堆栈下溢执行RETLW指令

你好，我一直在用浮点来处理数学表达式的问题。我的意思是，MikroC编译代码没有错误，但是在proteus模拟器上出现错误“堆栈下溢执行RETLW指令”。即使代码有大约3500个程序字，在ROM和RAM中也有足够的空间。该项目由PIC16F648A和16x2液晶显示器组成。它的功能之一是一个电动生态车原型的速度计。所以我必须处理一些“long int变量”和“float变量”。我一直使用MikroC的库函数将float变量转换为string（称为FloatToStr），以便在LCD显示器上显示它。即使浮标变量被给定为常数（如2.5），Proteus模拟器也产生相同的误差。另一方面，如果我创建一个程序，只使用浮点变量来计算一个长的数学表达式，并使用相同的函数来转换它，Proteus模拟器上没有错误。我的问题。所以，如果有人能帮助我，我很感激。我没有发布代码的任何部分，因为它太大了，即使使用常数也不能工作。所以，我认为问题不在代码上，但可能存在一些与程序大小有关的问题。毕竟，如果在完全专用于此的程序中使用相同的计算和转换函数，则没有错误。



      以下为原文

    Hello,

I've been having problems to work with math expression using 'float'. I mean, the MikroC compiles the code without errors, but on the Proteus simulator occurs an error "Stack underflow executing RETLW instruction". Even the code has around 3500 program words, there are enough space in ROM and RAM.

The project consists of PIC16F648A and an 16x2 LCD display. One of its functions is be a speedometer for a prototype of an electrical ecologic car. So I have to work with some 'long int variables' and 'float variables'.

I've been using an library function of MikroC to convert a float variable to string (called FloatToStr) in order to show it in LCD display. Even though the float variable is given as a constant (like 2.5), the Proteus simulator generates the same error. On the other hand, if I create a program only to calculate a long math expression using float variables and use this same function to convert it, there is NO error on Proteus simulator.

I had looked for solutions on the internet and books for a long time before I created this post, but I hadn't solved my problem. So, I appreciate if someone can help me

PS.: I didn't post any part of the code because it's so big and it isn't working even using a constant. So, I believe that the problem isn't on the code, but maybe there is some issue related with the size of the program. After all, if I use the same calculation and convert function in a program totally dedicated for this, there is no error.

Simas,