Request for comments on forthcoming blog posting about Unicode inWindows consoles

[Alf P. Steinbach]

http://alfps.wordpress.com/?p=356&shareadraft=4ece5380818f3

You may have a chance to affect the final posting...

The link is just to an unfinished draft (maybe I'll complete this this
evening), but much read-worthy there, I think.


Cheers,

- Alf

 

[Alf P. Steinbach]

Off-topic.

I wouldn't have posted here if I thought it was off topic here.

Despite the lack of a charter for this newsgroup.

It's about using the C++ language and about limitations in the C++
standard library.

You may of course think that it's environment specific, like,
"Windows!". And in one very limited sense it is. But on the other hand,
it is necessary to discuss these details in order to be able to write
portable C++ code.

As a minimum example, the copy program in Josutti's book, while intended
to be portable and by many people believed to be portable, is not. It is
actually environment specific, while appearing to be portable. While the
code I'm discussing is opposite: it may appear to be
environment-specific, but is actually in support of portability ;-).

Cheers & hth., & thanks for the comment!,

- Alf

 

[Alf P. Steinbach]

Your code *is* environment specific; it doesn't just *appear* to be
environment specific hence it is off-topic here; try posting it in a
Micro$oft newsgroup instead.

You didn't get the point.

But thank you for the advice about posting to Microsoft groups.

Although I am sure that you were suggesting that for entirely wrong reasons.

Posting to a Microsoft group could in principle be useful for getting
some action going so that the same C++ code could be used portably in
Windows and Linux (at least). However, I know from experience that such
postings are met with hate and disdain from the fanboys that frequent
the MS groups. Similarly, Microsoft's very practical suggestion to have
source code files start with a BOM, is ridiculed in the *nix Unicode FAQ
and other *nix documents. So there's much hate and idiocy on both sides
of that historical conflict. I'd rather avoid that.

Folks who read this group can instead learn a lot about C++, and help to
influence the future, by reading good blogs such as mine. ;-)

http://alfps.wordpress.com/?p=356&shareadraft=4ece5380818f3


Cheers & hth.,

- Alf

 

[Liviu]

http://alfps.wordpress.com/?p=356&shareadraft=4ece5380818f3

You may have a chance to affect the final posting...

Off topic here, as noted already. I'll just add that the title you chose
is also off topic relative to the actual contents of your blog post ;-)

You seem to be writing about wide standard/stream I/O in console
apps compiled with VC++ v10 run under Windows' builtin command
line interpreter cmd.exe. Neither the "stream" notion, nor the compiler
and not even the CLI itself are part of the "Win32 console subsystem".
proper. You'd get different behavior compiling with ICC and running
under 4NT (now TCC) for example, though in both cases you'd be
effectively using the same underlying "Windows console subsystem".

Bye & hth,
Liviu

 

[jacob navia]

Le 24/11/11 17:10, Alf P. Steinbach a écrit :

So there's much hate and idiocy on both sides
of that historical conflict. I'd rather avoid that.

You can't.

You can see that the answers here have all this underlying

"MY OS IS THE BEST"

type of idiocy. GNU fans consider Microsoft the evil empire
and Windows guys consider the "freetards" as mentally retarded.

Anything trying to be PORTABLE from one to the other
would imply trying to work on BOTH, an heresy for the fabs
of either system...

Note that "fan" comes from the word FANATIC.

jacob

 

[Alf P. Steinbach]

Off topic here, as noted already. I'll just add that the title you chose
is also off topic relative to the actual contents of your blog post ;-)

That's rubbish. You're an idiot playing with social signals, that's all.

You seem to be writing about wide standard/stream I/O in console
apps compiled with VC++ v10 run under Windows' builtin command
line interpreter cmd.exe. Neither the "stream" notion, nor the compiler
and not even the CLI itself are part of the "Win32 console subsystem".
proper. You'd get different behavior compiling with ICC and running
under 4NT (now TCC) for example,

Maybe, maybe not. Since *all* your technical assertions have been wrong
so far, I take this new assertion with a grain of salt (to put it
mildly). However, as noted in the first blog post in that series, g++
has only C level support for Microsoft's _O_U8TEXT mode.

