Re: About pastebin applet

2010-02-02 Thread Will Stephenson
On Thursday 28 Jan 2010 17:41:09 Artur Souza (MoRpHeUz) wrote:
 So, what do you think ? Pastebin turns into Share It! with a better
 architecture.

I dream of a desktop where these kinds of sharing are all accessible in the 
same way:

Sources:
local files
clipboard
URLs

Destinations:
Paste into widgets
Online paste/image/file bins
Online galleries (flickr/picasa/...)
Contacts (mail attachment, IM file xfer, bluetooth obex, zeroconf file xfer eg 
kepas[1])
'Facebook Wall' (for URLs), digg

On the face of it you would need a list of mimetypes that each destination 
would accept, config UI for destinations requiring registration, and a UI 
plugin to provide additional metadata needed (eg user name for *bins, url 
comment text for facebook wall, image title for galleries.  For the 
heavyweight media like email and IM I assume the plugin would just call KMail 
or Kopete and tell them to open a message composer or file transfer window 
primed with the content to send, and let them use their existing UI to select 
the contact to send it to (I implemented the Send To-Contact function in KDE 
3's kuick konqueror plugin and it was always a bit slow to fetch the list of 
online contacts in a slot connected to the context menu's aboutToShow() 
signal).

You could then use this Share it! broker in a plasma drop widget, Klipper and 
Copy To... in file managers.

So while you're rethinking things, go for the biggest picture, it's the KDE 
way! 

Will
___
Plasma-devel mailing list
Plasma-devel@kde.org
https://mail.kde.org/mailman/listinfo/plasma-devel


About pastebin applet

2010-01-28 Thread Artur Souza (MoRpHeUz)
Hello all,

I've been thinking about a nice way to improve the current pastebin applet and 
it's dataengine/service. So I'll share my ideas with you and you can say 
whatever comes to your mind :)

First of all, pastebin started as a plasmoid to just post text on pastebin.ca 
service. Willing to allow more than one server and at the same time don't 
change it to much (what actually happened looking backwards) I went to a 
simple and not elegant solution but that actually solved the issue and Sebas 
even gave some love to the UI to look better. Imagebin support came easily 
after this.

However, besides have been elected one of the top plasmoid by linux magazine it 
actually sucks in terms of architecture. Last tokamak while one compilation or 
another of KDE for the mysterious device(TM) I tried to create a 
dataengine/service from the plasmoid code to make it better and even allow 
other plasmoids to make use of the service itself.

Currently, in order to expand it making it support more services (each user on 
Earth seems to have it's preferred service) is painful. Most of the imagebin 
services are duplicating the http POST code got from KIPI plugins source code 
and to enable more services we need to recompile the whole thing and the 
support comes out on next KDE release.

Here comes the new idea: rewrite it to be based on plugins (easily achievable) 
and change it's name to something more user friendly like Share It! or 
something else.

Motivation: based on simple plugins we can provide stuff like GHNS for new 
services making it easier to expand it, rewrite the plasmoid code to use the 
new QStateMachine as it's basically a state machine, fix the dataengine/service 
code to provide one function for http POST based services so the plugins 
doesn't need to actually duplicate code and even make it work on different to 
avoid bugs of people pasting too much content at once (there is this bug where 
the user posted by mistake(?) a 12MB file and it froze plasma).

Another clear motivation is how user friendly Pastebin is ? It's more of a 
share tool and people that actually don't know what pastebin means looses a 
good chance/tool to share their stuff. It would even make sense to create 
plugins for flickr and picasa for example (maybe using the already present KIPI 
plugins for that).

So, what do you think ? Pastebin turns into Share It! with a better 
architecture.

/me waiting for the no's and don't's now :)

Cheers,

--
Artur Duque de Souza
openBossa
INdT - Instituto Nokia de Tecnologia
--
Blog: http://blog.morpheuz.cc
PGP: 0xDBEEAAC3 @ wwwkeys.pgp.net
--


signature.asc
Description: This is a digitally signed message part.
___
Plasma-devel mailing list
Plasma-devel@kde.org
https://mail.kde.org/mailman/listinfo/plasma-devel


Re: About pastebin applet

2010-01-28 Thread Sebastian Kügler
On Thursday 28 January 2010 17:41:09 Artur Souza (MoRpHeUz) wrote:
 Hello all,
 
 I've been thinking about a nice way to improve the current pastebin applet
 and it's dataengine/service. So I'll share my ideas with you and you can
 say whatever comes to your mind :)
 
 First of all, pastebin started as a plasmoid to just post text on
 pastebin.ca service. Willing to allow more than one server and at the same
 time don't change it to much (what actually happened looking
 backwards) I went to a simple and not elegant solution but that actually
 solved the issue and Sebas even gave some love to the UI to look better.
 Imagebin support came easily after this.
 
 However, besides have been elected one of the top plasmoid by linux
 magazine it actually sucks in terms of architecture. Last tokamak while
 one compilation or another of KDE for the mysterious device(TM) I tried to
 create a
 dataengine/service from the plasmoid code to make it better and even allow
 other plasmoids to make use of the service itself.
 
 Currently, in order to expand it making it support more services (each user
 on Earth seems to have it's preferred service) is painful. Most of the
 imagebin services are duplicating the http POST code got from KIPI plugins
 source code and to enable more services we need to recompile the whole
 thing and the support comes out on next KDE release.
 
 Here comes the new idea: rewrite it to be based on plugins (easily
 achievable) and change it's name to something more user friendly like
 Share It! or something else.
 
 Motivation: based on simple plugins we can provide stuff like GHNS for new
 services making it easier to expand it, rewrite the plasmoid code to use
 the new QStateMachine as it's basically a state machine, 

Yeah, the code actually already uses QStateMachine concepts, it was inspired by 
it 
(but QSM wasn't released at that point). :)

 fix the
 dataengine/service code to provide one function for http POST based
 services so the plugins doesn't need to actually duplicate code and even
 make it work on different to avoid bugs of people pasting too much content
 at once (there is this bug where the user posted by mistake(?) a 12MB file
 and it froze plasma).

I assume you mean making those plugins scripts (Python, or something), 
otherwise 
sharing via GHNS can be pretty painful.

 Another clear motivation is how user friendly Pastebin is ? It's more of
 a share tool and people that actually don't know what pastebin means
 looses a good chance/tool to share their stuff. It would even make sense
 to create plugins for flickr and picasa for example (maybe using the
 already present KIPI plugins for that).
 
 So, what do you think ? Pastebin turns into Share It! with a better
 architecture.

Sounds good. :)

 /me waiting for the no's and don't's now :)
-- 
sebas

http://www.kde.org | http://vizZzion.org | GPG Key ID: 9119 0EF9
___
Plasma-devel mailing list
Plasma-devel@kde.org
https://mail.kde.org/mailman/listinfo/plasma-devel