>It also fixed the scan bug that I had earlier reported. It did so, at least in >my environment, by displaying a question mark in place of the strange Unicode >character. nmh 1.6 displayed the actual character, though it messed up the >spacing. Either approach is fine with me. I prefer the 1.7 approach, though >others may differ.
Well, we were trying to get smarter about these things. It messed up the spacing in 1.6 because wcwidth() returned -1 (which means "not printable") so that confused the spacing calculator. That was "improved" by adding an assertion for 1.7, but we didn't realize that characters that passed other tests might not be printable, so that's why you had that assertion. Now we correctly handle this case and if wcwidth() returns -1 we decide it's unprintable and return the substitution character. --Ken _______________________________________________ Nmh-workers mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/nmh-workers
