BaCon (3.2)

[alexfish]

Hi Peter

Not sure about later versions of BaCon > 3.2

Can check this

BR
Alex
' test for string or number
char str[30] = "20.30300 This is test"
LOCAL * str2 TYPE char

PROTO strod
ret# = strtod(str,&str2)

PRINT ret# FORMAT "%lf\n"
PRINT str2 FORMAT "%s\n"

PRINT "char test number : " , ret# FORMAT "%s %c\n"
PRINT "char test string : " , str2 FORMAT "%s %c\n"

PRINT "char test number : ";
PRINT ret# FORMAT " %c\n"
PRINT "char test string : ";
PRINT  str2 FORMAT "%c\n"

[Pjot]

Hi Alex,

If you use explicit formatting then you have to make sure the format string matches the variable type.

Your program cannot compile because "%c" cannot show a 'double' type.

The below should work:

' test for string or number
char str[30] = "20.30300 This is test"
LOCAL * str2 TYPE char

PROTO strod
ret# = strtod(str,&str2)

PRINT ret# FORMAT "%lf\n"
PRINT str2 FORMAT "%s\n"

PRINT "char test number : " , ret# FORMAT "%s %f\n"
PRINT "char test string : " , str2 FORMAT "%s %s\n"

PRINT "char test number : ";
PRINT ret# FORMAT " %f\n"
PRINT "char test string : ";
PRINT  str2 FORMAT "%s\n"


HTH
Peter

[alexfish]

Hi Peter

on my system Raspberry PI2 there seems to be a problem with the print , yes your demo works ,

look at the two examples ,

' test for string or number
char str[30] = "20.30300 This is test"
LOCAL * str2 TYPE char

PROTO strod
ret# = strtod(str,&str2)

PRINT ret# FORMAT "%lf\n"
PRINT str2 FORMAT "%s\n"

PRINT "char test number : " , ret# FORMAT "%s %f\n"
PRINT "char test string : " , str2 FORMAT "%s %s\n"

PRINT "char test number : ";
PRINT ret# FORMAT " %f\n"
PRINT "char test string : ";
PRINT  str2 FORMAT "%s\n"

Yes the above works on Raspy with the %f & %s :: Results'::
20.303000
This is test
char test number :  20.303000
char test string :   This is test
char test number :  20.303000
char test string :  This is test

Now My code

' test for string or number
char str[30] = "20.30300 This is test"
LOCAL * str2 TYPE char

PROTO strod
ret# = strtod(str,&str2)

PRINT ret# FORMAT "%lf\n"
PRINT str2 FORMAT "%s\n"

PRINT "char test number : " , ret# FORMAT "%s %c\n"
PRINT "char test string : " , str2 FORMAT "%s %c\n"

PRINT "char test number : ";
PRINT ret# FORMAT " %c\n"
PRINT "char test string : ";
PRINT  str2 FORMAT "%c\n"


20.303000
This is test
char test number :  �  <<< this bit ' �' does not show visually in the terminal
char test string :  
char test number :  !
char test string :

so can compare the inline format ; to the single print format
all things been = then the inline Format should show at least something

Yet can see that the single print shows that that float is Number And that the String is not a Number By the presence of '!'
and nothing as in NULL for the String.

BR
Alex

[Pjot]

Hi Alex,

Please note that C is a type-sensitive language, and needs to match formatting correctly during compile time (so not on-the-fly, during runtime).

Forcing a float value to be printed as unsigned char is literally asking for problems.

What will happen is that the first byte of the float value will be printed, but this first byte can contain any value between 0-255 and therefore may produce some bad random output.

Similar problems will occur when printing a pointer to a sequence character (=string). Because you print the first byte of the pointer, and not the first byte of the string itself.

On higher level, what is it that you try to achieve?

Thx
Peter

[alexfish]

Hi Peter

as the dem shows 'detect if number or string'

yes the Pointer = string but both will show as a number
it is suggested the 'strtod( )' can be used as a method check that string is not a number.
in this case can be seen. as in ! . IE the number can be of any value , but mod(255) .
and the string = Null since it be a Pointer.

at the other high level , as seen by the results of the code

PRINT with Format is different is different to the single PRINT x , yet as mentioned @ Not sure if it be a BaCon Problem I would have thought the Two results be ==

Or is it System problem. or compiler problem. on Raspy.

how do your results stand.

