Monday, November 21, 2016

Digest for comp.lang.c++@googlegroups.com - 22 updates in 7 topics

ruben safir <ruben@mrbrklyn.com>: Nov 21 03:54PM -0500

On 11/21/2016 08:17 AM, Chris Vine wrote:
> unnecessarily confused by the basics - in this case between a type and
> a pointer to the type. You may also be confused by the syntactic
> difference between a pointer and a reference.
 
 
I am relearning C++ after a 6 year lay off. And BTW, the compiler is
probably the best instructor for C and C++
 
 
ruben safir <ruben@mrbrklyn.com>: Nov 21 03:56PM -0500

On 11/21/2016 08:17 AM, Chris Vine wrote:
> ng to the parent class in his post "the new {} syntax", your
> problem is that you are trying to initialize a parent* object 'tmp2'
> with something other than an address of a parent object.
 
 
right. I thought that is the syntax, but I misunderstood. Thank you,
 
Ruben
"Öö Tiib" <ootiib@hot.ee>: Nov 21 02:07PM -0800

On Monday, 21 November 2016 22:54:49 UTC+2, ruben safir wrote:
> > difference between a pointer and a reference.
 
> I am relearning C++ after a 6 year lay off. And BTW, the compiler is
> probably the best instructor for C and C++
 
Compiler is definitely worst instructor of those two languages whose
*every* feature contains a number of related undefined behaviors that
tools are not obliged to diagnose.
 
You can not find out about undefined behavior using compiler.
Undefined behavior sometimes results with a program that sometimes
appears to work or even that most of the time appears to work. So
the result of your voodoo programming is often somewhat unstable
program that acts in somewhat insane manner and your own poisoned
brains with misinformation about what voodoo spells work in C++.
Louis Krupp <lkrupp@nospam.pssw.com.invalid>: Nov 21 03:18PM -0700

On Mon, 21 Nov 2016 15:54:40 -0500, ruben safir <ruben@mrbrklyn.com>
wrote:
 
<snip>
>I am relearning C++ after a 6 year lay off. And BTW, the compiler is
>probably the best instructor for C and C++
 
The compiler might have better social skills than some of us on this
newsgroup, but I wouldn't call it a good instructor. It's kind of
terse.
 
I have a copy of Stephen Prata's C++ Primer Plus, Sixth Edition, 2012,
and it covers at least some of what's changed in C++11. Amazon lists
other titles. Read the reviews, buy whatever looks good from Amazon or
locally, and happy reading.
 
Louis
JiiPee <no@notvalid.com>: Nov 21 10:29PM

On 21/11/2016 22:18, Louis Krupp wrote:
> The compiler might have better social skills than some of us on this
> newsgroup
 
 
Yes, at least it is always 100% truthfull :). It does not tell lies...
JiiPee <no@notvalid.com>: Nov 21 10:29PM

On 21/11/2016 22:18, Louis Krupp wrote:
> I have a copy of Stephen Prata's C++ Primer Plus,
 
 
that was my first C++ book back in 1996 I bought, to learn C++. It was
good then... I liked it. I learned well using it. i tried also others
but this was best for me.
"Christopher J. Pisz" <cpisz@austin.rr.com>: Nov 21 05:12PM -0600

On 11/21/2016 4:29 PM, JiiPee wrote:
>> The compiler might have better social skills than some of us on this
>> newsgroup
 
> Yes, at least it is always 100% truthfull :). It does not tell lies...
 
I learned C++ from this newsgroup, Victor, Alf, and Oo Tiib.
Christian Steins <cs01@quantentunnel.de>: Nov 21 06:08PM +0100

Am 20.11.2016 um 18:01 schrieb Öö Tiib:
 
> Linux, iOS and Android. Also there are more people who use it and
> so more job offers where Qt framework is used than where wxWidgets is
> used.
 
Sorry for the german OP, clicked on wrong group.
Anyway thanks for your answer. So, QT seems to be preferred over wx by
most c++ people.
 
Christian
"Öö Tiib" <ootiib@hot.ee>: Nov 21 11:15AM -0800

On Monday, 21 November 2016 19:09:01 UTC+2, Christian Steins wrote:
 
> Sorry for the german OP, clicked on wrong group.
> Anyway thanks for your answer. So, QT seems to be preferred over wx by
> most c++ people.
 
It is unclear what most C++ people prefer. Lot of C++ people write embedded
software that is in devices without graphics output or server-side services
that have some network interface or command-line tools with text I/O or
games that use graphics rendering libraries built upon OpenGL or Direct3D.
 
I just say that Qt seems more useful for me and more popular than wxWidgets
among C++ programmers whom I know.
 
