That change was done.

The problem is not that the value 2 is missing.

It is there, it simply fell in my copy-and-paste.

The problem is that the meaning of 'entered' and 'exited'

in this context is completely unclear.

I'm going back to the 'cc' man page to find out the intended meaning.

Shalom

 

Adam Krolnik wrote:
<blockquote TYPE=CITE>Funny, there was a change request made on this line.
It looks as

 though the line was corrupted in the process somehow.

 From the archives:

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

 Date: Wed, 15 Dec 1999 10:34:44 -0600

 From: Adam Krolnik <krolnik@lsil.com>

 Subject: BTF - Draft 4 requested change.

 

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

 Was:

 Section 19.7 `line, last sentence of paragraph 2.

 "The level parameter indicates whether an include file has

 been specified (value is 1), an include file is exited (value
is 2),

 or neither has been done (value is 0).

 Proposed:

 "The level parameter indicates whether an include file has

 been entered (value is 1), an include file is exited (value is
2),

 or neither has been done (value is 0).

 

 Proposed change:

 Switch 'specified' to 'entered' - removes ambiguity.

-- 
 **************************************************************************
 Shalom Bresticker                           Shalom.Bresticker@motorola.com
 Motorola Semiconductor Israel, Ltd.                  Tel #: +972 9 9522268
 P.O.B. 2208, Herzlia 46120, ISRAEL                   Fax #: +972 9 9522890
 **************************************************************************

Subject:        Re: B29
Line and File compiler Directives.

   From:        Adam Krolnik
<adamk@cyrix.com>

     To:       
eli@avanticorp.com
<blockquote TYPE=CITE>You mention that the `line directive having 2 additional
optional parameters.

I only have access to a sun/solaris CPP man page. These are the relevant

sections:

     #line integer-constant "filename"

          Generate line
control information for the next pass  of

          the C compiler. 
integer-constant is interpreted as the

          line number
of the next line  and  filename  is  inter-

          preted 
as the file from where it comes.  If "filename"

          is not given,
the current filename  is  unchanged.   No

          additional 
tokens  are permitted on the directive line

          after the optional
filename.

 Output

     Output consists of a copy of the input file,
with  modifica-

     tions, plus lines of the form:

          #lineno " filename"
"level"

     indicating the original source line number
and  filename  of

     the following output line and whether this
is the first such

     line after an include file has been entered 
(level=1),  the

     first  such  line  after 
an  include  file  has been exited

     (level=2), or any other such line (level is
empty).

Are you asking for allowing the 'level' field from non specifyable

directive?  It would have to be optional since we are effectively

merging two directives into one for Verilog.

It looks to me as though there was some confusion about the `line
directive.

I looked at both CPP  and Gnu CPP.

- The worst is the "level" parameter, which is a CPP output
! It is not a user directive at all ! It is INFORMATION

for the compiler which DESCRIBES the file hierarchy, not CONTROLLING it.
Even worse, it became a REQUIRED parameter in 1364-2001. No wonder
I did not understand its meaning and use in 1364!

- Why is "filename" required ? Neither CPP nor Gnu CPP require
it !  I might want to use the same file

in different places under different names. I should not have to
play with this directive each time.

-  The Gnu CPP man page says that "filename" is a string
constant. Neither CPP nor Gnu CPP say that "filename" can
be a full or relative path name. Why here ? (I'm devil's advocate
here.) What happens if "filename"

is something like ".." ? Is the Verilog compiler/interpreter supposed
to relate to this as just a filename or is it

supposed to interpret file location as well ?

- By the way, the Gnu CPP man page specifies that the number parameter
"is a non-negative decimal integer constant". (Interesting. "Non-negative"
means that 0 is permitted.)

- Also by the way, the CPP man page description of "level" is
much clearer than what entered 1364.

In my opinion, "level" is a mistake and should be eliminated, and
"filename" should be "optional".

Comments ?

Shalom

 

Shalom.Bresticker@motorola.com wrote:
<blockquote TYPE=CITE>In 19.7 (`line): "The level parameter indicates whether
an

include file has been entered (value is 1), an include file

is existed (value is), or neither has been done(value is 0)."

This sentence is incomprehensible to me.

At the very least, an example or examples must be added.

-- 
 **************************************************************************
 Shalom Bresticker                           Shalom.Bresticker@motorola.com
 Motorola Semiconductor Israel, Ltd.                  Tel #: +972 9 9522268
 P.O.B. 2208, Herzlia 46120, ISRAEL                   Fax #: +972 9 9522890
 **************************************************************************

> - Why is "filename" required ? Neither CPP nor Gnu CPP require it
!  I >might want to use the same file

> in different places under different names. I should not have to play
with >this directive each time.

This came up in the discussions. To simplify parsing, all parameters

became required.
I don't understand the parsing problem.

It seems to have easy solutions.

The 'linenum' parameter is required.

"filename" is always surrounded by parentheses.

'level' is not surrounded by parentheses.

Further, you could make "filename" required if you want to give
'level'.

So parsing should be easy.

 
<blockquote TYPE=CITE>Since this directive was envisioned to be used by
a

source code generator (preprocessor, source to source translator, etc.)

the requirement of the parameters was not considered a hardcase. You
are

correct, it is not controlling anything; nor would you want it to.
Again, it looks to me to be a confusion between a compiler DIRECTIVE,

which TELLS the compiler to DO something, on the one hand,

and an informative preprocessor output on the other.

If, for CPP, 'linenum' and "filename" were enough, why does Verilog
need 'level' as well ?

What does it add ? Why can't it just be a comment ?

And a preprocessor may not know, for example, whether a file is included
or not.

 
<blockquote TYPE=CITE>> -  The Gnu CPP man page says that "filename"
is a string constant. >Neither CPP nor Gnu CPP say that "filename" can
be a full or

> relative path name. Why here ? (I'm devil's advocate here.) What
happens

> if "filename"

> is something like ".." ? Is the Verilog compiler/interpreter supposed
to

>relate to this as just a filename or is it supposed to interpret file

>location as well ?

The filename is a string - whatever you want to put there. The

compiler's

only use for this data is to inform the user of something pertaining

to it. So your filename could be any of your examples. What the user

puts in is what he will get out.
But if the LRM says it can be a full or relative path name, that seems
to imply that it has real meaning.

 

-- 
 **************************************************************************
 Shalom Bresticker                           Shalom.Bresticker@motorola.com
 Motorola Semiconductor Israel, Ltd.                  Tel #: +972 9 9522268
 P.O.B. 2208, Herzlia 46120, ISRAEL                   Fax #: +972 9 9522890
 **************************************************************************