| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 185.1 | passed along | GOSTE::CALLANDER | | Mon Jul 16 1990 09:52 | 3 |
|
Your request has been sent along to the team owning this file.
|
| 185.2 | | TOOK::SWIST | Jim Swist LKG2-2/T2 DTN 226-7102 | Mon Jul 16 1990 21:45 | 12 |
| We're going to clean some of this up when we port to Ultrix, but it's
probably too late to do other than minor things for MCC v1.0/vms.
With regard to the specific case cited, it should be noted that neither
the VAX nor the MIPS Ultrix C compilers complain about nested comments
(and the VAX/Ultrix pcc is generally regarded as a lowest common
denominator).
As a matter of general policy I'm not sure we should change something
if ONLY the VAX C STANDARD=PORTABLE qualifier complains about it,
particularly if ANSI C (where we're all supposed to be going) allows
it.
|
| 185.3 | | TOOK::R_LANE | | Tue Jul 17 1990 09:50 | 13 |
| I find the STANDARD=PORTABLE qualifier very useful for debugging coding
errors (including missing "*/", which results in "nested comment"
informational messages). However, it usefulness is diminished somewhat if
the compiler output is cluttered with nested comment messages caused by
comments that contain an extra "/*".
C certainly allows the type of "nested comments" that are in mcc_vea_def.h,
but I think this is a bad practice that should be avoided, especially in
common header files that are used by many programmers.
The next time someone adds something to mcc_vea_def.h I think that an extra
5 minutes spent on removing the "/*" from within the comments is worth the
effort.
|
| 185.4 | Check Kernighan and Ritchie... | SMAUG::BELANGER | Quality is not a mistake! | Tue Jul 17 1990 10:43 | 12 |
|
In regard to nested comments and ANSI C. According to Kernighan's
and Ritchie's ANSI C book (usually considered the final word in ANSI
C), on page 192 under the heading A2.2 Comments it states:
The characters /* introduce a comment, which terminates with
the characters */. Comments do not nest, and the do not occur
within string or character literals.
Therefore, nested comments are NOT accepted by ANSI C.
~Jon.
|
| 185.5 | which is nested? | TOOK::JESURAJ | | Tue Jul 17 1990 13:38 | 37 |
| ref .-1
Consider the examples:
example A:
/*
/*
/* */
example B:
/*
/*
*/
*/
Which one of these is nested?
Clearly Example A follows the rule stated in .4 , and hence should be
ok with the standard, and should not give any error message.
Example B does not follow the rule stated in .4. Therefore it should
generate error message.
The comments in MCC_VEA_DEF.H are similar to example A above. I wonder
why portable option gives error messages.
... jesuraj
|
| 185.6 | | TOOK::R_LANE | | Wed Jul 18 1990 13:15 | 37 |
| Re: .5
I think what is meant by the statement "comments do not nest" is that you
can't "comment out" a section of code that contains one or more comments.
It's clear that the comments in your example "B" are nested.
Why the comment in example "A" is considered to be a nested comment isn't
as clear because "/*" can be included in a comment intentionally by the
programmer without causing any problems during compilation. The
compilation is successful because once the first "/*" is encountered, all
following characters (including the second "/*") are ignored until a "*/"
is found.
The following is also equivalent to your example "A":
/* comment... programmer forgot to end comment
<code>
.
.
.
/* next comment... */
I consider this to be a nested comment also. This type of mistake can
sometimes cause subtle bugs that are hard to find. I would rather
have the compiler issue a message warning me about a possible nested
comment than have to find the problem myself during a debugging session.
Compiler messages generated due to comments that intentionally contain "/*"
may obscure messages that indicate erroneous "open" comments. This is why
I think it's a bad practice to use "/*" within comments, especially in
public header files. There are plenty of other character combinations that
can be used as separators or "comment block borders".
Roy
|