Old bug in macros.h

[ASM]

There is a bug in CHANCE16 macros which causes industry growth miscalculation with smooth economy. Its in SVN since revision 1.
Patch for 0.4.7 attached, should apply to trunk too.

[Ryalla]

There is a bug in CHANCE16 macros which causes industry growth miscalculation with smooth economy. Its in SVN since revision 1.
Patch for 0.4.7 attached, should apply to trunk too.

If it's been in macros.h since revision 1, how do you know it's an error? :P

[ASM]

If it's been in macros.h since revision 1, how do you know it's an error?

Because it makes smooth economy unusable
With it we have 20/1024 chance of production decrease and same chance of production increase, so industries doesn't grow at all.

[SirkoZ]

If it's been in macros.h since revision 1, how do you know it's an error?

Because it makes smooth economy unusable
With it we have 20/1024 chance of production decrease and same chance of production increase, so industries doesn't grow at all.

Indeed it is - smooth economy without my improved patch (didn't touch the macros though) is totally unusable - industries mostly decrease if they are being used and increase if they are not used... :-S

At any rate - I'll test your patch...

[SirkoZ]

Wow - it's already been included into trunk...

It works - quite nicely - now 2 more bugs with the current smooth_economy - IT_FARM_2 = tropical maize_farm doesn't plant any fields and once industry production decreases to 32/40 units per month it does not increase anymore, which is annoying - here is a quick fix for those 2 bugs.

[henrik]

And this is why macros are evil and should rarely be used. Inline functions are so much better for this kind of thing!

[MeusH]

Can anyone explain me the patch by ASM? How do a and (a) differ?

[SirkoZ]

And this is why macros are evil and should rarely be used. Inline functions are so much better for this kind of thing!

Oh cry me a river - it's fixed now, ain't it?

[SirkoZ]

Can anyone explain me the patch by ASM? How do a and (a) differ?

It takes into account all of the a, especially if a is something like
x + ((20 * pct_transported) >> 8 ).

[hertogjan]

Can anyone explain me the patch by ASM? How do a and (a) differ?

It's about what happens at the precompiling stage. Suppose you have the code

Code: Select all

#define SQUARE(a) a*a

which you want to use to square a number.
Now if you put in something like

Code: Select all

SQUARE(2+3);

it is changed to 2+3*2+3 (it just replaces a by 2+3), which equals 11, and not 25 which you were after. To prevent this from happening, you'll have to define

Code: Select all

#define SQUARE(a) (a)*(a)

Now

Code: Select all

SQUARE(2+3);

will be changed to (2+3)*(2+3), yielding the correct result 25.

[edit]Whoops. Don't forget to check when copy-pasting next time [/edit]

[glx]

Code: Select all

#define SQUARE(a) (a)*(a)

return the correct value

[Darkvater]

Code: Select all

#define SQUARE(a) (a)*(a)

return the correct value

Let's turn it into this, shall we?

Code: Select all

#define SQUARE(a) ((a)*(a))

[MeusH]

Thank you so much for answers