[Bug 47762] New: LoadStringA fails instead of returning const pointer to string if buffer length is zero

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=47762

Bug ID: 47762
           Summary: LoadStringA fails instead of returning const pointer
                    to string if buffer length is zero
           Product: Wine
           Version: unspecified
          Hardware: x86
                OS: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: user32
          Assignee: [e-mail]
          Reporter: [e-mail]
      Distribution: ---

https://docs.microsoft.com/en-us/windows/win32/api/winuser/nf-winuser-loadst...

According to MSDN doc, LoadStringA should be able to return const pointer
without using allocated buffer if the buffer size is set to zero.

Here we see that Wine just returns -1 in that case:

https://source.winehq.org/git/wine.git/blob/refs/heads/master:/dlls/user32/r...

Since there is something that looks legit in the LoadStringW just above, it
shouldn't be too hard to fix, amirite?

[Bug 47762] LoadStringA fails instead of returning const pointer to string if buffer length is zero

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=47762

Zhiyi Zhang  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #1 from Zhiyi Zhang  ---
Some tests would be required to fix this. For example, what happens when
Unicode/ASCII string resources are loaded by LoadStringA/W?

[Bug 47762] LoadStringA fails instead of returning const pointer to string if buffer length is zero

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=47762

--- Comment #2 from Alexandre Julliard  ---
String resources are stored in Unicode, so only LoadStringW is able to return a
direct pointer. That's how it works on Windows. Is there an actual app that
needs a different behavior?

[Bug 47762] LoadStringA fails instead of returning const pointer to string if buffer length is zero

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=47762

--- Comment #3 from Stanislav  ---
(In reply to Alexandre Julliard from comment #2)
> String resources are stored in Unicode, so only LoadStringW is able to
> return a direct pointer. That's how it works on Windows.

Ok, now I see. There is some kind of fixed-length Unicode and not UTF-8 used in
the compiled resources so we cannot just take ANSI-only Unicode string and say
it's a valid ASCII one.

But then how does Windows do it? Converting string to ASCII in some kind of
implicit temporary buffer? Telling resource compiler that it should pack string
data in non-Unicode mode if we will be using A-suffixed functions?

I mean, MS doc says black on white that LoadStringA should somehow work with
zero buflen, so this one probably shouldn't be completely ignored even if I'm
misunderstanding something.
> Is there an actual app that needs a different behavior?

No, not really. In fact, I've been just making one as my uni assignment and
stumbled upon this issue.

[Bug 47762] LoadStringA fails instead of returning const pointer to string if buffer length is zero

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=47762

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |INVALID
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #4 from Alexandre Julliard  ---
I see no evidence that Windows does this, and our existing tests don't show it
either. Unless you can write a test case that shows differently, I'd just
assume that MSDN is incorrect, it happens fairly often.

[Bug 47762] LoadStringA fails instead of returning const pointer to string if buffer length is zero

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=47762

Stanislav  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|INVALID                     |NOTOURBUG

--- Comment #5 from Stanislav  ---
(In reply to Alexandre Julliard from comment #4)
> I'd just assume that MSDN is incorrect, it happens fairly often.

Wow, MSDN sure sucks. I actually found a fellow with visual studio and checked
out LoadStringA on his PC, it behaves precisely as the Wine version.

Thanks for pointing it out, case closed.

[Bug 47762] LoadStringA fails instead of returning const pointer to string if buffer length is zero

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=47762

Zebediah Figura  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|NOTOURBUG                   |INVALID
             Status|RESOLVED                    |CLOSED
                 CC|                            |[e-mail]

--- Comment #6 from Zebediah Figura  ---
Closing invalid. NOTOURBUG is for bugs in upstream components that Wine depends
on.