🚀 go-pugleaf

In article <45fjo5$2gs@usenet.pa.dec.com> diamond@jrd.dec.com (Norman Diamond) writes:

> From: diamond@jrd.dec.com (Norman Diamond)
>
> Now, I look at the wording of translation phase 2.  It does not specify
> that each backslash followed by a new-line character is deleted; it
> specifies that each new-line character and an immediately preceding
> backslash are deleted.  This wording suggests that a backwards examination
> is required, similar to detection of a preprocessing directive which
> depends on examining any white space that immediately precedes a #.
> Thus, it seems that the following four lines are spliced by TP2:
> tok\\\
> 
> 
> en
> After each splice is performed, a newline character is observed to be
> immediately preceded by a backslash.

I read that too in the Standard and I am inclined to agree.  Here is a
short list of what some other (assuming you may be one) compiler writers
understood:

HP PA-RISC - gcc : leaves two \\ after tok (version 2.6.3)
HP PA-RISC - c89 : leaves two \\ after tok
SUN        - cc  : removes all three \\\ and yields only one `token'

--Daniel