parser-devel

Re: [Parser-devel] [token.c] Test case update

[Noam P. <npo...@uw...>]

Jinesh K J wrote:
>> So you're saying escapes only occur within quotes? I'm not sure that
>> this makes sense: what if you wanted to put in a quote at then end of a
>> token, eg [hello\" world] -> [hello"]
> If someone want to use a quote, then let them use them inside a quote.
> For the above case that you've mentioned, the solution would be:
> 
> ["hello\""] -> [hello"]
> or
> ['hello"'] -> [hello"]
> 
> What do you think?
> 
Ok, I guess I was confused because I was looking at the test cases with 
the shell, (using echo "foo") to get rid of the backslashes. I've 
updated the repository so all the test cases now work.

I think the next thing to do is handle invalid input: eg unclosed 
quotes, right now that causes a segfault. I'm not sure what the 
appropriate way of signalling an error would be.

Noam

Re: [Parser-devel] [token.c] Test case update

[Jinesh K J <jin...@gm...>]

> Ok, I guess I was confused because I was looking at the test cases with
> the shell, (using echo "foo") to get rid of the backslashes. I've
> updated the repository so all the test cases now work.
>
Yes! Now everthing is fine. I'll now start working on test cases for puttoken().

> I think the next thing to do is handle invalid input: eg unclosed
> quotes, right now that causes a segfault. I'm not sure what the
> appropriate way of signalling an error would be.
Ohh! That's a very serious issue. Good that you reminded me. Don't worry.
If the quote doesn't close, let the '\0' at the end of string be taken
for granted.

>
> Noam

BTW, I've added -Wall to cc and some warnings have been removed. One
is still left,
which is in token.c. See below:

[jinesh@jinesh np]$ make
cc -Wall -g -D DETAILED=0 token.c main.c
token.c: In function 'gettoken':
token.c:78: warning: ignoring return value of 'realloc', declared with
attribute warn_unused_result

I hope that the realloc return value could be used. What so you say?

Jinesh.