[Reportbug-maint] Bug#848729: fails with UnicodeDecodeError: 'utf-8' codec can't decode byte 0xa3 in position 4341: invalid start byte

[Sven Hartge]

Hi!

I want to add some other observations to this bug. reportbug seems to
die during the parsing of /var/lib/dpkg/status, hitting some incorrectly
encoded characters. But if I launch reportbug with

   LC_ALL=de_DE at euro reportbug netdata

switching to the old 8bit locale, I am able to report a bug against netdata.

| UnicodeDecodeError: 'utf-8' codec can't decode byte 0xa3 in position
5333: invalid start byte

The character "0xa3" would be the "£" sign in ISO-8859-1

I am seeing this character on i386 in the following package:


Package: libfsplib0
Status: install ok installed
Priority: optional
Section: libs
Installed-Size: 76
Maintainer: £Ø­Ù<85>د اÙ<84>Ù<85>Ø­Ù<85>Ù<88>دÙ<8a> (Ahmed
El-Mahmoudy) <aelmahmoudy at sabily.org>
Architecture: i386
Source: fsplib
Version: 0.11-2
Depends: libc6 (>= 2.2)

There is another package from Ahmed El-Mahmoudy in the file, but here
his name is encoded differently:

Package: libharfbuzz0b
Status: install ok installed
Priority: optional
Section: libs
Installed-Size: 985
Maintainer: أحÙ<85>د اÙ<84>Ù<85>Ø­Ù<85>Ù<88>دÙ<8a> (Ahmed
El-Mahmoudy) <aelmahmoudy at users.sourceforge.net>
Architecture: i386
Multi-Arch: same
Source: harfbuzz (1.2.7-1)
Version: 1.2.7-1+b1


Looking at the status file with "od -t cx1 /var/lib/dpkg/status" shows
the following byte sequences:

a) wrong:

243 330 255 331 205 330 257
a3  d8  ad  d9  85  d8  af

b) right:
330 243 330 255 331 205 330 257
d8  a3  d8  ad  d9  85  d8  af

So far I only see this problem in my i386 Sid system, the files on the
amd64 systems seem correct and reportbug does not show any errors there.

Grüße,
Sven.