|Current situation||Wil Alvarez||6/18/13 2:41 PM|
First, I want to thank you all for taking the time to report all the issues with Turpial 1.6.9 and try to find a fix. Let me clarify a little bit the current situation.
Turpial 1.6.9 is no working anymore because of the new Twitter API version. I should implement support for the API 1.1 but honestly I don't want to do it on 1.6.9 (in fact, I don't want to support that version anymore)
Why? Well, because of the greater things that can be achieved with libturpial.
Ok but, where is this amazing library and why it has been taking so long to see the light? I've been working with libturpial to make some cool refactors and add support for Twitter API 1.1 but I got lot of problems. Among them, a very annoying memory leak that took me more than two months to fix and when I thought I was done I started to have issues with OAuth. I wrote on Twitter forum  to see if I got some answer but got nothing. It was a very frustrating situation (I even thought in give up) but before taking any action I decided to give it one more shot. So, I tested yesterday and with no apparent explanation everything worked, like magic, without reason, without changes on source code. Shit happens. Finally, I've decided to continue with the project.
Why I don't use an existing library for twitter? Because since the beginning of the times, libturpial has been designed to support multiple accounts, multiple protocols, multiple columns and a lot of more features that make the programmers life better and there is no library that let you do such things at the moment. Maybe it was a very ambitious concept and probably I'm failing on the same errors than the GNU Hurd kernel but well, nobody said it would be easy.
So, taking this decision let me just a short period of time to develop a new version of Turpial that fixes the API support issue and implement more functionalities so it would be nice to get some help. Any help will be really appreciated, this will be my last try to rescue Turpial and if I can't make it I'll give up.
What needs to be done?
I will create issues on github to track pending tasks and I'll do my best to deliver (finally) a new release of Turpial. If you want to collaborate and have doubts, just ask, I'll answer any doubt gladly. Post a new thread, send me a mail, talk to me on Twitter or whatever you want.
Thanks for your attention and your support
|Re: [turpial-dev] Current situation||@orvtech||6/19/13 11:06 PM|
Noobish question and a bit off-topic.
I have started to mess with python and have written some simple scripts that interact with twitter through python-twitter . Where should I start if I wanted to play with libturpial? How do I install it, documentation, etc..? ...
Just realized all the answers to my questions were on https://github.com/satanas/libturpial but could I see some code that show how to a simple status update? sort like "hello word" kinda thing.
Being a noob to python for me some to understand which libraries I need to call (or how to) and some minor syntax stuff. I have (unpublished) some piece of python code that will determine when is best for one to tweet. My goal is to incorporate these as plugins (and learn python... so it is going to be a long way).
|Re: [turpial-dev] Current situation||Wil Alvarez||6/20/13 6:30 AM|
2013/6/20 Oliver Rivas <orv...@orvtech.com>
Yesterday I was working with libturpial and after all the changes I did, the documentation on the source code is not up-to-date, actually I've an issue created  to start updating doc when all changes are done. However I can guide you through a very simple step that will let you play a little bit with libturpial. Let's begin:
* First, clone my personal libturpial repo  and go to the development branch
* Install libturpial in development mode (for more information use  as reference):
# python setup.py develop
Now you need start coding. First, you need an account instance:
from libturpial.api.models.account import *
a = Account.new('twitter')
That way you have an empty twitter account. After this, you need to authenticate with OAuth, let's ask for a request_token:
url = a.request_oauth_access()
There you will see an url, open it in your browser, authorize Turpial and copy the pin. Then:
Where '1234567' is the pin returned by Twitter. Now we have a fully functional twitter account (but just in memory). If you want to save your credentials on disk use:
Let's suppose that you close Python terminal and at the next day you want to test libturpial. Well, you won't need to repeat all this process just load your account and you're done:
from libturpial.api.models.account import *
a = Account.load('youruser-twitter')
libturpial handles all account ids as "user-protocol", so it's easy to guess which is the id for any account. All the information is stored in ~/.config/turpial/accounts. Take a look if you want. Try any method you see in  to fetch timeline, directs, etc.
