On Sun, 15 Nov 2026 19:08:58 +0100,
Klemens Nanni <[email protected]> wrote:
> 
> 15.11.2026 17:59, Kirill A. Korinsky пишет:
> > On Sat, 14 Nov 2026 19:21:49 +0100,
> > Nazarenko Mykyta <[email protected]> wrote:
> >>
> >> +@@ -11,9 +11,9 @@
> >> + namespace Webview::WebKitGTK::Library {
> >> + 
> >> + ResolveResult Resolve(Platform platform) {
> >> +- const auto lib = base::Platform::LoadLibrary("libwebkitgtk-6.0.so.4", 
> >> RTLD_NODELETE)
> >>  -         ?: base::Platform::LoadLibrary("libwebkit2gtk-4.1.so.0", 
> >> RTLD_NODELETE)
> >>  -         ?: base::Platform::LoadLibrary("libwebkit2gtk-4.0.so.37", 
> >> RTLD_NODELETE);
> >> ++ const auto lib = base::Platform::LoadLibrary("libwebkitgtk-6.0.so", 
> >> RTLD_NODELETE)
> >>  +         ?: base::Platform::LoadLibrary("libwebkit2gtk-4.1.so", 
> >> RTLD_NODELETE)
> >>  +         ?: base::Platform::LoadLibrary("libwebkit2gtk-4.0.so", 
> >> RTLD_NODELETE);
> >>    const auto result = lib
> >>            && LOAD_LIBRARY_SYMBOL(lib, gtk_init_check)
> >> -          && LOAD_LIBRARY_SYMBOL(lib, gtk_widget_get_type)
> >> +          && LOAD_LIBRARY_SYMBOL(lib, gdk_set_allowed_backends)
> > 
> > Just spoiled it: we have installed /usr/local/lib/libwebkitgtk-6.0.so.0.7
> > probably base::Platform::LoadLibrary("libwebkitgtk-6.0.so", RTLD_NODELETE)
> > loads it, but I have no idea that tdesktop makes as LoadLibrary().
> 
> This library may not be used on OpenBSD, I don't remember.
> 
> The patch is part of a sweep over this pattern of loading libs from when one
> of them caused tdesktop to die on startup as it couldn't find the wrong name.
>

It can be used for something like show 3Ds page when you buy telegram
premium, or similar use cases.

I not sure, just guess.

> > 
> > Anyway, it builds here with new devel/wallet-engine and tdesktop works, I
> > had tested it briefly and nothing to report so far.
> 
> I'm currently wondering whether we should hold off with tdesktop 7.4.x
> and stick to 7.3.x where quickjs is not yet used.
> 
> Apparently, folks keep finding holes in the new javascript code...
>

Well, we can hold 7.3.x for a while, but here the catch: we have no
information how many holes are unpatched.

We may move forward with updates, and so far it reads that majority of them
was fixed.

-- 
wbr, Kirill

Reply via email to