[Users] [Bug 3889] address and quoted message inconsistent in reply

[Michal Suchánek]

On Tue, 18 Sep 2018 12:17:39 -0000
Paul <claws at thewildbeast.co.uk> wrote:

> On Tue, 18 Sep 2018 13:50:15 +0200
> Michal Suchánek <msuchanek at suse.de> wrote: 
> 
> > AFAICT claws does not have per-list reply settings so I use
> > reply-all all the time which works with all lists.
> 
> See /Comfiguration/Preferences/Compose/Writing, 'Reply invokes
> mailing list reply'.

What do you not understand about per-list? That's global setting, not
per-list. Some lists work like that, some do not.

> 
> See /Message/Reply to/Mailing List.
> 
> See /Message/Mailing List/...

So claws already knows what mailing list the message belongs to. If it
can set default reply option based on that it might work better.
However, it seems you can only choose between reply-sender and
reply-list as default. I want to choose between reply-list and
reply-all which does not seem to be supported at all (or making
reply-all the default for that matter). So I always invoke reply-all
manually overriding the default so what the default is is completely
irrelevant.

> 
> Failing those, you can create a folder for a mailing list, filter the
> messages into it, and then use the Folder Properties to set the
> correct address.

But that forces me to create separate folder which is not something I
want.

> 
> > It fixes sending mail to random people unrelated to the message
> > which can be considered security issue. On the other hand it breaks
> > some features in an obvious and easy to workaround way. I would say
> > it's an improvement.
> 
> That's where we differ. It's not about sending messages, but about
> composing them. 

And composing messages is about what? You stack them into your drawer
and never send them?

> Arguably, the current situation is easy to avoid.

It is not. claws gives you a reply window with bogus data - a message
with addresses from another message. Unless you carefully check each
message you compose you send messages to wrong people. The case when a
feature just fails completely is much easier to workaround.

Thanks

Michal