August 2024 - Wine-gitlab - list.winehq.org

Re: [PATCH v8 0/3] MR6130: shell32: Added stubs for the EnumerableObjectCollection COM class.
by Elizabeth Figura (＠zfigura) 14 Aug '24

14 Aug '24

On Wed Aug 14 17:21:18 2024 +0000, Alfred Agrell wrote: > - Simply implement EnumerableObjectCollection and IEnumObjects > - Add IObjectCollection > - Add remoting support (and run into that WIDL bug while we're at it) > - Move it to windows.storage.dll (Wine doesn't have that DLL yet > (closest match is windows.storage.applicationdata.dll), so that's yet > another pile of copypasted scaffolding) > The amount of scope creep in this MR is somewhere between concerning and > hilarious. I vote keep it in shell32 until and unless something cares. I don't think the last two are scope creep. Committing this without proper remoting support is simply broken, and putting the DLL in the right place is also a matter of correctness. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6130#note_78717

1 0

Re: [PATCH v8 0/3] MR6130: shell32: Added stubs for the EnumerableObjectCollection COM class.
by Elizabeth Figura (＠zfigura) 14 Aug '24

14 Aug '24

On Wed Aug 14 17:21:18 2024 +0000, Alfred Agrell wrote: > I can't see anything obviously wrong in v8 3/3(*). Do you want 3/3 > merged into 1/3? > And yeah, that looks strange. Not exactly wrong - S_OK means it fetched > as many items as it asked for, so that'll just write the value that's > already there (unless the Next impl is buggy or something) - but > certainly strange. Yeah, better remove it until and unless proven necessary. > (*) actually 3/4, but that last one's a merge commit so let's pretend it > doesn't exist. Rebase and remove it. Oh, right, I misremember, S_FALSE is returned if the count isn't equal. So this isn't really that important to fix. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6130#note_78716

1 0

Re: [PATCH v8 0/3] MR6130: shell32: Added stubs for the EnumerableObjectCollection COM class.
by Alfred Agrell (＠Alcaro) 14 Aug '24

14 Aug '24

On Wed Aug 14 17:21:18 2024 +0000, Elizabeth Figura wrote: > 3/3 still needs to be fixed up to 1/3 where the class is introduced, > including the actxprxy parts. > There's also a whitespace error introduced in 2/3. > Also, I'm just now noticing... > > ``` > > +HRESULT __RPC_STUB IEnumObjects_Next_Stub(IEnumObjects *This, ULONG > celt, REFIID riid, void **rgelt, ULONG *pceltFetched) > > +{ > > + HRESULT hr; > > + TRACE("(%p)->(%ld, %p, %p, %p)\n", This, celt, > debugstr_guid(riid), rgelt, pceltFetched); > > + *pceltFetched = 0; > > + hr = IEnumObjects_Next(This, celt, riid, rgelt, pceltFetched); > > + if (hr == S_OK) *pceltFetched = celt; > > + return hr; > > +} > > ``` > The "*pceltFetched = celt" line looks very wrong, and I don't think that > line should be necessary at all. It was copied from IEnumShellItems, but > that also looks wrong. I can't see anything obviously wrong in v8 3/3(*). Do you want 3/3 merged into 1/3? And yeah, that looks strange. Not exactly wrong - S_OK means it fetched as many items as it asked for, so that'll just write the value that's already there (unless the Next impl is buggy or something) - but certainly strange. Yeah, better remove it until and unless proven necessary. (*) actually 3/4, but that last one's a merge commit so let's pretend it doesn't exist. Rebase and remove it. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6130#note_78711

1 0

Re: [PATCH v8 0/3] MR6130: shell32: Added stubs for the EnumerableObjectCollection COM class.
by Alfred Agrell (＠Alcaro) 14 Aug '24

