幻影断点

更好地打破一个NOP，避免任何代码行可能转化为胸罩，GTO，呼叫或返回。



      以下为原文

    Better to break on a NOP and avoid any line of code that might translate into a BRA, GOTO, CALL or RETURN. 

1）这被称为断点集结，这正是8位处理器工作的方式。经验丰富的8位工程师知道尽早设置断点一行，或插入一个NOP，并打破那里作为咖啡评论家建议。注意，SKID是一种汇编语言指令，所以当处理C代码时，它可能是棘手的，因为大多数C代码行生成不止一个汇编代码。2）为了在C代码行上设置断点，该C代码行必须实际生成汇编操作码。各种C行实际上不生成程序集（例如空白或注释行、{}s、变量声明等）。3）不知道那里发生了什么。



      以下为原文

    1)  This is called breakpoint skidding and is just the way the 8 bit processors work.  Experienced 8 bit engineers know to set the breakpoint one line early, or to insert a NOP and break there as coffee_critic suggests.  Note that skid is one assembly language instruction, so it can be tricky when dealing with C code as most C lines of code generate more than one assembly code.
 
2) In order to set a breakpoint on a line of C code, that line of C code must actually generate assembly opcodes.  All sorts of C lines don't actually generate assembly (e.g. blank or comment lines, {}s, variable declarations etc.).
 
3)  No idea what's going on there. 

这是PIC硬件行为的方式，被称为断点打滑：http://MyCHIPPrdult.COM/TLS0201：一个8位PIC中的SkID Effice断点，是一个寄存器，观察程序指针，在地址匹配时生成调试中断，但是指令已经执行。因此，调试器只在下一个指令之前获得控制。这可能位于调用的函数内，或者在跳转的目标中。当准备调试一些复杂代码时，可以插入Nop-（）；宏指令如咖啡评论家所解释的。一些代码行不创建任何指令。NS。如果没有与一行连接的指令地址，则调试器将无法设置断点。同样，您可以插入NOP（）；指令宏。迈西尔



      以下为原文

    It is the way the PIC hardware behave, and have been called breakpoint skidding:
http://microchipdeveloper.com/tls0201:skid-effect
 
Breakpoint in a 8 bit PIC, is a register watching the program pointer,
and generating a Debug interrupt when the address match.
But then the instruction is already executing, so debugger get control only before the next instruction.
This may well be inside a function beeing called, or at the target of a jump. 
 
When preparing to debug some complicated code, you may insert NOP(); macro instructions as explained by coffee critic.
Some code lines do not create any instructions. If there is no instruction address connected with a line, then debugger will not be able to set the  breakpoint. Again, you may insert NOP(); instruction macros.
 
   Mysil

自7版以来，甚至在早期版本中，MPLAB都出现了幻像断点。就像它们出现的神秘一样，它们消失了。一种加速消失的方法是关闭MPLAB并重新打开它。



      以下为原文

    Phantom breakpoints have occurred in MPLAB since version 7 and possibly even in earlier versions.  As mysteriously as they appear, they disappear.  One way to speed their disappearance is to close MPLAB and reopen it.

同意。



      以下为原文

    agreed.

这是仿真器还是真正的调试工具？模拟器不应该滑行。我使用MPLAB ICE调试（汇编）代码，它有一个选项，消除了被称为“软件断点”的打滑。我使用PIC24/DSPIC33部分来完成我们的所有项目，所以我从未尝试过用MPLAB ICE调试8位的部分。我不知道这个选项是否适用于8位部件。我讨厌打滑，当我想出这个选择时，我很高兴。



      以下为原文

    Is this in the simulator or with a real debugging tool?  The simulator should never skid.  I use MPLAB ICE for debugging (assembly) code and it has an option that eliminates the skidding called "software breakpoints".  I use PIC24/dsPIC33 parts for all of our projects, so I have never tried debugging 8 bit parts with MPLAB ICE.  I don't know if that option is available for 8 bit parts.  I hated the skidding, and was happy when I figured this option out.
 

就像FIY一样。我切换了ToMPLAB-X版本3.61，这是我在更新之前使用的版本。（1）打滑仍在进行中）（3）断点现在正确地工作）。



      以下为原文

    Just as an FYI.  I switched to MPLAB-X version 3.61  This is the version I was using before updating.  Now
 
1) The skidding still occurs
2) The step over now works correctly
3) The broken breakpoints now works correctly
4) The phantom breakpoints are gone
 
Now debugging is at least usable.