:) I'd be willing to bet this particular feature won't bite you again.
This is a good point to note how useful test first development is. http://perlunit.sourceforge.net If you start with the test/validation before even writing the code, you'll seldom have an issue with this in production. -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of John Deighan Sent: Tuesday, October 17, 2006 9:13 AM To: perl-win32-users@listserv.ActiveState.com Subject: RE: what?!? At 09:37 AM 10/17/2006, George Gallen wrote: >Referencing the last element of an array (-1) without having to get >the # of elements first, is VERY useful. Using other negative positions >might not have as many uses, but when you find one, You'll be glad >it's there. The problem comes when you have a function that computes an index, it's buggy and returns a negative number. Using that index returns incorrect data instead of an undefined value, which makes it much harder to detect that bug. Besides, what's so hard about saying [EMAIL PROTECTED] as opposed to $array[-1] ? >As for bounds checking, that always (well almost always) needs to be >done by the routine for two reasons. the first is that you don't want >the system to throw out warnings and/or error messages while the program >is running (it scares the user), and secondly, you never know what the >user will enter as a reference number for an array position. You can trap errors and warnings. We usually inform the user that something went wrong, but send a detailed error message via an e-mail to ourselves. I think that most users would rather be told that something went wrong than to get incorrect data. Incorrect data that is taken as correct data can lead to very bad things. Lack of automatic bounds checking (at least in cases where the boundaries are fixed, so this doesn't really relate to Perl) is responsible for all the "buffer overflow" security holes we find outselves having to deal with. I totally agree that all data entered by users needs to be thoroughly checked, but that's a different issue than whether we programmers should have to implement bounds checking for every time we index into an array. >As they say, no program is fool proof, because fools can be so ingenious, >but the more boundry checking you have with inputs and such, then you >it gives you more of chance to tell the user what a fool they are.... > >George > > > it's in the docs. Also, I checked, and if you index past the > > beginning of an array with a negative number, you get undef, which is > > at least consistent with what happens when you index past the far end > > of an array. > > > > >Scanned for Spam and Viruses. >PCG Information Technology Services. _______________________________________________ Perl-Win32-Users mailing list Perl-Win32-Users@listserv.ActiveState.com To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs _______________________________________________ Perl-Win32-Users mailing list Perl-Win32-Users@listserv.ActiveState.com To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
