leo2moko's str2ind

[Das Signal]

Hi all,

As you may know, TI's toolchain, when building a firmware,
usually replaces the original strings with integers, and places
the strings in a file called "str2ind.tab". This has the
advantage of saving flash space, however without this strings
file the trace output is much less useful.

Recently I've made some progress on a Python script that
automates the subtitution of integers with the corresponding
strings in a binary firmware (such as Compal's); basically
it matches the functions, either with the .map files, or with
the bindiff tool. This so far has worked pretty well, and
the script is able to recover about 50% of the strings.

I've noted however some small inconsistencies, for example
in pb_read_sim_record() calls vso_i_func_itrace, with its
name as first argument; however in the disassembled leo2moko
the particular integer, 851, leads to a completely different
string in str2ind.tab. So I'm thinking something is wrong
somewhere, perhaps someone can shed some light on this?

--DS