(In reply to Jorg K (CEST = GMT+2) from comment #28) > Also, keep in mind that the outgoing delsp processing was only introduced > for ISO-2022-JP, so if you scrap that, the delsp processing can be removed, > for example here: I was going to say that's not entirely true, but then I went back searching through the bugs and remembered that we have this giant cesspool of interaction between format=flowed, plaintext serializer, charset, and Content-Transfer-Encoding selection that is overcomplicated and way too deferential to supporting legacy issues that may not matter anymore, and so what ought to be the design in a smart modern system has too little implications for our existing codebase. When we can watch all of this crap code go up in a merry blaze, I will be quite happy. I haven't looked at this patch in detail yet (splinter review is crapping out on me), but I will ask that any time you assign "UTF-8" to a nsString only to immediately pass it into a UTF-16-to-ASCII (or UTF-8) conversion function, just strip out the middleman and assign it to a nsCString instead. Better yet, try to go forward and pass UTF-8 literals to directly wherever the results are being used.
Bug 862292 Comment 33 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
(In reply to Jorg K (CEST = GMT+2) from comment #28) > Also, keep in mind that the outgoing delsp processing was only introduced > for ISO-2022-JP, so if you scrap that, the delsp processing can be removed, > for example here: I was going to say that's not entirely true, but then I went back searching through the bugs and remembered that we have this giant cesspool of interaction between format=flowed, plaintext serializer, charset, and Content-Transfer-Encoding selection that is overcomplicated and way too deferential to supporting legacy issues that may not matter anymore, and so what ought to be the design in a smart modern system has too little implications for our existing codebase. When we can watch all of this crap code go up in a merry blaze, I will be quite happy. I haven't looked at this patch in detail yet (splinter review is crapping out on me), but I will ask that any time you assign "UTF-8" to a nsString only to immediately pass it into a UTF-16-to-ASCII (or UTF-8) conversion function, just strip out the middleman and assign it to a nsCString instead. Better yet, try to go forward and pass "UTF-8" literal strings to directly wherever the results are being used.