Also for example on programmers Q/A site like Stack Overflow it seems to
be more popular. See:
http://stackoverflow.com/questions/tagged/wxwidgets gives 3K questions
http://stackoverflow.com/questions/tagged/qt gives 54K questions
However for example unity3d framework gives also lot of questions ...
http://stackoverflow.com/questions/tagged/unity3d gives 21K questions
... so who knows what most prefer.
David Brown <david.brown@hesbynett.no>: Nov 21 08:36PM +0100

On 21/11/16 20:15, Öö Tiib wrote:
> However for example unity3d framework gives also lot of questions ...
> http://stackoverflow.com/questions/tagged/unity3d gives 21K questions
> ... so who knows what most prefer.
 
It is always difficult to know what these sorts of statistics mean, or
if they mean anything at all. It could be that QT is 18 times as
popular as wx - but it could also mean that it is 18 times as hard to
use, or that the wx documentation is 18 times as clear as QT's, or that
the wx project forums are 18 times as good as the QT project forums so
that people don't have to ask on Stack Overflow.
 
I think both are popular toolkits, with their pros and cons - I
certainly would not choose between them based solely on this
information. I also would not dismiss other possible toolkits either.
My advice to the OP is to read a bit on the project webpages, especially
the tutorials. Check that the licensing suits your project, and try out
a "hello world" with each toolkit that looks interesting. Then ask
questions in newsgroups that are a bit more specific once you know what
to ask about.
Melzzzzz <mel@zzzzz.com>: Nov 21 08:50PM +0100

On Mon, 21 Nov 2016 20:36:23 +0100
> use, or that the wx documentation is 18 times as clear as QT's, or
> that the wx project forums are 18 times as good as the QT project
> forums so that people don't have to ask on Stack Overflow.
 
Qt has excellent documentation and examples. Heck has IDE as well.
Really easy peazy to quickly get results.
 
 
 
 
--
press any key to continue or any other to quit
"Öö Tiib" <ootiib@hot.ee>: Nov 21 01:44PM -0800

On Monday, 21 November 2016 21:36:36 UTC+2, David Brown wrote:
> > ... so who knows what most prefer.
 
> It is always difficult to know what these sorts of statistics mean, or
> if they mean anything at all. It could be that QT is 18 times as
 
Good questions are rare on SO, so these are novices who ask there. Count
of ignorant people in trouble has always been quite good indicator of
desire to use and therefore popularity.
 
> popular as wx - but it could also mean that it is 18 times as hard to
> use, or that the wx documentation is 18 times as clear as QT's, or that
 
C++ is not really made easy to use for novice. Novice is likely
struggling and most often with C++ itself, its syntax, tools and simple
logic errors. The framework or library just happen to be the frame of
reference for that story of "torment" and "agony" of gaining skills
with C++. ;D
 
> the wx project forums are 18 times as good as the QT project forums so
> that people don't have to ask on Stack Overflow.
 
Both have really good documentation and decent forums IMHO so there
can't be such differences.
 
> a "hello world" with each toolkit that looks interesting. Then ask
> questions in newsgroups that are a bit more specific once you know what
> to ask about.
 
Yes. One's own eye is the king.
http://doc.qt.io/qt-5/qtexamples.html
http://docs.wxwidgets.org/3.1/page_samples.html
bitrex <bitrex@de.lete.earthlink.net>: Nov 21 03:37PM -0500

How can I do something like this with a copy constructor where I'd like
a second default argument, where the default is a member variable of the
class the copy is being made from
 
MyClass(const MyClass& my_class, member=my_class._member) :
_member(member) { etc...
 
Currently I'm getting a "local variable 'my_class' may not appear in
this context" error as it isn't in the appropriate scope for a default
argument.
Marcel Mueller <news.5.maazl@spamgourmet.org>: Nov 21 10:06PM +0100

On 21.11.16 21.37, bitrex wrote:
 
> Currently I'm getting a "local variable 'my_class' may not appear in
> this context" error as it isn't in the appropriate scope for a default
> argument.
 
You need to overload the constructor.
 
MyClass(const MyClass& my_class) : _member(my_class._member) ...
MyClass(const MyClass& my_class, member) : _member(member) ...
 
 
Default members are no overloads. They are just default arguments that
are passed by the caller when no argument is given.
This might look similar to overloading but it is entirely different. If
you change a default parameter you need to recompile every caller. If
you change the overloaded implementation (without the additional
parameter) to assign another default value you need to recompile the
callee only.
 
