which
| introduced the various data structures of the shell.
| 
| Do you think it would be a good idea to write a TOUR file fot tcsh ?
| 

Are there many people who want to understand and change the tcsh source code?

christos


From ldarby at tuffmail.com  Tue Mar 18 20:29:10 2008
From: ldarby at tuffmail.com (Laurence Darby)
Date: Tue, 18 Mar 2008 18:29:10 +0000
Subject: A TOUR file for tcsh
In-Reply-To: <20080318173249.5621256539@rebar.astron.com>
References: <45459.194.254.136.2.1205858042.squirrel@webmail.reseau.eseo.fr>
	<20080318173249.5621256539@rebar.astron.com>
Message-ID: <20080318182910.0624afbd.ldarby@tuffmail.com>

Christos Zoulas wrote:

> On Mar 18,  5:34pm, karim.hamidou at reseau.eseo.fr
> (karim.hamidou at reseau.eseo.fr) wrote:
> -- Subject: A TOUR file for tcsh
> 
> | Hi,
> | 
> | I remember that there was in the ash distribution a file named TOUR
> | which introduced the various data structures of the shell.
> | 
> | Do you think it would be a good idea to write a TOUR file fot tcsh ?
> | 
> 
> Are there many people who want to understand and change the tcsh
> source code?

A TOUR file certainly would have helped me when I started on this:

http://bugs.gw.com/view.php?id=48

but I seemed to manage without one... Btw, is there any chance that can
be merged in?

Thanks,
Laurence


From karim.hamidou at reseau.eseo.fr  Wed Mar 19 14:04:06 2008
From: karim.hamidou at reseau.eseo.fr (karim.hamidou at reseau.eseo.fr)
Date: Wed, 19 Mar 2008 13:04:06 +0100 (CET)
Subject: A TOUR file for tcsh
In-Reply-To: <20080318173249.5621256539@rebar.astron.com>
References: <45459.194.254.136.2.1205858042.squirrel@webmail.reseau.eseo.fr>
	from karim.hamidou@reseau.eseo.fr (Mar 18,  5:34pm)
	<20080318173249.5621256539@rebar.astron.com>
Message-ID: <51982.194.254.136.2.1205928246.squirrel@webmail.reseau.eseo.fr>

> Are there many people who want to understand and change the tcsh source
> code?
>
> christos
>

Hi,

I was thinking that it may help hackers who would like to add a ponctual
functionnality to tcsh.