.
Last update: 1997-05-20
9945-2-62
_____________________________________________________________________________
Topic: ex/vi
Relevant Sections: 5.10.7.2.1, page 523, lines 1328 etc
Defect Report:
-----------------------
(From:bostic@vangogh.CS.Berkeley.EDU (Keith Bostic))
For each of the areas discussed below, I believe that
the intent of the POSIX standard is correct, but that
the standard itself is unclear or ambiguous.
(1) Command Descriptions
Section 5.10.7.2.1, page 523, lines 1328
Historically, the beginning/end of the insert was also considered
to apply in the test, not just whether or not it was preceded or
followed by character that can't be part of a word. For example,
if you have the string "ABCDEFGH", with the cursor on the 'C', and
you type ":ab foo bar<carriage-return>ifoo<esc>", the abbreviation
will be performed.
Suggestion:
Reword to make this clear.
(2) Command Descriptions
Section 5.10.7.2.37, page 532, lines 1683-1688
Historically, the ! did not have to immediately precede the
file, i.e. "w ! foo" was legal.
Suggestion:
Reword to make this clear.
(3) Command Descriptions
Section 5.10.7.2.44, page 534, line 1753
The current breakup of the substitute commands is somewhat confusing.
In particular, the substitute command specified on line 1755 is a
special case of the main one specified on page 530, line 1597.
Suggestion:
Merge the two sections into one section.
(4) Extended Description
Section 5.35.7, page 628, lines 4863-4865
The wording in this paragraph requires that if a % or # appears as part
of a pathname entered as a command argument that it's expanded to the
current or alternate pathname. It's unclear to me that this wording
doesn't require that a '%' in a substitute pattern be expanded. Is
"pathname" a defined term for command arguments? The synopsis lines
for things like the ex edit command use "file", not pathname.
Suggestion:
Reword as necessary; the intent is correct.
WG15 response for 9945-2:1993
-----------------------------------
Q1:
The standard states what sequence of characters make up the word, and
conforming implementations must conform to this. However, concerns have
been raised about this which are being referred to the sponsor.
Q2:
The standard clearly states, in the synopsis, a form which corresponds to your
example, and conforming implementations must conform to this.
Q3:
The standard clearly states a definition for the syntax of substitute and
resubstitute, and conforming implementations must conform to this. Concerns
about this issue are being referred to the sponsor.
Q4:
The standard clearly states the behavior for the occurance of a '%' or '#'
in a pathname, and conforming implementations must conform to this.
Rationale for Interpretation:
-----------------------------
None.
_____________________________________________________________________________