Quote Originally Posted by Saffith View Post
It seems to work fine in 2.50.3 RC1 and 2.53 beta 2, so I'm inclined to think that's not really the problem. But I don't know why it makes any difference if that's the case.
I really don't know right now. I rolled this stuff back, and I'm letting it sit on simmer. I had complaints that stuff broke, before I did anything remotely related to those functions, which is why I tinkered with them in the first place. I'm inclined to label them as 2.50 glass, and to implement something better in versions after 2.53, that are saner, cleaner, and more well-wrought. Both ComboData, and ScreenData are part of the master plan at this point, so, it might just be better to leave them alone until we have that stuff done.

I don't know of any quests that would break by reverting them to ~build 1793, but I know of four that break in different ways, by not reverting them, and deducing why, at this point, has drained me entirely of emotion of any sort. I'm in a sort of mechanical fugue state at present, so unless there is some overwhelming reason to further tinker with these things before 2.54 or 2.55 or something truly new, I will probably leave them be in a state that I know is working, and not erratic.

I may look at how they were set up again in 2.50.3RC1, but AFAIK, nothing changed from 2.53b2 to 2.53b7 that would have made them break. What I did notice, is a completely erratic behaviour pattern; at least when compiled with MSVC9. I'll test the files in beta 2 though, just to see what happens.

Apparently, MSVC also wanted to completely wreck DrawInteger, so I had to fix that. The static cast of sdci[9] to float, followed by a cast to ( int ) was --somehow-- causing values to output at value-1 when were passed to sprintf(). (ie.g. A value of 5 was printing as 4.)

How, why? IDK. Casting from long to float to int is ridiculous anyway; so it no longer does that. The only time it static casts to float now, is when you are trying to display decimals.

Let's see though... I woke at around 03:00GMT, and it's 19:50GMT. Discounting the time that I spent going to market, and preparing my daily meal, I have been working on this mire for at least twelve hours today. I committed all of the changes, applying them to master and 2.50.x branches; except for the reversion of these functions. That is sitting in its own branch, for a day. It was beastly to reconstruct without accurate SVNs of 2.50.2.

Luckily, I did find some old files that had some ffscript.cpp stuff in them on my RAID array. Now I just want to have me whiskey and stare at something mindless for a while.