Your message dated Tue, 10 Jun 2008 10:17:54 +0100
with message-id <[EMAIL PROTECTED]>
and subject line Re: Bug#485506: emdebian-tools: Should not prompt user in bug
script
has caused the Debian Bug report #485506,
regarding emdebian-tools: Should not prompt user in bug script
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [EMAIL PROTECTED]
immediately.)
--
485506: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=485506
Debian Bug Tracking System
Contact [EMAIL PROTECTED] with problems
--- Begin Message ---
Package: emdebian-tools
Version: 1.1.0
Hi,
packages should not in general ask the user for feedback in bug scripts.
This is both disruptive to the user and hard to support by graphical
frontends.
Please consider using a warning in the presubj instead, the user can
then review the included stuff in their editor.
thanks,
Michael
--- End Message ---
--- Begin Message ---
Package: emdebian-tools
Version: 1.1.0
On Tue, 2008-06-10 at 00:14 +0200, Michael Banck wrote:
> Hi,
>
> packages should not in general ask the user for feedback in bug scripts.
? is that documented somewhere ?
I see lots of prompting when filing bugs.
yesno is explicitly implemented as an internal function inside reportbug
itself. I see no reason why the reportbug implementation cannot be
reimplemented when the bug script is read in by some other handler.
It isn't as if I'm forcing a prompt using a command line utility, I call
a reportbug function.
Feel free to check the reportbug source code - I see no prompt code in
the emdebian-tools bug script.
> This is both disruptive to the user and hard to support by graphical
> frontends.
I initially implemented the questions after viewing other reportbug
scripts on my own system and I've implemented the same question format
in each of my reportbug scripts. (At least on this system, other
packages doing the same are exim4*, hibernate, installation-report and
apt (upon which I based my own bug scripts because of the similarity of
the data requested)).
Of the scripts that don't use 'yesno', others still use getkey (texlive*
and libkpathsea4). Again, getkey is defined within the reportbug
handle_bugscript handler.
This isn't a debconf situation, prompts don't halt an installation and
reporting bugs is a long sequence of prompts.
The current implementation may be a little awkward to handle in a GUI
(albeit not impossible) but maybe the "asking questions" bit isn't
actually the problem and any sane bug reporting tool should simply use
the defacto standard way of implementing the reportbug shell functions
so that the user is still asked before possibly sensitive/confidential
data is included in a public bug report.
Banning prompting in bug scripts is, IMHO, madness.
> Please consider using a warning in the presubj instead, the user can
> then review the included stuff in their editor.
I think a question mechanism should still be supported in any handler of
the reportbug scripts.
The emdebian-tools bug scripts (and all my other bug scripts) do *NOT*
require prompting, they merely call a reportbug function that
*IMPLEMENTS* prompting.
There is nothing to say that a different handler cannot reimplement the
yesno and getkey functions in a different handler.
Closing the bug report because the emdebian-tools bug script does not
actually *do* any prompting, it merely calls a reportbug function which
currently implements the function using a prompt. (Not reassigning
because prompting in reportbug is clearly not a bug itself.)
--
Neil Williams
=============
[EMAIL PROTECTED]
signature.asc
Description: This is a digitally signed message part
--- End Message ---
