converting variables

[Nate]

I am working on an Oakley parser and want to call an fopen function to
read in the log file. I can use this to read the file, but only if I
pass a "const char" variable to fopen. Since I would like the end
user to specify the name of the log file, I have a scanf in there to
allow the user to do so, but it isn't a const at that point. I have
seen reference to calls like static_cast, dynamic_cast and
reinterpret_cas t in C that will allow me to convert to another type,
but I am having a little trouble using them.

This is what I tried, with no luck:

printf("Enter the log file name: ");

scanf ("%c", &filename);

reinterpret_cas t <const char>( filename );

and get this error:

OakleyParser.cp p(14) : error C2440: 'reinterpret_ca st' : cannot
convert from 'char' to 'char'

So VS.Net recognizes both as Char, but to use the char variable
‘filename' in fopen, I have to have a constant.

file_in = fopen(filename, "r");

OakleyParser.cp p(20) : error C2664: 'fopen' : cannot convert parameter
1 from 'char' to 'const char *'

Any ideas?

[Sam Dennis]

Nick Keighley wrote:

[for reading a line]
scanf ("%s", filename);
int result = scanf( "%" STR( FILENAME_MAX ) "[^\n]", filename );
or better, avoid scanf() (its error recovery is poor) and use fgets()

Actually, when used in a few certain ways, it can replace fgets and
simplify error checking at the cost of a little hassle if one needs
to change the field width(s). (This would be a much more difficult
decision if fgets returned the number of characters read; as things
stand, though, (f)scanf seems the clear winner.)

I feel strangely enlightened.

--
++acr@,ka"

[Dan Pop]

In <8a************ **************@ posting.google. com> ni***********@m arconi.com (Nick Keighley) writes:

"Mabden" <ma****@sbcglob al.net> wrote in message news:<qH******* *********@newss vr27.news.prodi gy.com>...

"Nick Keighley" <ni***********@ marconi.com> wrote in message
news:8a******** *************** ***@posting.goo gle.com...

> na**@nateharris .com (Nate) wrote in message

news:<7a******* *************** ****@posting.go ogle.com>...

> or better, avoid scanf() (its error recovery is poor) and use fgets()
> (you'll have to check for '\n' and embedded spaces.

What's wrong with embedded spaces?

the scanf() solution won't have embedded spaces (as someone else pointed out
%s stops at the first whitespace).

There's more to scanf than %s... Shall we also discard printf for
floating point values because %d expects an int value?

Dan
--
Dan Pop
DESY Zeuthen, RZ group
Email: Da*****@ifh.de

[CBFalconer]

Dan Pop wrote:

ni***********@m arconi.com (Nick Keighley) writes:

.... snip ...

the scanf() solution won't have embedded spaces (as someone
else pointed out %s stops at the first whitespace).

There's more to scanf than %s... Shall we also discard printf
for floating point values because %d expects an int value?

It would be nice. However we don't have a standardized set of
output routines, while we do have some suitable input routines in
strto**.

%s is what the newbie (to scanf) reaches for. Surprise. What is
needed is a standard set of i/o routines for primitive types
to/from text streams.

Eschew variadic functions.

--
Chuck F (cb********@yah oo.com) (cb********@wor ldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home .att.net> USE worldnet address!

[Barry Schwarz]

On Wed, 19 May 2004 10:37:57 +0800, Mitchell
<ch************ ***@inaHATESPAM me.coHATESPAMm> wrote:

On 18 May 2004 09:35:22 -0700, na**@nateharris .com (Nate) wrote:

printf("Enter the log file name: ");

scanf ("%c", &filename);

reinterpret_cas t <const char>( filename );

How did you declare filename?

char filename[1024];

or something similar? In this case, your scanf line should read

scanf("%s",fil ename);

because filename is already the address of the character array.
&filename points to the sky. Also, use %s !!!!! %c is for reading in a
character.

If filename is declared as an array as you suppose, &filename points
to the exact same address filename does. The problem is it would have
the wrong type. The %s format requires a pointer to char (which
filename is) while &filename would have pointer to array of char.

OakleyParser. cpp(14) : error C2440: 'reinterpret_ca st' : cannot
convert from 'char' to 'char'

a=reinterpret_ cast<char>(x);

is c++'s way of saying

a=(char) x;

for nonpolymorphic types.

So VS.Net recognizes both as Char, but to use the char variable
‘filename' in fopen, I have to have a constant.

file_in = fopen(filename, "r");

Oh, there's no need! =) Just feed it a char. It will be implicitly
casted to const char.

The first parameter of fopen must be a pointer to char, not a char.
<<Remove the del for email>>

[Barry Schwarz]

On Wed, 19 May 2004 20:04:36 GMT, CBFalconer <cb********@yah oo.com>
wrote:

Mabden wrote:

"Nick Keighley" <ni***********@ marconi.com> wrote in message

na**@nateharris .com (Nate) wrote in message

or better, avoid scanf() (its error recovery is poor) and use
fgets() (you'll have to check for '\n' and embedded spaces.)

