Re: [hatari-devel] Symbols visibility in debugger code

[Thorsten Otto <admin@xxxxxxxxxxx>]

On Sonntag, 18. Juli 2021 13:24:32 CEST Troed Sångberg wrote:

> Apologies, but I do think your example is an edge case

 

Unfortunately it isn't. The first instruction might just set up a loop counter, the 2nd loads the address of a table or something.

 

 

>you do not know which subroutine (for example) you're right now deciding to >step into or run through.

 

That's what your heuristic would be about ;) For jsr, bsr etc. you can assume that the address is some program address (and will most likely be some symbol). But this not always the case, move.l #<something>,Dx might be a symbol, or something else. Same for other instructions, like sub.l. So in general, the disassembler cannot always know what to use. I think the best is to let the user decide what to display, via an option.