Monday, November 16, 2015

Digest for comp.lang.c++@googlegroups.com - 25 updates in 7 topics

Geoff <geoff@invalid.invalid>: Nov 15 04:13PM -0800

On Sat, 14 Nov 2015 10:01:14 -0800 (PST), Vlad from Moscow
 
>This code does not satisfy the question because in the question variable key is defined the following way
 
>char key[5]="april";
 
>So your code is wrong and does not solve the problem described in the question.
 
On the contrary, I answered the question posed in the title: How to
compare strings.
 
My solutions addressed his misunderstanding of how C strings are
initialized and how to compare strings for exact match.
 
The char key[5] = "april" is a frequent novice mistake when
initializing C strings and the question is clearly from a novice who
is intending to use strings as arrays rather than character arrays. So
he forgets the extra null character in the key, then he gets a C
string from scanf and tries to compare it to the key invalid C string
in key using ==, probably because he comes to C from another language
that knows what to do with == when comparing strings, such as C++, for
instance, and is surprised when his program doesn't behave as he
expected when handling what he thinks are strings.
 
Incidentally, you missed an opportunity to criticize my C solution
since it uses fgets and fails to take into account the '\n' returned
in the string. This was a carry-over from some experimentation I was
doing and was posted by mistake.
 
Correct C solution:
 
#include <stdio.h>
#include <string.h>
 
int main(void){
char key[] = "april"; // auto size array as string
char ckey[10];
 
printf("Enter the key: ");
scanf("%9s", ckey);
 
// strings must match exactly.
if(strlen(key) == strlen(ckey) && !strcmp(key, ckey)) {
printf("Correct.\n");
}
else {
printf("Wrong.\n");
}
return 0;
}
Geoff <geoff@invalid.invalid>: Nov 15 04:17PM -0800

On Sat, 14 Nov 2015 09:56:21 -0800 (PST), Vlad from Moscow
 
>> Your answer was not absolutely correct. It fails to say "Wrong." when
>> the input does not exactly match the key. A serious defect.
 
>It is interesting to me when does it fail to say "wrong" when the input does not exactly match the key?:)
 
Your solution fails to say "Wrong" and says "Correct" when the input
matches the first five characters but doesn't exactly match. Input
such as aprila, aprilaaaaaaaaaa, all produce "Correct when in fact
this would be erroneous since the arrays don't exactly match.
Vlad from Moscow <vlad.moscow@mail.ru>: Nov 16 12:25AM -0800

On Monday, November 16, 2015 at 3:17:52 AM UTC+3, Geoff wrote:
> matches the first five characters but doesn't exactly match. Input
> such as aprila, aprilaaaaaaaaaa, all produce "Correct when in fact
> this would be erroneous since the arrays don't exactly match.
 
 
I am sorry but you are very low qualified programmer. You are even unable to understand what is written in questions and what you can do is only read titles of questions.:)
 
Usually bubbies also are able to read only titles.:)
 
I'll help you. You can enter neither "aprila" nor "aprilaaaaaaaaaa" because the character array used to store the entered string has only 6 characters.:)
 
Such a task occurs very often when for example you read some data from a file or other source and the record contains an identification field. And you need to check that identification field. So usually programs contain valid identification strings that have the required size and compare them with this identification field. Or they need to compare a part of a string with the identification field that for example to be sure that the string starts with the identification's characters.
 
So your answer has nothing common with the code presented in the question.
 
I advice you to think about that titles does not reflect entirely a problem. There is no enough space in the title that to describe a problem. So titles are very often have too common meaning. For example a title can sound like "How to delete an element in a vector".:)
How are you going to answer a question with such a title? What is the vector? Is it standard container std::vector? Or is it just an array? Neither the first nor the second! It is a user-defined single-linked list that was called vector in some question.:)
 
It is the content of the question it is the code presented in the question that has to be taken into account.
 
And moreover in fact the title of this question does not contradict to the code example shown in the question. Indeed he need to compare a string with another in fact string that does not have the terminating zero.
 
It can be considered as a task to compare of a substring with a given number of characters with other string that has no more than the given number of characters plus one character for the terminating zero.
 
So you are able to understand neither the question nor the correct code I showed in my answer.
 
