J
James Kuyper
Joachim said:
Mine (Thunderbird) does.
Mine (Thunderbird) does.
T
Tim Rentsch
Ben Bacarisse said:
This point is addressed in 6.8p2:
A block allows a set of declarations and statements to be grouped
into one syntactic unit. The initializers of objects that have
automatic storage duration, and the variable length array
declarators of ordinary identifiers with block scope, are
evaluated and the values are stored in the objects (including
storing an indeterminate value in objects without an initializer)
each time the declaration is reached in the order of execution,
as if it were a statement, /and within each declaration in the
order that declarators appear./ [emphasis added]
The last clause specifies evaluation order for declarators within
a single declaration for automatic variables, including (even
if somewhat obliquely) declarators in for() statements.
Presumably file scope declarations and 'static' declarations
within functions don't need such guarantees because the
requirements for their initializers (being constant expressions)
prevent any side effects from taking place.
S
Stephen Sprunk
Joachim said:
You mean the embedded declarations? That is indeed a C99ism (though I
learned it as a GCCism).
The second comma is clearly a comma operator; there is a sequence point
between "x++" and "y--".
The reason this is a common use for the comma operator is that, unlike
in most other contexts, a semicolon cannot be used here for that purpose.
No, but IIRC there is a sequence point before and after each of the
three expressions in a for() anyways; that a semicolon is used to
separate them just makes it a bit more confusing as to which rule applies.
S
H
Harald van Dijk
So it is. Thank you, I had missed that.
Right. That explains why this order of evaluation can be specified under
statements rather than declarations: it's specified as a property of the
statements that the declarations appear in, rather than of the
declarations themselves, because all declarations it applies to -- those
with automatic storage duration -- are part of a statement.
K
Keith Thompson
James Kuyper said:
Really? So if I write, say, a directory path /etc/ it puts "etc" in
italics rather than showing you the '/' characters? If my newsreader
did that, I'd look for a way to turn it off or a new newsreader.
R
Richard
Keith Thompson said:
Gnus does indeed do this. Well, in non antiquated versions. You should
consider moving up to Emacs 22 at least.
S
Stephen Sprunk
Keith said:
It doesn't remove the / characters; they and the text they enclose are
simply italicized, which is easy enough to ignore when it's incorrect.
Also, the italics (or *bold* or _underline_) only apply when the special
characters appear at the start and end of words; most folks write /etc,
not /etc/, and the former will not get italics.
The feature can be disabled easily enough if you don't like it, just
like one can disable conversion of emoticons (which does, on occasion,
erase valid punctuation). It's a popular feature, though.
S
P
Phil Carmody
MisterE said:Its interesting that the comma operator is so seldom used, even in its only
usual place in for(;.
I once saw some code that , joined every possible line it could. Not sure
why.
That was a response to a company coding standard which said "A
function may only have one return statement" and which was
interpreted ultra-literally. I hope you enjoyed maintaining
that code as much as I enjoyed writing it!
Phil
K
Keith Thompson
Phil Carmody said:
What does using comma operators rather than semicolons have to do with
having only one return statement? (Unless you mean that it was used
to ensure that the function had only one *statement* which isn't an
ultra-literal interpretation at all, just a wrong one.)
P
Phil Carmody
Keith Thompson said:
Au contraire. For example, if, as an external contractor in a
free staff canteen, you "may only have one plateful of main
course", then you appear to think that you can have an unbounded
quantity of dessert too to accompany said single plate of main
course. Such is not the common interpretation of the English
phrase "may only have one <X>".
You appear to be confusing it with "may have only one <X>",
where the "only" modifies a different semantic element, and
consequently has a different meaning.
Phil
J
James Kuyper
Keith said:
No, it displays it in italics while also showing the '/' characters.
I've never noticed this until Joachim's question made me actually check.
M
Mark Wooding
Joachim Schmitz said:
Gnus for one, if you run it under a window system. (I run my Gnus in
screen, so I get visible slashes.)
-- [mdw]
J
J. J. Farrell
Stephen said:
It's a very long-standing concept ... these three have been standard
Usenet markup since the dawn of time (or soon after). I don't know when
newsreaders started displaying the formats, but for terminals capable of
doing so I assume it was about the same time as the markup came in.