What's wrong with embedded spaces?

from N869, conversion specifiers for fscanf:

s Matches a sequence of non-white-space
characters.228)

But the question was what is wrong with spaces in a filename, which
fgets will allow but NK seems to think need special attention.
<<Remove the del for email>>

[CBFalconer]

Barry Schwarz wrote:

CBFalconer <cb********@yah oo.com> wrote:

Mabden wrote:

"Nick Keighley" <ni***********@ marconi.com> wrote in message
na**@nateharris .com (Nate) wrote in message

or better, avoid scanf() (its error recovery is poor) and use
fgets() (you'll have to check for '\n' and embedded spaces.)

What's wrong with embedded spaces?

from N869, conversion specifiers for fscanf:

s Matches a sequence of non-white-space
characters.228)

But the question was what is wrong with spaces in a filename, which
fgets will allow but NK seems to think need special attention.

The original code stopped input at the first trailing white space,
so the fgets call needs to be treated differently. As to what's
wrong with spaces in a filename, that is another large diatribe
which includes letting software firms without ethics control
standards.

--
"The most amazing achievement of the computer software industry
is its continuing cancellation of the steady and staggering
gains made by the computer hardware industry..." - Petroski

[Nick Keighley]

Da*****@cern.ch (Dan Pop) wrote in message news:<c8******* ***@sunnews.cer n.ch>...

In <8a************ **************@ posting.google. com> ni***********@m arconi.com (Nick Keighley) writes:

"Mabden" <ma****@sbcglob al.net> wrote in message news:<qHNqc.137 2
$i******@newssv r27.news.prodig y.com>...

"Nick Keighley" <ni***********@ marconi.com> wrote in message
news:8a******** *************** ***@posting.goo gle.com...
> na**@nateharris .com (Nate) wrote in message
> news:<7a******* *************** ****@posting.go ogle.com>... > or better, avoid scanf() (its error recovery is poor) and use fgets()
> (you'll have to check for '\n' and embedded spaces.

What's wrong with embedded spaces?

the scanf() solution won't have embedded spaces (as someone else pointed out
%s stops at the first whitespace).

There's more to scanf than %s... Shall we also discard printf for
floating point values because %d expects an int value?

I wasn't criticising %s format, just pointing the behaviour of scanf("%s")
is different from fgets(). Not better or worse, just different. I've no idea
if the OP wants to allow embedded spaces or not- I'm just saying he has to
allow for a change in behaviour.
--
Nick Keighley

[Dan Pop]

In <8a************ **************@ posting.google. com> ni***********@m arconi.com (Nick Keighley) writes:

Da*****@cern.c h (Dan Pop) wrote in message news:<c8******* ***@sunnews.cer n.ch>...

In <8a************ **************@ posting.google. com> ni***********@m arconi.com (Nick Keighley) writes:

>"Mabden" <ma****@sbcglob al.net> wrote in message news:<qHNqc.137 2
> $i******@newssv r27.news.prodig y.com>...
>> "Nick Keighley" <ni***********@ marconi.com> wrote in message
>> news:8a******** *************** ***@posting.goo gle.com...
>> > na**@nateharris .com (Nate) wrote in message
>> > news:<7a******* *************** ****@posting.go ogle.com>... >> > or better, avoid scanf() (its error recovery is poor) and use fgets() ^^^^^^^^^^^^^ >> > (you'll have to check for '\n' and embedded spaces.
>>
>> What's wrong with embedded spaces?
>
>the scanf() solution won't have embedded spaces (as someone else pointed out
>%s stops at the first whitespace).

There's more to scanf than %s... Shall we also discard printf for
floating point values because %d expects an int value?

I wasn't criticising %s format, just pointing the behaviour of scanf("%s")
is different from fgets(). Not better or worse, just different. I've no idea
if the OP wants to allow embedded spaces or not- I'm just saying he has to
allow for a change in behaviour.

Who wrote: "avoid scanf()" above?

Dan
--
Dan Pop
DESY Zeuthen, RZ group
Email: Da*****@ifh.de

[Dan Pop]

In <40************ ***@yahoo.com> CBFalconer <cb********@yah oo.com> writes:

Dan Pop wrote:

ni***********@m arconi.com (Nick Keighley) writes:

... snip ...

the scanf() solution won't have embedded spaces (as someone
else pointed out %s stops at the first whitespace).

There's more to scanf than %s... Shall we also discard printf
for floating point values because %d expects an int value?

It would be nice. However we don't have a standardized set of
output routines, while we do have some suitable input routines in
strto**.

Which of them is suitable for getting the name of an input file from the
user? The strtox routines are useful *only* after successfully obtaining
the input while this thread was focused on getting that input.

Have I ever recommended you to engage your brain *before* replying?

Dan
--
Dan Pop
DESY Zeuthen, RZ group
Email: Da*****@ifh.de

[Mabden]

"Dan Pop" <Da*****@cern.c h> wrote in message
news:c8******** **@sunnews.cern .ch...

In <40************ ***@yahoo.com> CBFalconer <cb********@yah oo.com> writes:

Dan Pop wrote:

ni***********@m arconi.com (Nick Keighley) writes:

... snip ...

Hey! you snipped my "What's wrong with embedded spaces?" comment! I was
getting a lot of mileage out of that one!

--
Mabden