| Add Proposal | Add Analysis | Edit Class, Environment, or Release |
| Number | 583
|
| Category | errata
|
| Synopsis | comment_text allows \n and */
|
| State | open
|
| Class | errata-discuss
|
| Arrival-Date | May 11 2004
|
| Originator | Shalom Bresticker <Shalom.Bresticker@freescale.com>
|
| Release | 2001b
|
| Description |
Geoffrey.Coram@analog.com wrote me the following: I'm working in the Verilog-AMS committee and I noticed a small glitch in the LRM definition for comments. Both the AMS LRM and 1364-2004-d2 contain the following: one_line_comment ::= // comment_text \n block_comment ::= /* comment_text */ comment_text ::= { Any_ASCII_character } However, a newline should be excluded from one_line_comment. It's not clear what the termination is. // this should not match the syntax but "this should not \n match the syntax" is comment_text because I can use it in between /* and */ Maybe I'm thinking too much like a "greedy" parser. A greedy parser is one that tries to fit as many tokens into the pattern as possible. Thus, if I had a pattern like "a*a" where * was to match any ascii character, then from the following input stream afooabarafoofoo the parser would match the pattern afooabara and leave "foofoo" on the stack to look for the next pattern. For the case of the comments, we want the \n (or the */) to terminate the pattern. A non-greedy parser would pick out "afooa" from the stream to match the pattern and pass that as a token, and then pick from from "barafoofoo" to find the next pattern. -Geoffrey Similarly, comment_text allows for */ to appear in it. Strictly speaking, there probably should be two different types of comment text. Shalom -- Shalom Bresticker Shalom.Bresticker @freescale.com Design & Reuse Methodology Tel: +972 9 9522268 Freescale Semiconductor Israel, Ltd. Fax: +972 9 9522890 POB 2208, Herzlia 46120, ISRAEL Cell: +972 50 5441478 [ ]Freescale Internal Use Only [ ]Freescale Confidential Proprietary |
| Fix |
Unknown |
| Audit-Trail |
From: Steven Sharp <sharp@cadence.com> To: etf-bugs@boyd.com, Shalom.Bresticker@freescale.com Cc: Subject: Re: errata/583: comment_text allows \n and */ Date: Tue, 11 May 2004 15:14:01 -0400 (EDT) Is anybody really going to have trouble understanding what is meant? I think we have enough real issues to deal with, without nitpicking. Steven Sharp sharp@cadence.com From: Michael McNamara <mac@verisity.com> To: Shalom Bresticker <Shalom.Bresticker@freescale.com> Cc: etf-bugs@boyd.com Subject: RE: errata/583: comment_text allows \n and */ Date: Tue, 11 May 2004 13:14:02 -0700 Good points. Something else to consider is an enhancement that allows non-ascii characters in comments. (Japanese, Chinese, French, German, Hebrew, et cetera) -- On May 11 2004 at 02:50, Shalom Bresticker sent a message: > To: etf-bugs@boyd.com > Subject: "errata/583: comment_text allows \n and */" > >Number: 583 > >Category: errata > >Originator: Shalom Bresticker <Shalom.Bresticker@freescale.com> > >Description: > > Geoffrey.Coram@analog.com wrote me the following: > > I'm working in the Verilog-AMS committee and I noticed a > small glitch in the LRM definition for comments. Both > the AMS LRM and 1364-2004-d2 contain the following: > > one_line_comment ::= // comment_text \n > block_comment ::= /* comment_text */ > comment_text ::= { Any_ASCII_character } > > However, a newline should be excluded from one_line_comment. > > It's not clear what the termination is. > > // this should not > match the syntax > > but "this should not \n match the syntax" is comment_text > because I can use it in between /* and */ > > Maybe I'm thinking too much like a "greedy" parser. > > A greedy parser is one that tries to fit as many tokens > into the pattern as possible. > > Thus, if I had a pattern like "a*a" where * was to match > any ascii character, then from the following input stream > afooabarafoofoo > the parser would match the pattern > afooabara > and leave "foofoo" on the stack to look for the next pattern. > > For the case of the comments, we want the \n (or the */) to > terminate the pattern. > > A non-greedy parser would pick out "afooa" from the stream > to match the pattern and pass that as a token, and then > pick from from "barafoofoo" to find the next pattern. > > -Geoffrey > > Similarly, comment_text allows for */ to appear in it. > > Strictly speaking, there probably should be two different > types of comment text. > > Shalom > > -- > Shalom Bresticker Shalom.Bresticker @freescale.com > Design & Reuse Methodology Tel: +972 9 9522268 > Freescale Semiconductor Israel, Ltd. Fax: +972 9 9522890 > POB 2208, Herzlia 46120, ISRAEL Cell: +972 50 5441478 > > [ ]Freescale Internal Use Only > [ ]Freescale Confidential Proprietary |
| Unformatted |
|
Hosted by Boyd Technology