Programming is not your speciality.:)
SG <s.gesemann@gmail.com>: Nov 16 04:09AM -0800

On Monday, November 16, 2015 at 9:25:59 AM UTC+1, Vlad from Moscow wrote:
 
> I'll help you. You can enter neither "aprila" nor "aprilaaaaaaaaaa"
> because the character array used to store the entered string has
> only 6 characters.:)
 
Of course, a user can enter "aprila" or even "aprilaaaaaaaaaa". You
just ignore anything beyond 5 characters by using a short array and
"%5s" as scanf string. And that's the issue because it makes your
program accept "aprila" and "aprilaaaaaaaaaa" as correct key
regardless of whether you store the user input in full or just up to
the first five characters in the array named key.
 
While
 
char ckey[5] = "april";
 
is valid in C, I don't think it's good practice. You did not address
the danger of this approach and the intricacies of C string handling
w.r.t. null terminators.
 
> [...]
> Programming is not your speciality.:)
> [...]
 
One major factor in why your answer got downvoted and you got blocked
is this kind of attitude of yours, probably. You chose to engage in
an edit war about whether or not it's OK to insult other SO members as
part of your answer.
Vlad from Moscow <vlad.moscow@mail.ru>: Nov 16 04:42AM -0800

On Monday, November 16, 2015 at 3:10:25 PM UTC+3, SG wrote:
> is this kind of attitude of yours, probably. You chose to engage in
> an edit war about whether or not it's OK to insult other SO members as
> part of your answer.
 
It is very interesting can you say what are there in the array of 6 characters that contains a string beyond the five characters?:)
 
As for practise then i will repeat for such unexperienced weak programmers as you that it occurs very othen when you deal with records read from fro example a file or network. Usually some raw data contains identification fields.
 
Consider for example a record that contains a field either with "XML " or "HTML" (onlky four characters neither zero character). And you need to identify the record. In this case you can declare a character array of 5 characters initialized it with either "XML",'\0' or "HTML",'\0' and check the identification field as I shwoed in the answer.
 
One more for such unquakified programmers as you. This is valid correct code.
 
You could suggest to redesign the application (though I doubt because the record structure might exist apriori). But in any case it does not mean that my answer is incorrect.
 
My answer was down-voted for two reasons.
 
The first one is that in the West there are very many low-qualified programmers that even do not know that such an initialization for character arrays in C is valid. It is seen very well from the comment to my answer. There were other similar comments but they were deleted by the moderator.
 
And the second reason is the fascism that thrives in the West including the Stackoverflow.
 
It is so obvious that what you are doing is trying to find a reason to accuse me and to invent something that to say that my entirely correct answer is wrong.:)
Christian Gollwitzer <auriocus@gmx.de>: Nov 16 02:10PM +0100

Am 16.11.15 um 09:25 schrieb Vlad from Moscow:
 
> I am sorry but you are very low qualified programmer. You are even unable to understand what is written in questions and what you can do is only read titles of questions.:)
 
> Usually bubbies also are able to read only titles.:)
 
... and this is why you got banned from the "fascist" web site.
 
 
Christian
"Alf P. Steinbach" <alf.p.steinbach+usenet@gmail.com>: Nov 16 02:12PM +0100

On 11/16/2015 1:42 PM, Vlad from Moscow wrote:
>> an edit war about whether or not it's OK to insult other SO members as
>> part of your answer.
 
> As for practise then i will repeat for such unexperienced weak programmers as you
 
Google is your friend Vlad, even if you think we're not. Judging from
his postings in clc++, Sebastian has quite extensive experience, and is
not at all a weak programmer. Googling him & his research bears that out.
 
I think the problem is that you're trying to argue that you were
technically right, believing that nobody here UNDERSTAND your answer
(just because one or two didn't), when in fact the technical has been
cleared up long ago and people such as Sebastian are trying in a very
friendly non-adversary way to help you understand what you did wrong.
 
So, what you did wrong:
 
• You focused on a single, minor technical aspect; one that considering
the OP's level of expertise was unlikely to be relevant.
 
• You engaged in a flame war.
 
• You did not properly adjust to changing circumstances, such as
recognizing that the people you talk to here, as opposed to on SO,
UNDERSTAND the technical point about C versus C++ rules etc.
 
