On Tue, 14 Dec 1999 [EMAIL PROTECTED] wrote: > Ok, I just commited the modifs to check compression version and abort > with an explicit message if there is a mismatch. This will avoid > unexplicit errors like the core Vadim found. We need to be very careful about messing with database formats. The general rule I've been using is that 3.1 -> 3.2, it's OK to change the format (like 3.0.x -> 3.1) but that "point releases" should be at least backwards-compatible. > * How are catastrophic errors supposed to be dealt with in htdig? You're not supposed to have them. ;-) More seriously, there haven't been cases where there's anything catastrophic. Let's say it can't find the database that was specified--it quits. For example htsearch puts up a variety of "catastophic errors." > * This leads to a bigger question of error processing.... is the > standard C++ error mechanism portable enough to be used? Do you mean exceptions? I think you'll need to ask a compile guru--I don't know what compilers support exceptions well. (I learned C++ with pretty modern compilers.) -Geoff ------------------------------------ To unsubscribe from the htdig3-dev mailing list, send a message to [EMAIL PROTECTED] You will receive a message to confirm this.
