Alright, here’s the deal. I’ve got good news, and bad news. Bad news first? Yeah, I’m a rip the band-aid off quickly guy myself. Here it is.
So I mentioned that I was seeing some weird stuff in some of the stories while testing my inserter code for the stories. The only way to do this is to actually insert the generated translations into the game and give it a shot. For pacific high school, at one point, all the dialog is one step ahead of the pictures. It got me thinking that there must have been an additional, empty string pointer to nothing for some reason. So, I tried inserting an unknown string, and turns out that fixed it. At that part. Later on, the same thing happens again.
The Taiyo and Gorin stories appear to be pretty clean. Each dialog appears in the story and has a sequential pointer for it. There are a lot of unnecessary end string codes at the end of the stings though, like up to 3 or 4 sometimes, when one is sufficent. This isn’t a problem for Taiyo and Gorin, but it is for Pacific. Which means, a pointer could technically point to one of those, and it would consume a spot in the table, but never be used, My current method of dumping the text doesn’t use the pointers at all. I just calculate them. Big mistake, apparently. I should know better than to think people did a perfect job on all this stuff, right?
What this means is that in order for me to rewrite the pointer table, I need to account for every string it handles, whether it “should” be there or not.
Essentially, I neeed to add more work to the insertion process, which begins with dumping properly. That will require also shifting around my current spreadsheet stuff in addition which I’ll have to repeat when Noel is done with his own spreadsheet.
The good news? Once I get that fixed, I’m very confident that my inserter will handle it like a rock star. All my testing so far seems to show that it should be just fine. I’ve got tools that convert the csv translation spreadsheet to json, and then shove that data into the binary files exactly where they’re told to, so once I get accurate table dumps, I can provide that data accurately now.
It really is a two steps forward, one step back kind of thing. Kind of frustrating, but also an interesting learning experience.
| Process | Progress |
|---|---|
| Translation | 14% |
| Proof reading | 0% |
| Insertion | 5% |
| Testing | 0% |
Comments
Post a Comment