On Mon, Jan 05, 2004 at 11:35:12AM +0100, Klaus Schmidinger wrote: > Hmmm, I wonder why the check against 0xFF was introduced in the first place? > It wasn't there in the original 'libdvt' code written by Rolf Hakenes, > and I'd say it isn't necessary, anyway, since 'char' is assumed to be 8 bit. That depends on your level of paranoia. As far as I read the C standard documents there is no part that claims char to be 8 bit. It just says that char must be "large enough to store any member of the basic execution character set" (ISO/IEC 9899:1999 6.2.5 par.3). On the other hand side I would expect vdr (like most software) to break anyway on systems with different sizes for char data type. > I guess I'll just remove the (*from <= 0xFF) check. I think, that's ok for that case. If vdr was software to run in some security critical area I would object, but I do not consider vdr such type of software. ;-) > (Sorry, Robert, this is the second fix in a row that you have suggested > but I have done differently - no offense ;-) I don't feel offended by such things. This might have been the case if you did something completely stupid instead and claimed to have reinvented the wheel. --- And even then, most likely, I wouldn't have cared. ;-) But to come back to another issue: If you add -W to the command line many unused parameter, missing initializer, and such stuff warnings do appear. If you are interested in having cleaned them up, I can do a patch for that. But I will do a patch only if there is real interest. Because if you don't care about unused parameters or uninitialized variables it is pointless to do a patch. I personally would recommend cleaning that up because it often shows hidden bugs, but this is your decision. Robert -- Robert Schiele Tel.: +49-621-181-2517 Dipl.-Wirtsch.informatiker mailto:rschiele@uni-mannheim.de
Attachment:
pgp00002.pgp
Description: PGP signature