So that's what to improve: (1) try much harder to adjust to continually
changing circumstances, (2) opt out of flame wars that you are not
intentionally trying to prolong, e.g. enlist the help of moderators in
such cases, and (3) look beyond your first impression and first obvious
issue, take a wider view where you consider the relevance of each thing.
 
Cheers & hth.,
 
- Alf
asetofsymbols@gmail.com: Nov 16 07:08AM -0800

The OP in initial post said:
"Is it possible to solve the problem without using other libraries?"
The answer is yes
So answer that use strcmp or
fgets or other different from
scand()
has to be not full voted...
Daniel <danielaparker@gmail.com>: Nov 16 07:02AM -0800

> One presumes class data members would be private, else what's the point of having a class?
 
> Given that, putting the data members first would be contrary to the generally accepted public, protected, private ordering with a class.
 
There's at least one practical c++ specific reason for putting the data members first, it makes it easier to line up the constructor initializer lists with the order in which data members are declared, which is a good thing to do in C++.
 
Public, protected, private ordering is just a convention, it's usefulness to the developer is vanishing with modern practices using templates, inline functions etc., try quickly groking any of the standard library or boost classes by glancing at their headers. At the same time, tooling has become quote good for introspecting classes within an editor environment, and for producing class documentation.
 
> It is the same problem of "const"...
 
The problem with const is that it doesn't mean pure, which it would have to mean were compilers able to take advantage of it, or programmers to reason about it.
 
Daniel
agent@drrob1.com: Nov 16 07:59AM -0500

Hi. I am learning C++ after many years of using modula-2 on windows.
 
I now am working my way thru C++ all in one for dummies, by John Paul
Mueller and Jeff Cogswell.
 
On the chapter about getting the contents of a directory, they
describe io.h and _findfirst(), _findnext(), and struct _finddata_t.
 
This is not working under ubuntu 14.04. I located a version of io.h
in the /usr/src/linux-headers tree, copied it to my working directory,
but I'm still getting a lot of undeclared symbol errors.
 
How do I read the contents of a directory in Ubuntu 14.04?
 
Thanks,
Rob
"Alf P. Steinbach" <alf.p.steinbach+usenet@gmail.com>: Nov 16 02:17PM +0100

> in the /usr/src/linux-headers tree, copied it to my working directory,
> but I'm still getting a lot of undeclared symbol errors.
 
> How do I read the contents of a directory in Ubuntu 14.04?
 
Sounds like an ungood book, because those `_xxx` identifiers are
internal to the implementation, i.e. they're compiler and platform specific.
 
For portable code, use the Boost file system library.
 
It will probably be part of C++17.
 
 
Cheers & hth.,
 
- Alf
scott@slp53.sl.home (Scott Lurndal): Nov 16 02:44PM

>in the /usr/src/linux-headers tree, copied it to my working directory,
>but I'm still getting a lot of undeclared symbol errors.
 
>How do I read the contents of a directory in Ubuntu 14.04?
 
man readdir
Paavo Helde <myfirstname@osa.pri.ee>: Nov 15 05:44PM -0600

jacobnavia <jacob@jacob.remcomp.fr> wrote in
 
> Remember, this is C++, not C. Everything must be a lot more complex
> than it needs to!
 
>:-)
 
In C++ one can choose the exact level of complexity one wants. Literal
magic constants are valid C++ and have lots of drawbacks. The #define
constants are valid C++ and have lots of drawbacks. The enum constants
are valid C++ as well and have some drawbacks. Enum classes are valid C++
11 and have other kind of drawbacks (the complexity). In C++ one can
choose the needed level of safety-maintainability-simplicity as desired
all over the full scale, that's its bless and curse.
 
Short one-off hacks could best do with direct numeric literals, whereas
in large long-living projects robustness, safety, ease-of-maintenance
issues trump all other concerns like the effort of writing some more
lines for each enum. The thing is, one-off hacks can be written in any
language, most of which are simpler than C++, while for large long-living
projects C++ is one of the few truly viable options.
 
Cheers
Paavo
"Öö Tiib" <ootiib@hot.ee>: Nov 15 04:30PM -0800

On Monday, 16 November 2015 00:19:37 UTC+2, jacobnavia wrote:
 
> I can now say
 
> if (val &(Sent|Forwarded))
 
> to test for some bit.
 
