D
Dan Pop
In said:
Now, this makes sense, considering the number of newsreaders that don't
give their users control over the headers of their posts...
Dan
Now, this makes sense, considering the number of newsreaders that don't
give their users control over the headers of their posts...
Dan
D
Dan Pop
^^^^^^In said:
When, *exactly*?
Dan
R
Ralmin
Prashanth Badabagni said:
It simplifies to:
int main(void)
{
if(putchar(59), putchar(10), exit(0), 1)
{
}
}
The value 59 corresponds to the semicolon in ASCII. The value 10 corresponds
to the newline in ASCII.
So, on a system with a character set based on ASCII (which includes most
modern computers) it is equivalent to:
int main(void)
{
puts(";");
return 0;
}
It is not valid code on C99 because there is no declaration in scope for the
putchar (<stdio.h>) or exit (<stdlib.h>) functions.
A
Andrey Tarasevich
Dan said:
See the message from CBFalconer in this very branch.
D
Dan Pop
In said:
I can see no message from CBFalconer in this very branch.
Dan
J
Joona I Palaste
Andrey must be talking about this:
(written by CBFalconer)----------------------------------------------
Either version works, because it is unofficial and the servers
choose whether or not to observe it. Google does. Thus use of
the silly thing keeps the article off the whole google system.
D
Dan Pop
^^^^^^^^^^^^^^^^^^^^^^^^^^In said:
Are you sure? I wasn't aware of a single thread having multiple subject
lines...
Dan
J
Joona I Palaste
Dan Pop said:
It's a definition question. My definition of a thread is based on the
References: NNTP header. A message with no article IDs in the header
starts a thread, and any message with its article ID in the header
belongs to that thread.
D
Dan Pop
In said:
That's a bogus definition. A thread is defined by its subject line: this
is what you specify in your killfile when you want to plonk a thread.
Furthermore, many people use to start a new thread by "recycling" the
headers of an unrelated article. This is why certain new threads are
born with a non-empty References: header line. My newsreader displays
the reference count of each article in overview mode and I easily spot
them.
That's sheer naivety. Apart from the above, there are *many* cases of
articles belonging to a thread that no longer have the original articled
ID referenced. Some people's posts never contain more than three IDs
in the References: line (presumably the most recent three) and I am
often forced to trim my References: line because it exceeds whatever
limit my server enforces and the article is rejected.
The subject line is not affected by any such issues... Furthermore,
it provides a definition that is consistent on both Usenet and mailing
list discussions.
Dan
A
Arthur J. O'Dwyer
What exactly do you mean by "subject line"? Obviously the messages
with headers
Subject: foo
Subject: Re: foo
belong in the same "thread," according to your non-References-based
definition (which I don't agree with, even though Google does). How
about
Subject: RE: foo
Subject: Fwd: foo
Subject: [OT] Re: foo
Subject: Re: [OT] Re: foo
Subject: Re: [OT] foo
Subject: Re: fo...
Subject: bar, was Re: foo
Subject: [bar] Re: foo
I've ordered those subject lines roughly in the order I'd expect to
see a hypothetical Subject-threading-based newsreader accept them
as co-threaded with Subject: foo. IMO all of them would be accepted
as co-threaded by the average human reader.
And then Google accepts Subject headers like these:
Subject: foo
Subject: foo
Subject: FOO
Subject: Re: FOO
as co-threaded with Subject: foo, even though any human or
References-based reader would see immediately that capitalization is
significant in Subject headers, and no thread can have more than one
"root" message (indicated by the lack of Re: on the first three
example subject lines).
The major problem with Subject-based threading, of course, is that
you end up with "threads" like this one:
http://groups.google.com/[email protected]
This is *exactly* why the concept of the unique Message-ID header
was invented[1], and why References-based "tree" threading is
intrinsically better than Subject-based threading. YMOV, but I
really don't know why.
-Arthur
[1] - If it's not, well, it should be.
A
Alan Balmer
M
Mark McIntyre
and Subject: Re: Réf: Re: foo
if you're lucky enough to correspond with lots of french people.
I took one luck and curled up into a ball...
D
Dan Pop
Not in mine. The next to last is obviously another thread, caused by
topic shift. The connection between the last and all the others is
impossible to predict, a priori. The ones with [OT] may also be different
threads, if the original thread was topical.
Dan
D
Dave Thompson
According to your headers you mean OE, not full Outlook.
I don't think you can; except by cracking and patching OE, which
wouldn't be worth it. Or by using a (separate) proxy. But as already
noted for X-No-Archive you don't need to; the only system it's aimed
at, gooja, accepts it in body. (I am among those that think no-archive
is a bad idea anyway, but that's a different question.)
Huh? When? OE puts stupid X-headers, and likes to put inappropriate
and at best useless "rich" formats in both headers and body, but I've
never heard of a malware it even *could* put in headers much less did.
I don't know any headers even M$ can unwisely and unsafely give
execution semantics; only content-disposition even comes close.
Quite a few viruses use(d) the OE (or O?) *address book* to find
targets, and apparently spam zombies do so for camoflage; but AFAIK
not going through the actual software, and even if they do, the virus
if any is in the body not the headers.
I don't even recall any exploits of OE header *receipt* processing,
though there are plenty of body ones, especially HTML and MIME parts.
- David.Thompson1 at worldnet.att.net
I don't think you can; except by cracking and patching OE, which
wouldn't be worth it. Or by using a (separate) proxy. But as already
noted for X-No-Archive you don't need to; the only system it's aimed
at, gooja, accepts it in body. (I am among those that think no-archive
is a bad idea anyway, but that's a different question.)
Huh? When? OE puts stupid X-headers, and likes to put inappropriate
and at best useless "rich" formats in both headers and body, but I've
never heard of a malware it even *could* put in headers much less did.
I don't know any headers even M$ can unwisely and unsafely give
execution semantics; only content-disposition even comes close.
Quite a few viruses use(d) the OE (or O?) *address book* to find
targets, and apparently spam zombies do so for camoflage; but AFAIK
not going through the actual software, and even if they do, the virus
if any is in the body not the headers.
I don't even recall any exploits of OE header *receipt* processing,
though there are plenty of body ones, especially HTML and MIME parts.
- David.Thompson1 at worldnet.att.net