Wireshark-bugs: [Wireshark-bugs] [Bug 12845] First start with non-empty extcap folder after inst
Comment # 24
on bug 12845
from Roland Knall
(In reply to Guy Harris from comment #23)
> (In reply to Pavel Sindelka from comment #20)
> > It seems even English-speaking Windows are not a guarantee that the names
> > will be pure ASCII. The other guy of the two has provided his
> > --extcap-config output (see the full dump below), and there are two items
> > using non-ASCII characters, one per each root hub:
> >
> > value {arg=99}{value=1_7}{display=Jeppe SchoubyeÔÇÖs Mouse
> > #1}{enabled=false}{parent=1_3}
> > value {arg=99}{value=9_1}{display=Dell Wireless 5809e GobiÔäó 4G LTE Mobile
> > Broadband Card}{enabled=false}{parent=9}
>
> There's a bit of mojibake:
>
> https://en.wikipedia.org/wiki/Mojibake
>
> there - is this some issue where a string in some "ASCII" code page is being
> treated as UTF-8 or *vice versa*?
Sadly no, we where able to reproduce the issue with plain-english texts. Has
nothing to do with character coding. The real issue here is that Windows has a
far shorter limit on the stdout buffer in comparison to Linux or OS/X. All
three systems have the same issues, but Windows strikes first.
This did not happen in the past, because we used g_spawn_sync, which handles
that internally. We will have to change the _sync call to an _async call and
handle the buffer ourselves. I am working on a fix, but as Windows is not my
operating system of choice, it takes me a while.
You are receiving this mail because:
- You are watching all bug changes.