14 Aug '24

On Wed Aug 14 17:21:18 2024 +0000, Elizabeth Figura wrote: > > Also, what version of Windows are you seeing vend this class from > shell32? My Windows 10 vends it from windows.storage.dll. > FWIW, I do see this vended from shell32 in Windows 7. It's probably > better to match more recent Windows here, though I'm not sure if it's > worth caring. - Simply implement EnumerableObjectCollection and IEnumObjects - Add IObjectCollection - Add remoting support (and run into that WIDL bug while we're at it) - Move it to windows.storage.dll (Wine doesn't have that DLL yet (closest match is windows.storage.applicationdata.dll), so that's yet another pile of copypasted scaffolding) The amount of scope creep in this MR is somewhere between concerning and hilarious. I vote keep it in shell32 until and unless something cares. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6130#note_78712

1 0

[PATCH v4 0/5] MR6260: mshtml: Misc improvements and fixes.
by Gabriel Ivăncescu 14 Aug '24

14 Aug '24

-- v4: mshtml/tests: Accept rare return value from ReportResult on native. mshtml: Don't process tasks in message loop during parse complete mshtml: Implement EmulateIE* modes for X-UA-Compatible. mshtml: Don't return default ports from location.host in IE10+ modes. mshtml: Avoid calling remove_target_tasks needlessly. https://gitlab.winehq.org/wine/wine/-/merge_requests/6260

2 8

Re: [PATCH v8 0/3] MR6130: shell32: Added stubs for the EnumerableObjectCollection COM class.
by Elizabeth Figura (＠zfigura) 14 Aug '24

14 Aug '24

On Wed Aug 14 16:29:53 2024 +0000, Kevin Martinez wrote: > Ok, was just following the example of other entries. Yeah, it's unfortunately one of those things that's been cargo-culted, but I think it's better to avoid it going forward. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6130#note_78697

1 0

Re: [PATCH v8 0/3] MR6130: shell32: Added stubs for the EnumerableObjectCollection COM class.
by Elizabeth Figura (＠zfigura) 14 Aug '24

14 Aug '24

> Also, what version of Windows are you seeing vend this class from shell32? My Windows 10 vends it from windows.storage.dll. FWIW, I do see this vended from shell32 in Windows 7. It's probably better to match more recent Windows here, though I'm not sure if it's worth caring. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6130#note_78695

1 0

Re: [PATCH v8 0/3] MR6130: shell32: Added stubs for the EnumerableObjectCollection COM class.
by Elizabeth Figura (＠zfigura) 14 Aug '24

14 Aug '24

3/3 still needs to be fixed up to 1/3 where the class is introduced, including the actxprxy parts. There's also a whitespace error introduced in 2/3. Also, I'm just now noticing... > ``` > +HRESULT __RPC_STUB IEnumObjects_Next_Stub(IEnumObjects *This, ULONG celt, REFIID riid, void **rgelt, ULONG *pceltFetched) > +{ > + HRESULT hr; > + TRACE("(%p)->(%ld, %p, %p, %p)\n", This, celt, debugstr_guid(riid), rgelt, pceltFetched); > + *pceltFetched = 0; > + hr = IEnumObjects_Next(This, celt, riid, rgelt, pceltFetched); > + if (hr == S_OK) *pceltFetched = celt; > + return hr; > +} > ``` The "*pceltFetched = celt" line looks very wrong, and I don't think that line should be necessary at all. It was copied from IEnumShellItems, but that also looks wrong. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6130#note_78694

1 0

[PATCH v2 0/3] MR6276: winebus.sys: Improve udev device information parsing.
by Rémi Bernon 14 Aug '24

14 Aug '24

Supersedes https://gitlab.winehq.org/wine/wine/-/merge_requests/6268. -- v2: winebus.sys: Read vendor/product/serial strings from the usb subsystem. winebus.sys: Always parse uevent PRODUCT= line on input subsystem. winebus.sys: Add devtype parameter to get_device_subsystem_info(). https://gitlab.winehq.org/wine/wine/-/merge_requests/6276

3 4

[PATCH 0/3] MR6276: winebus.sys: Improve udev device information parsing.
by Rémi Bernon 14 Aug '24

14 Aug '24

Supersedes https://gitlab.winehq.org/wine/wine/-/merge_requests/6268. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6276

4 10