 
Marcel
ruben safir <ruben@mrbrklyn.com>: Nov 21 03:58PM -0500

CXX:=g++
CXXFLAGS:=-Wall -ggdb -pg -pthread
 
LDFLAGS:=-L/usr/local/lib/mysql -lmysqlpp -lmysqlclient
 
maxpath : maxpath.o main.o
${CXX} ${CXXFLAGS} -o maxpath maxpath.o main.o
 
main.o : main.cpp
${CXX} ${CXXFLAGS} -o main.o -c main.cpp
 
maxpath.o : maxpath.cpp maxpath.h
${CXX} ${CXXFLAGS} -c maxpath.cpp
 
clean :
rm maxpath *.o make.deps
touch *.cpp *.h
 
include make.deps
make.deps: *.cpp ; gcc -M *.cpp >$@
ruben safir <ruben@mrbrklyn.com>: Nov 21 03:50PM -0500

On 11/21/2016 08:59 AM, Alf P. Steinbach wrote:
 
> You were introducing needless and misleading uncertainty.
 
Alf, your no gem either. Reconsider what your saying because your way
off and if you were one of my children at the table, would be sent
upstairs to mull over things.
 
Reuvain
"Ahmet Can ÜN" <ahmetcanun@gmail.com>: Nov 21 10:54AM -0800

Hi,
 
This function rests in fbstring.h header from facebook folly library.
 
Link: https://github.com/facebook/folly/blob/master/folly/FBString.h#L434
 
I didn't see why this code casts p back and forth to void*?
static RefCounted * fromData(Char * p) {
return static_cast<RefCounted*>(
static_cast<void*>(
static_cast<unsigned char*>(static_cast<void*>(p))
- sizeof(refCount_)));
}
 
Why can't it be written like this? Can anyone see a logical reason?
 
static RefCounted * fromData(Char * p) {
return static_cast<RefCounted*>(
static_cast<unsigned char*>(p)
- sizeof(refCount_));
}
Gareth Owen <gwowen@gmail.com>: Nov 21 07:19PM

> static_cast<unsigned char*>(p)
> - sizeof(refCount_));
> }
 
What did your compiler say?
 
Clue: Any pointer type can be static_cast to a void*,
and a void* can be static_cast to any pointer type, but ...
"Ahmet Can ÜN" <ahmetcanun@gmail.com>: Nov 21 11:51AM -0800

On Monday, November 21, 2016 at 10:20:01 PM UTC+3, gwowen wrote:
 
> What did your compiler say?
 
> Clue: Any pointer type can be static_cast to a void*,
> and a void* can be static_cast to any pointer type, but ...
 
Thank you. I'm still under the influence of C.
"Öö Tiib" <ootiib@hot.ee>: Nov 21 11:54AM -0800

On Monday, 21 November 2016 20:54:36 UTC+2, Ahmet Can ÜN wrote:
> static_cast<unsigned char*>(p)
> - sizeof(refCount_));
> }
 
Because static_cast is not meant for type punning between unrelated
types like that.
 
It is legal to use reinterpret_cast like that:
 
return reinterpret_cast<RefCounted*>(
reinterpret_cast<unsigned char*>(p)
- sizeof(refCount_));
 
It is legal to use C style cast like that:
 
return (RefCounted*)(
(unsigned char*)(p)
- sizeof(refCount_));
Chris Vine <chris@cvine--nospam--.freeserve.co.uk>: Nov 21 05:31PM

On Mon, 21 Nov 2016 14:48:15 +0000
> ODR-using it. Casting it to lvalue of another type is another. I
> don't at the moment have the C++ standard in front of me to give you
> the reference.
 
I do now. It's §9.4.2/3. "If a non-volatile const static data member
is of integral or enumeration type, its declaration in the class
definition can specify a brace-or-equal-initializer in which every
initializer-clause that is an assignment-expression is a constant
expression (5.20). ... . The member shall still be defined in
a namespace scope if it is odr-used (3.2) in the program and the
namespace scope definition shall not contain an initializer." This
implies that the initialised const static integral member need not be
defined in namespace scope (its declaration is sufficient) if it is not
ODR-used.
 
Chris
Chris Vine <chris@cvine--nospam--.freeserve.co.uk>: Nov 21 06:18PM

On Mon, 21 Nov 2016 17:31:03 +0000
> Chris Vine <chris@cvine--nospam--.freeserve.co.uk> wrote:
> > On Mon, 21 Nov 2016 13:26:07 GMT
> > scott@slp53.sl.home (Scott Lurndal) wrote:
[snip]
> implies that the initialised const static integral member need not be
> defined in namespace scope (its declaration is sufficient) if it is
> not ODR-used.
 
However I do now recall having experiences similar to the one you
mention. Some forms of ODR-use, such as assigning to a const
reference, may be optimized out with a sufficiently high level of
optimization, and will compile and link even though technically the
code in question is not conformant. That may be what you had in mind.
But where the static member is not ODR-used, you should be safe from
that problem.
You received this digest because you're subscribed to updates for this group. You can change your settings on the group membership page.
To unsubscribe from this group and stop receiving emails from it send an email to comp.lang.c+++unsubscribe@googlegroups.com.
Posted by at

No comments:

Post a Comment

Subscribe to: Post Comments (Atom)