[Eric] [eric5.0] a few bugs, possibly

[Angus]

<Forgot to set the destination email address to this list>

>> Settings > Preferences > Editor > General > Tab width
>> Settings > Preferences > Editor > Edge Mode > Column Number
>> Are devided by 2 in the applied output (e.g. setting a tab width of 4
>> creates an actual tab width of 2 in the editor).
>> I'm using the standard Python3 encoding (without declaration): utf-8.
>
> That should not happen. If these value are correct in the eric5.ini file, it
> is a QScintilla issue. Maybe this is related to using a non monospaced font.

Actually this happens with the "Use monospace font as default"
selected in the settings.

>>> *) Py3Flakes doesn't understand the 'with' statement even though
>>> that's recommended way of handling files according to the official
>>>
>>> Python tutorial:
>>> >>>with open('foo', 'r') as f:
>>> >>>    l = f.readlines()
>>>
>> >causes: "py3flakes warning: Undefined name 'f'."
>>
>> That is a bug and should be fixed.
>
>That issue will be fixed in the next release/snapshot.

Nice.

>> *) This one is a bit hard to describe, but the only one giving me
>> grief. It happens in various widgets containing text, but I'll just
>> try to give one specific example of the treeview in Settings >
>> Preferences:
>> Click the triangle next to an item (e.g. "Debugger"). Often there will
>> be no response. Nor will the second click. But then, moving the
>> mousing pointer down will cause the screen to update that area. That
>> is, the positioning of text will change "on the fly", which looks like
>> the screen is "tearing" wherever you move the mouse pointer to. This
>> also regularly happens in the help viewer when reading documentation,
>> but I've trouble reproducing it consistently (some combination of
>> scrolling, selecting text and/or moving the mouse pointer).
>
> I have seen something similar on Windows systems. However, I doubt that it is
> an eric issue but rather Qt or PyQt.

Forgot to mention I'm only using Gnome in Arch Linux. I'm also
guessing it's most likely a Qt bug. I'll try to investigate more... I
love Qt. Shame it's a bit buggy.

>> *) Finally, I consider it a bug to have no public bug tracker (or if
>> there is one, it's too hard to find). Having one would it make it
>> easier for the developer to keep track of open issues. Furthermore, it
>> would make it a lot easier for users to find out what issues are
>> already known to the developer, saving wasted energy on redundant
>> reports by users and replies by the developer.
>
> It is a bug but probably one that won't get fixed in the near future. Doing
> eric mostly as a one man project doesn't leave me time to manage a bug
> tracker. Bugs should be reported via the built-in bug reporting feature or
> through this mailing list.

I think a bug tracker would actually save you time though. Managing
one is pretty fast and easy, and the effort would be made up for by
having fewer messages to reply to that report/ask the same thing (and
to those failing to notice the tracker you could simply reply with an
url pointing to the issue). But of course that's just my POV and you
should use whatever method works best for you (I'm impressed that you
code and maintain all of eric yourself anyway).

BTW, just encountered another bug, albeit just a little obscure edge
case. In the editor, at the *end of the last line* of the text, if you
type an "opening character" (e.g. '[') the corresponding closing
character (e.g. ']') appears automatically as it should, but if you
then delete/backspace that closing character and then retype it
manually, this warning message appears: Warning: An unhandled
exception occurred. Please report [...] <class 'IndexError'>: string
index out of range --- File
"/usr/lib/python3.1/site-packages/eric5/QScintilla/TypingCompleters/CompleterPython.py",
line 126, in charAdded --- if char == self.editor.text(line)[col]:

Regards,
Ang.