BR
Alex

[Pjot]

Hi Alex,

As far as I know there is no way in plain C to find out which variable contains a number, and which contains a string.

Actually, I have the same problem in BaCon. During conversion time, BaCon registers all variables and their types. So when needed, it can lookup the type when needed (see also TYPEOF$ which makes use of the internal BaCon registration).

Next to that, BaCon can check the '$' suffix.

So if you find a good solution, let me know!

Best regards
Peter

[alexfish]

Hi Peter


do not Know for sure . here hacked up a little bit to test address and the physical
BR
Alex
SUB test(long testbit)

PRINT testbit
REM "can do something here maybe"
ENDSUB


' test for string or number
LOCAL *str1 = " TYP 20.30300 This is test" TYPE char
LOCAL * str2   TYPE char
test((long)&str1)
test((long)str1)
PRINT "-----------------------------"
str1= "56"

test((long)&str1)
test((long)str1)

PRINT "-----------------------------"
str1= "567858585858858858585858585858858558858585555555555555555555555555555555555555555555555555555555555555"

test((long)&str1)
test((long)str1)

PRINT "-----------------------------"
str1= ""

test((long)&str1)
test((long)str1)

REM now we know SIZE OF double and float that be easy

REM test the integer long will it be the same

LOCAL testint TYPE int
PRINT "-----------------------------"

test((long)&testint)
test((long)testint)

[alexfish]

Hi Peter here a slight reverse . hope U compare & hope U be seeing the same as me
BR
Alex
SUB test(long testbit)

PRINT testbit
REM "can do something here"
ENDSUB
LOCAL testint TYPE int
PRINT "-----------------------------"

test((long)&testint)
test((long)testint)

' test for string or number
LOCAL *str1 = " TYP 20.30300 This is test" TYPE char
LOCAL * str2   TYPE char
test((long)&str1)
test((long)str1)
PRINT "-----------------------------"
str1= "56"

test((long)&str1)
test((long)str1)

PRINT "-----------------------------"
str1= "567858585858858858585858585858858558858585555555555555555555555555555555555555555555555555555555555555"

test((long)&str1)
test((long)str1)

PRINT "-----------------------------"
str1= ""

test((long)&str1)
test((long)str1)

ADDED :: think there be two left to sort out:: on these basic bits

[alexfish]

Here have cleaned up the code

although there is a variation of the numbers , here on the Raspy there is a defined pattern to them

SUB test(long testbit)

PRINT testbit
REM "can do something here"

ENDSUB

PRINT "---STRING -------------------"

LOCAL *str = " TYP 20.30300 This is test" TYPE char

test((long)&str)
test((long)str)

LOCAL testint TYPE int
LOCAL testlong TYPE long

PRINT "----INT---------------------"

test((long)&testint)
test((long)testint)

PRINT "----LONG----------------------"

test((long)&testlong)
test((long)testlong)


PRINT "---STRINGS -------------------"

LOCAL *str1 = " TYP 20.30300 This is test" TYPE char
LOCAL * str2   TYPE char
test((long)&str1)
test((long)str1)

PRINT "-----------------------------"

str1= "56"

test((long)&str1)
test((long)str1)

PRINT "-----------------------------"
str1= "567858585858858858585858585858858558858585555555555555555555555555555555555555555555555555555555555555"

test((long)&str1)
test((long)str1)

PRINT "-----------------------------"
str1= ""

test((long)&str1)
test((long)str1)

[alexfish]

Hi Peter

can also test literal , but also mentioned can doubles and float

test((long)"Hello World")

test((long)9)

test((long) 1.9)

 test((long) 1.0f)

PRINT SIZEOF(9)
PRINT SIZEOF(10.9)

[Pjot]

Hi Alex,

Not sure if this really means something - what is the variables actually contain a value?

E.g.:

LOCAL testint = 12345678 TYPE int
LOCAL testlong = 34251678 TYPE long

PRINT "----INT---------------------"

test((long)&testint)
test((long)testint)


Also, the SIZEOF is architecture dependent can can give similar results for different types. For example, a SIZEOF(char*) will give the same result as SIZEOF(void*).