Too cryptic.
I like it better to read as 'if (val[Sent] or val[Forwarded])'.
Code:
 
#include <bitset>
#include <iostream>
 
int main()
{
// Jacob's enum with magic numbers removed
enum DocState
{
Archived,
Pending,
Received,
Sent,
Forwarded,
Count_DocStates
};
 
// typedef the combined bitfield type DocStates
typedef std::bitset<Count_DocStates> DocStates;
 
// we have our bifield "val"
DocStates val;

// we can put some flags into it
val.set(Archived).set(Pending);
 
// Irrelevant crap that might be useful sometimes with such values
std::cout << "bitfield of val: " << val << '\r'
<< "number of bits: " << val.count() << '\r'
<< "val reversed: " << ~val << '\r';
 
// what Jacob wanted was
// if (val &(Sent|Forwarded))
// I consider it about as good to read
if (val[Sent] or val[Forwarded])
{
std::cout << "It was sent or forwarded \r";
}
else
{
std::cout << "It wasn't sent or forwarded \r";
}
}
 
Output:
 
bitfield of val: 00011
number of bits: 2
val reversed: 11100
It wasn't sent or forwarded
David Brown <david.brown@hesbynett.no>: Nov 16 01:51AM +0100

On 15/11/15 23:43, jacobnavia wrote:
 
> I am enumerating the different states that a message can have: archived,
> sent, whatever.
 
> I do not see why you suppose that it is not an enumeration!
 
No, you are defining names for a set of bits - that is not the same as
an enumeration.
 
However, an "enum" is often the clearest and neatest way to achieve this
in C, and I think it is not unreasonable to do it that way.
 
You are right that a C++ enum class can't do that, because you can't
convert the enumeration constants to integers. Fair enough, a C++ enum
class won't cover that usage. But for more normal enumerations, an enum
class is a stronger and better scoped alternative to using the old, weak
enums. So in C++, use enum classes where possible, and old enums
otherwise. (Or use a selection of constexpr const's inside a class or
struct.)
Ian Collins <ian-news@hotmail.com>: Nov 16 03:05PM +1300

jacobnavia wrote:
>> }
 
> That's 13 lines of code to be done for each of those enums to do what
> plain enum gives me for free.
 
To be fair, the operators could be on one line and only one of them is
required for your example... So that's only one extra line (not to
mention your "13 lines" included an example!).
 
Most enum types don't require logical operators and they don't need to
be converted to int.
 
> Great Progress!
 
Indeed it is. Being able to specify the size of an enum type is
something that has been missing for way too long so has giving
enumerators their own scope.
 
> Look this is C++, I am no expert. I will just go on using enums...
 
Which is one of the beauties of C++: you have that choice.
 
--
Ian Collins
Martin Shobe <martin.shobe@yahoo.com>: Nov 15 10:17PM -0600

On 11/15/2015 6:51 PM, David Brown wrote:
> enums. So in C++, use enum classes where possible, and old enums
> otherwise. (Or use a selection of constexpr const's inside a class or
> struct.)
 
Is everyone forgetting that you can overload the & operator (and friends)?
 
Martin Shobe
David Brown <david.brown@hesbynett.no>: Nov 16 08:38AM +0100

On 16/11/15 05:17, Martin Shobe wrote:
>> otherwise. (Or use a selection of constexpr const's inside a class or
>> struct.)
 
> Is everyone forgetting that you can overload the & operator (and friends)?
 
It's a balance between convenience and type safety. Yes, you /can/ make
an enum class and overload & and |, or perhaps make a conversion
operator to int. But that's quite a bit of extra coding compared to
simply using an old-fashioned enum. Is it really worth that extra
effort? With C++, you can choose either way - whatever makes most sense
for /your/ code.
Christian Gollwitzer <auriocus@gmx.de>: Nov 16 10:54AM +0100

Am 15.11.15 um 23:37 schrieb Ian Collins:
> class was to prevent inadvertent conversion to int. Given bit fields
> are unsigned types, enum class allows you to specify the underlying
> type.
 
This is still misusing enums for bitfields. There is a syntax for
bitfields in C++
 
struct foo {
bool sent : 1;
bool read : 1;
};
 
and then you can test for multiple conditions without the bitoperator hack:
 
foo x;
 
if (x.sent && x.read) { do_something(); }
 
 
I would expect that the compiler can optimize this to a bitoperation.
 