I think other readers understand that it's not possible to have
discussed every issue in detail in the middle of a series of postings
addressing the problem area. It would be a variation of Zeno's paradox.

But I think other readers understand that when I mentioned e.g. g++'s
limited support, it was because I knew about it, because I have done
everything that I write about, not because I (like you, so far) guessed
that that might be the case and treated the plausible guess as a fact.

though in both cases you'd be
effectively using the same underlying "Windows console subsystem".

Yes, that's right, and I'm baffled reading something you have written
that makes sense.


Cheers & hth.,

- Alf

 

[Liviu]

That's rubbish. You're an idiot playing with social signals, that's
all.

Lighten up, Francis. Just because you are confused about Unicode
in Windows consoles, and seemingly bitter that you don't have a clue
as to how it all works, doesn't mean you can't write about it ;-)

However, get used to the idea that those who actually cared to learn
and understand won't stop writing actual Win32 Unicode console apps
(of which there are many real life examples) just because you personally
consider it too hard to comprehend.

That said, holidays call and I am out of here for good. You've got
the entire usenet and blogosphere playground all to yourself.

Liviu

 

[Johannes Schaub]

http://alfps.wordpress.com/?p=356&shareadraft=4ece5380818f3

You may have a chance to affect the final posting...

The link is just to an unfinished draft (maybe I'll complete this this
evening), but much read-worthy there, I think.

That's mean. Next time please post a preview about a C++ topic so that we
all can help you and get fame in your "Thanks to" list.

 

[Alf P. Steinbach]

That's mean. Next time please post a preview about a C++ topic so that we
all can help you and get fame in your "Thanks to" list.

I know that you like formal standards issues.

And in that blog posting there are a few quotes from C+11 and C99 (just
search for those terms).

Where you might conceivably disagree with my reading?

For the purely formal there's also the issue that Leigh Johnston
unwittingly brought up.

Namely, whether imbuing a locale with some custom codecvt facet, in
`std::wcout`, is guaranteed to have any effect?


Cheers,

- Alf

 

[gwowen]

Le 24/11/11 17:10, Alf P. Steinbach a écrit :


You can't.

You can see that the answers here have all this underlying

Also, the idea that comp.lang.c++ is restricted to the standardised
language and everything else is off-topic is simply not supported by
the groups charter, or its common usage in the early days of the
newsgroup. Topicality fundamentalists like Leigh are a relatively new
occurence.

Of course, that was in the days when being a polite, helpful, decent
human being was considered preferable to being a language lawyer /
usenet cop.

 

[88888 Dihedral]

Well, I tested C++ that objects might eat heap memory very quickly.

Even I write assembly code to link with C++, it is still slow down with objects referenced BY MANY OBJECTS.

I suspect that in modern cpus, any language that allocate too much in the heap will be slow down due to the cache swapping beyond programs use the memory in the heap beyond L1 and L2 in the long time.

Is there any better way to avoid this by writing all programs in C++?

 

[Tobias Müller]

Bullshit; the closest there is to a "groups charter" is the FAQ and it says:

"Only post to comp.lang.c++ if your question is about the C++ language
itself. For example, C++ code design, syntax, style, rules, bugs, etc.
Ultimately this means your question must be answerable by looking into
the C++ language definition as determined by the ISO/ANSI C++ Standard
document, and by planned extensions and adjustments. Operating-specific
questions (e.g., about Windows NT / 95 / 3.x, UNIX, etc.) should go to an
operating-system-specific newsgroup (see below), not to comp.lang.c++."

However, a newsgroup is not only meant for asking a question and getting
the correct answer, but also for discussions.

So Alf's post is off-topic due to the fact that it is a platform specific post.


If you want to ignore the concept of a "topic" then we might as well just
have a single newsgroup into which anyone could post any old garbage
about any subject imaginable and live with the resultant "noise".

But questions or discussions can cover more than one topic (topic, as
defined by the usenet). Personally, I find a discussion that is partially
off topic much less disturbing than those "You're an idiot" - "shut up and
learn programming first" - "You're an idiot" - "shut up" discussions that
happen regularly on this list.

That said, I don't like posts that are completely off topic either.

Tobi