length of char array

**zacs7** · 09-17-2007

See the FAQ for flusing the input buffer then, http://faq.cprogramming.com/cgi-bin/...&id=1043284392

**ssharish2005** · 09-17-2007

Originally Posted by juststartedC

Hey taurus

have you finished the psuedocode yet???

Taurus your class mate is asking you somenthing LOL

ssharish

**keira** · 09-17-2007

You have to understand that user input is kept in a buffer called stdin. It is dynamic and resizes itself so you don't have to worry about making things fit into it. When you store 20 bytes (characters) of data into an array you are using a function to read 20 bytes from stdin and store them into the array. If you don't tell it to read 20 bytes (if you use a function like scanf() where the array size is not passed to the function, it will read bytes from stdin until it encounters a NULL or a byte represented in hex as 00 meaning the end of the string). If you understood nothing I just said, just remember to use fgets and specify the length of your array and you will never run into problems concerning overflow.

**CornedBee** · 09-17-2007

1) The data is read from a stream, not a buffer. A buffer may be in there, but it is irrelevant to the concept of reading from a stream.
2) You can tell scanf() the buffer size:

Code:

scanf("&#37;20s", buffer);

It even works for dynamically sized buffers:

Code:

scanf("%*s", size, buffer);

3) The limit character for fgets() is the newline, '\n', not the NUL. A text stream will in fact never emit a NUL; at best it will emit an EOF.

**Dave_Sinkula** · 09-17-2007

Originally Posted by CornedBee

It even works for dynamically sized buffers:

Code:

scanf("%*s", size, buffer);

Hmm? In the scanf family, the * is for assignment suppression.

**CornedBee** · 09-18-2007

Oh, sorry. In that case, you have to create a format string.

Code:

sprintf(buffer1, "&#37;%%is", size);
scanf(buffer1, buffer2);

Stupid omission, IMO.

**Dave_Sinkula** · 09-18-2007

So how big do you make buffer1? [/nitpick]

**taurus** · 09-18-2007

Originally Posted by keira

You have to understand that user input is kept in a buffer called stdin. It is dynamic and resizes itself so you don't have to worry about making things fit into it. When you store 20 bytes (characters) of data into an array you are using a function to read 20 bytes from stdin and store them into the array. If you don't tell it to read 20 bytes (if you use a function like scanf() where the array size is not passed to the function, it will read bytes from stdin until it encounters a NULL or a byte represented in hex as 00 meaning the end of the string). If you understood nothing I just said, just remember to use fgets and specify the length of your array and you will never run into problems concerning overflow.

ahhhhhhh i see, so instead of using all that buffer thing i can use fgets?

**CornedBee** · 09-18-2007

Originally Posted by Dave_Sinkula

So how big do you make buffer1? [/nitpick]

ceil(log10(UINT_MAX)) + 3

**Dave_Sinkula** · 09-18-2007

Originally Posted by CornedBee

ceil(log10(UINT_MAX)) + 3

So how ugly does the scanf version get if you want its fgets equivalent?

**CornedBee** · 09-18-2007

Very ugly. I never said it was nice, only that it was possible.

**taurus** · 09-20-2007

jus a quick question. Is there a command to erase the buffer in C. say i enter a string. now i can say if string>20 then erase buffer and reinput?

**matsp** · 09-20-2007

You mean like this FAQ question:
http://faq.cprogramming.com/cgi-bin/...&id=1043284392

--
Mats

**taurus** · 09-21-2007

ya i was doing the buffer thing: scanf("%20s", buffer);

but once i enter more than 20 characters i get a segmentation fault. How can i create an error message instead?

**matsp** · 09-21-2007

Originally Posted by taurus

ya i was doing the buffer thing: scanf("%20s", buffer);

but once i enter more than 20 characters i get a segmentation fault. How can i create an error message instead?

Yes, becasue scanf is "stupid" and can't deal with things going wrong very well - there is A REASON we suggest using fgets() instead - it is more fault-tolerant and unless you do things wrong, you can't make it overwrite other variables and such.

But unfortunately, you are in a class where you can't use that [unless I'm confused], so you have to live with the application not being particularly fault-tolerant.

--
Mats