Christian
Martin Shobe <martin.shobe@yahoo.com>: Nov 16 08:26AM -0600

On 11/16/2015 1:38 AM, David Brown wrote:
> simply using an old-fashioned enum. Is it really worth that extra
> effort? With C++, you can choose either way - whatever makes most sense
> for /your/ code.
 
As I tend to prefer the type safety, I do consider adding the couple of
lines of code needed worth the effort. I understand that others might
not consider it worth the effort and I'm find with that. My response was
intended to address the claim that, "a C++ enum class won't cover that
usage." (I also had read the other threads that mentioned the
possibility at that point.)
 
Martin Shobe
David Brown <david.brown@hesbynett.no>: Nov 16 03:33PM +0100

On 16/11/15 15:26, Martin Shobe wrote:
> intended to address the claim that, "a C++ enum class won't cover that
> usage." (I also had read the other threads that mentioned the
> possibility at that point.)
 
Fair enough. I should have written that a C++ enum class won't cover
that usage without additional code. (It is also possible to static_cast
the enum class element to an int or other underlying type directly in
the code. But that's ugly.)
scott@slp53.sl.home (Scott Lurndal): Nov 16 02:41PM

>> and C++.
 
>I think you missed the point. In both C and C++ a better alternative to
>#defines for error codes exists and has existed for decades: a simple enum.
 
Although a simple enum is a signed integer. I'd prefer unsigned values
to be declared as unsigned and one can't do that with enum, unless one is
using C++11 or higher.
Juha Nieminen <nospam@thanks.invalid>: Nov 16 09:43AM

> size_type
> max_size() const _GLIBCXX_NOEXCEPT
> { return _M_get_Node_allocator().max_size(); }
 
Curiously, in libc++ (which is mainly used by clang) it's defined as:
 
size_type max_size() const _NOEXCEPT
{return numeric_limits<difference_type>::max();}
 
which, I believe, is the correct implementation.
 
--- news://freenews.netfront.net/ - complaints: news@netfront.net ---
Ramine <ramine@1.1>: Nov 15 06:57PM -0800

Hello,
 
 
I feel that i must explain to you how do work my inventions that are my
SemaCondvar and SemaMonitor objects, you will find those classes inside
the SemaCondvar.pas file inside the zip file, SemaCondvar and
SemaMonitor are new and portable synchronization objects , SemaCondvar
combines all the charateristics of a semaphore and a condition variable
and SemaMonitor combines all charateristics of a semaphore and an
eventcount , they only use an event object and a very fast and efficient
and portable FIFO fair Lock , so they are fast and they are FIFO fair.
 
When you set the first parameter of the constructor to true it will add
the characteristic of a Semaphore to the condition variable or to the
Eventcount, so the signal will not be lost if the threads are not
waiting for the SemaCondvar or SemaMonitor objects, but when you set the
first parameter of the constructor to false it will not behave like a
Semaphore because if the threads are not waiting for the SemaCondvar or
SemaMonitor the signal will be lost..
 
Now you can pass the SemaCondvar's or Semamonitor's initialcount and
SemaCondvar's or SemaMonitor's MaximumCount to the construtor, it's like
the windows Semaphore`s InitialCount and the Semaphore's MaximumCount.
 
Like this:
 
t:=TSemaMonitor.create(true,0,4);
You have 5 options in the defines.inc file for setting the kind of
locks, just look inside defines.inc , if you want to set it for the
Mutex that is energy efficient because it blocks the threads, uncomment
the option Mutex, if you want to set it for my node based scalable Lock,
uncomment the option MLock, if you want to set it for my scalable array
based lock called AMLock just uncomment the option AMLock inside
defines.inc, if you want to set it for Ticket Spinlock just uncomment
the option TicketSpinlock ,If you want to set it for Spinlock just
uncomment the option Spinlock.
That's all.
 
You can download my SemaMonitor and SemaCondvar from:
 
https://sites.google.com/site/aminer68/light-weight-semacondvar-semamonitor
 
and from:
 
https://sites.google.com/site/aminer68/semacondvar-semamonitor
 
 
Please feel free to port it to C++...
 
 
Thank you,
Amine Moulay Ramdane.
bleachbot <bleachbot@httrack.com>: Nov 16 12:56AM +0100

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)