[Reportbug-maint] Bug#857794: reportbug: crash when encountering some non-ASCII characters

[Dima Kogan]

On March 18, 2017 12:24:12 PM PDT, Evgeni Golov <evgeni at debian.org> wrote:
>Hi Dima,
>
>> Hi. I just tried to send an unblock, and reportbug crashed. Session:
>> 
>>     dima at scrawny:~$ sudo apt install reportbug
>>     ...
>>     Please enter the name of the package: geda-gaf
>>     Checking status database...
>>     Traceback (most recent call last):
>…
>>         return codecs.ascii_decode(input, self.errors)[0]
>>     UnicodeDecodeError: 'ascii' codec can't decode byte 0xd8 in
>position 327: ordinal not in range(128)
>
>The package has "أحمد المحمودي (Ahmed El-Mahmoudy)
><aelmahmoudy at users.sourceforge.net>" as uploader.
>And that is not parsable by your non Unicode environment.
>
>> Presumably there's some non-ascii something somewhere that's tripping
>it
>> up. It is 100% reproducible, and as you can see I'm using the latest
>> available reportbug. Thanks.
>
>I can reproduce it with LC_ALL=C, but not with C.UTF-8.
>
>I would argue that reportbug can't do anything here, as your
>environment just can't display the string correctly.

Hi. Reportbug can do some set of these:

1. Tell me what is wrong and how I can make it work instead of throwing an opaque exception. I never did report that bug

2. Turn on whatever setting it needs. Is it just an incompatible environment setting? If I set LC_ALL to C.UTF-8 then this will magically work? What does this setting mean? Can reportbug set this?

3. If reportbug can't display some unicode string, can it simply not display it, or display an ascii version of it (this was available here)? The bug report it generates can still contain the unicode.

Any of these would be an improvement. Thanks