-
Notifications
You must be signed in to change notification settings - Fork 9
Description
My decomp project contains a large amount of SJIS-encoded strings, and while mwccgap has a -target-encoding switch (i think to pass along to mwcc?) it doesn't use the encoding specified while preprocessing asm files. As an example, one of my .rodata files contains a four kanji string, and without further processing mwccgap inserts half of the ascii representation into the resulting code (four kanji, represented by twelve bytes in ascii (plus one for nul ending), or by eight bytes in sjis (plus one for nul), but in the output only the six bytes from the first two kanji as ascii are present.)
i've tried running the files through iconv first to convert them to sjis prior to sending them through mwccgap, but this results in errors like "Exception: Failed to preprocess (file): 'utf-8' codec can't decode byte 0x8d in position 80: invalid start byte". looks like more codecs (or at least sjis) should be implemented, chosen through that -target-encoding switch.