https://bugs.documentfoundation.org/show_bug.cgi?id=166128
--- Comment #8 from Michael Weghorn <[email protected]> --- (In reply to Hossein from comment #6) > Sometimes it becomes important, when the splash screen becomes too big, or > too small. I have seen both cases. If that's a problem, maybe just using a fixed size could be a way forward? As far as I remember, it always showed in the same size for me, but maybe something can go wrong in the size calculation. (By now, it lives in the ctor of the SplashScreenWindow class.) > At least part of the UI is from picture, and that is not localize-able. I > see "Development Build". This comes from something like > icon-themes/colibre/brand_dev/intro-highres.png vs > icon-themes/colibre/brand/intro.png True, if there is no way to use localized icons, that can be a problem. > The oosplash executable > > still exists and is used on Linux during startup, even if it's not > > responsible for the splash screen any more. > That is a good progress! Thanks for that. What are the reasons to keep that? > If it is because of javaldx, I think that is one source of slowness in > startup, which can probably moved to some step after startup. javaldx is indeed there and some other quite low-level logic that I didn't dig into that much. Maybe some of it is obsolete/no longer needed by now, would need a closer look. If you're interested: The implementation is in `SAL_IMPLEMENT_MAIN_WITH_ARGS` in desktop/unx/source/start.cxx . (In reply to Buovjaga from comment #7) > Re: removing the splash, do we know what are some things that delay startup? > Thousands of fonts on the system? I don't know. To see what happens during the phase that the splash screen is shown: The implementation is in Desktop::Main (in desktop/source/app/app.cxx). Splash screen starts here: // Startup screen OpenSplashScreen(); (I see now that e.g. also the mar updater runs on Windows in that phase, which possibly might take a while?) -- You are receiving this mail because: You are the assignee for the bug.
