[PATCH v4 2/8] reg.exe: Add path_get_key() to remove boilerplate

[Jonathan Vollebregt]

> Right, I didn't keep that in mind. Still there's the question of what
> should happen if the user passes a key path like
> HKEY_LOCAL_MACHINE_IBREAKYOU\Software\Wine .

>> Assuming you expect errors using RegCloseKey on a predefined key
>> like HCKU: In practice it works just fine. I imagine it's supposed
>> to work this way since testbot doesn't throw any errors either.
> According to the RegCloseKey implementation in
> dlls/advapi32/registry.c you are correct. I can't find any tests for
> this behavior though. A comment may be in order if you intentionally
> relying on this behavior.

Fixed and commented:
> https://github.com/jnvsor/wine/commit/1f6012dcee7ee58e7206c1a6510cc77972755508

> This
> patch is easier to read, but I am still wondering what the
>
>> +    if (key[0] == '\\') +        key++;
>
> and
>
>> +    i = strlenW(key) - 1; +    while (i >= 0 && key[i] == '\\') +
>> key[i--] = 0;
>
> parts are for. What do leading backslashes in a key do if they don't
> specify a remote host? Can there really be any number of trailing
> backslashes? What about trailing spaces like "
> \\remote_host\HKLM\foo\bar"?

The examples at the following URL indicate that even if a single leading 
backslash isn't a valid key name, it's a valid reg.exe input.
> https://www.microsoft.com/resources/documentation/windows/xp/all/proddocs/en-us/reg.mspx?mfr=true

Trailing spaces are valid key names (At least in wine, I imagine in 
native too)

> There are plenty of things that need additional tests.

The closest thing I have to a native windows system is testbot and I 
doubt you'd want me clogging it up just to find out what the proper reg 
inputs are.

If anyone can write tests for the rest, that would be great!