[Webkit-unassigned] [Bug 193571] [GTK][WPE] Add API to add paths to sandbox

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=193571

Michael Catanzaro <mcatanzaro at igalia.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |youennf at gmail.com

--- Comment #10 from Michael Catanzaro <mcatanzaro at igalia.com> ---
(In reply to Patrick Griffis from comment #9)
> > Maybe we should attempt a g_mkdir_with_parents(), so we don't have to worry
> > about this if the directory is writable? Bad idea?
> >
> > We could add a flags parameter to control the behavior?
> 
> Well its blocking IO...

Before the web context is initialized, though. This is basically the only time it's OK! I just wonder whether applications will be more annoyed that we create directories for them, or more annoyed when they break because the directory just happens to not exist early enough.

> > Why are you doing this instead of just adding a second HashMap to struct
> > LaunchOptions? Would be much simpler, right?
> 
> Because Youenn was very strongly against adding things to LaunchOptions last
> review since an extra data field already existed... I just avoided touching
> it from the start but I agree just adding it there would be much cleaner.

Oh, I forgot about that, sorry.

Adding platform-specific members to cross-platform structures is good to avoid, but we do it when needed. Here it really makes sense, because the alternative you've implemented -- serializing what should be a HashMap into a string and then manually deserializing it -- isn't great. Youenn? (Anyway, if that's what Youenn wants, it's what we'll do.)

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20190121/668d8758/attachment.html>