David Golden wrote: > On Sat, Jul 10, 2010 at 7:10 PM, karl williamson > <public@khwilliamson.com> wrote: >> Here is a revised patch based on people's feedback. >> >> I think it is a waste of time to worry too much about the wording of the >> \000 part, as I'm close to submitting patches that clean up a number of the >> problems, and so the wording is in flux anyway. > > After reading through the revised patch, two thoughts came to mind: > > * I'd like to see octal escapes moved to the bottom of the list. The > most confusing one should not be first. I agree. Revised patch attached doing this and a couple other changes; see below. > > * Should we just flag octal escapes as "discouraged" (note -- not > "deprecated") due to the legacy confusion? Are your other patches > coming going to fix this octal mess? The next patch adds \o{} which side steps the whole problem. Later patches fix the extraneous NUL in [\8], and otherwise clean up the misleading warnings. I guess I'll fix the misleading \x warnings as well. I'd also like to change \xg to not generate a NUL, but to be just 'g', but I have a feeling that will be a backwards compatibility issue. But to that end, I added the word 'currently' in describing the behavior, to put the reader on notice that they shouldn't be relying on that. I also changed the order so that \x{} is first. And, I realized it isn't just for wide characters, but any characters, and in fact we prefer it over \x without the braces for several reasons, discussed recently here. \x without the braces is only for 'narrow' or 'short' characters. I'm not sure which would be the better term. 'wide' and 'long' are used variously in other pods.Thread Previous | Thread Next