BR
Peter

[alexfish]

Hi Peter

here is code run + the results of Raspberry PI2 of Note as regards the string the numbers IE 73320, 39504

look through the different values esp on first two digits . then as mention the second part of the string + the literal string value

although not tested :: are certain areas set aside for allocating the different TYPES , as shown by the number scheme , if one could call it
then here on these bits , a string could be detected , and an integer is different to the long and as is the double to the float

like U mention things could be different on other systems , but sitting here I know what I be looking at , Yes further testing would be required.
esp with other vars
forgot to mention have done two string set at LOCAL *str1 = " TYP 20.30300 This is test" TYPE char , check the four numbers esp the second of each as in the short
numbers ==.
BR
Alex

SUB test(long testbit)


PRINT testbit

REM "can do something here"

ENDSUB

PRINT "---STRING -------------------"

LOCAL *str = " TYP 20.30300 This is test" TYPE char

test((long)&str)
test((long)str)

LOCAL testint TYPE int
LOCAL testlong TYPE long

PRINT "----INT---------------------"

test((long)&testint)
test((long)testint)

PRINT "----LONG----------------------"

test((long)&testlong)
test((long)testlong)


PRINT "---STRINGS -------------------"

LOCAL *str1 = " TYP 20.30300 This is test" TYPE char
LOCAL * str2   TYPE char
test((long)&str1)
test((long)str1)

PRINT "-----------------------------"

str1= "56"

test((long)&str1)
test((long)str1)

PRINT "-----------------------------"
str1= "567858585858858858585858585858858558858585555555555555555555555555555555555555555555555555555555555555"

test((long)&str1)
test((long)str1)

PRINT "-----------------------------"
str1= ""

test((long)&str1)
test((long)str1)
PRINT "-----------------------------"
LOCAL dub TYPE double
PRINT "-Double------------------------\n"
test((long)&dub)
test((long)dub)
LOCAL flt TYPE double
PRINT "-Float------------------------\n"
test((long)&flt)
test((long)flt)
PRINT "-----------------------------"
test((long)"Hello World")
PRINT "-----------------------------"
test((long)9)
PRINT "-----------------------------"
test((long) 1.9)
PRINT "-----------------------------"
test((long) 1.0f)
PRINT "-----------------------------"
PRINT SIZEOF(9)
PRINT SIZEOF(10.9)
PRINT "DONE"

/*

tests an Rasberry PI2
---STRING -------------------
73316
39504
----INT---------------------
88492
0
----LONG----------------------
75376
0
---STRINGS -------------------
73320
39504
-----------------------------
73320
39760
-----------------------------
73320
39764
-----------------------------
73320
39528
-----------------------------
-Double------------------------

88904
0
-Float------------------------

75384
0
-----------------------------
39936
-----------------------------
9
-----------------------------
1
-----------------------------
1
-----------------------------
4
8
DONE
/*



[alexfish]

Hi Peter here is an interesting one

SUB test(long  testbit)

PRINT testbit

ENDSUB

A$ = "2500TTY"
test((long)&A$)' Check this one out see bottom
test((long)A$) : ' Check this one out see bottom

PRINT "*********************************"
REM can check these with GCC like so

FOR t = 1 TO LEN(A$)-1
 PRINT A$[t]  FORMAT "%c\n"
 test((long)&A$[t])
   test((long)A$[t])
 PRINT "---------------------------"
NEXT

LOCAL * A= "2500TTY" TYPE char
test((long)&A)
test((long)A)
 PRINT "---------------------------"
LOCAL AB= "2500TTY" TYPE char *
test((long)&AB)
test((long)AB)

71728
8298592
*********************************
5
8298593
53
---------------------------
0
8298594
48
---------------------------
0
8298595
48
---------------------------
T
8298596
84
---------------------------
T
8298597
84
---------------------------
Y
8298598
89
---------------------------
71676
38164
---------------------------
2127440180
38164

[alexfish]

A separate test of the last AB string

LOCAL AB= "2500TTY" TYPE char *
test((long)&AB)
test((long)AB)
 PRINT "---------------------------"
FOR t = 1 TO LEN(AB)-1
 PRINT AB[t]  FORMAT "%c\n"
 test((long)&AB[t])
   test((long)AB[t])
 PRINT "---------------------------"
NEXT


2123519108
38564
---------------------------
5
38565
53
---------------------------
0
38566
48
---------------------------
0
38567
48
---------------------------
T
38568
84
---------------------------
T
38569
84
---------------------------
Y
38570
89
---------------------------