That is not all. The final way to use libturpial will be through core.py module, which is responsible for handle multiple accounts, columns and all that stuff. However it is under heavy development and it will be changing a lot. When core be more stable I can post some instructions to use it.
It would be nice too create a page in the development section at the Wiki  with this kind of information, I hope somebody could help with this.
Well, that's all for the moment. Just let me know if you have any issue testing and I'll help you gladly
|Re: [turpial-dev] Current situation||X3MBoy||6/20/13 8:24 AM|
I can help with wiki edition, just give me an account to do it.
2013/6/20 Wil Alvarez <wil.al...@gmail.com>
GNU/Linux User #403161
Ubuntu User #8749
|Re: [turpial-dev] Current situation||Andre Jaenisch||6/20/13 11:41 AM|
Am Donnerstag, 20. Juni 2013 15:30:47 UTC+2 schrieb Wil Alvarez:Hi Oliver,
Sounds pretty easy.
Would it be possible to define the API at the moment? That is, which format (JSON, Arrays, Dictionaries etc.) expect some other functions and so on.
Oh, and is the suffix "-twitter" mandatory? I've read the example as, well, an example. So I thought I could simply ignore the -twitter.
Can the documentation be recreated via Sphinx?
One last thing: Python 2.7 or, say, 3.3?
|Re: [turpial-dev] Current situation||Wil Alvarez||6/20/13 12:55 PM|
2013/6/20 Andre Jaenisch <andrej...@googlemail.com>
Am Donnerstag, 20. Juni 2013 15:30:47 UTC+2 schrieb Wil Alvarez:
Of course, would you like to open a new thread to start discussing this?
Yep, right now is mandatory because is the way libturpial can identify if the account belongs to twitter, identica or any other service. It's part of the account id can't be ignored at the moment. Probably we could work in a better way of do this.
Totally! The current documentation is under Sphinx format ;)
Python 3.3 is a must but not right now. I think that we should make it work first, so let's go with Python 2.7. When we are done with the first stable release we could start to porting to Python 3.3
Sounds fair to you?
|Re: Current situation||Andre Jaenisch||6/22/13 2:20 AM|
Am Dienstag, 18. Juni 2013 23:41:50 UTC+2 schrieb Wil Alvarez:
Oh, yeah, fortune and fame ... Will we get at least a gold star (or bird)? ^.^
As far as I understand the current partition, Turpial.git is responsible for the view (e.g. Qt, GTK3, ncurses), whereas libturpial is the model (handling requests and so on). I've read your blog post why you dislike GTK. Is it though possible, to have a dynamically defined layout? Say, the choice between one, two or three columns. For me, I'd prefer Timeline+Mentions.
Shall I start a new thread for this?
Links, please. To GitHub, mailto and Twitter account (shorter way to get in touch.)
May you edit your post to place the links like in your blog post, too?
|Re: [turpial-dev] Current situation||Andre Jaenisch||6/22/13 2:24 AM|
Am Donnerstag, 20. Juni 2013 21:55:26 UTC+2 schrieb Wil Alvarez:
Ehm, wait a minute. I do. But joining my wife's birthday party this weekend, so no further contributions ;-)
What about placing this information in the [section] of .cfg file? One can determine which [section] to choose on login. Sounds more intuitively. Or are there shared information between twitter and identi.ca / pump.io? Have no account on the latter ...
Yepp, it does. Although I prefer the "latest and greatest" (hence using Sabayon Linux based on Gentoo testing) I'm familiar with Python 2.7 (will make Debian users happy, I guess :))
|Re: [turpial-dev] Current situation||Wil Alvarez||7/9/13 7:32 PM|
2013/6/20 Eduard Lucena <eduard...@gmail.com>
I tried to create your account and suddenly I noticed that somebody (presumably a bot) created more than 40.000 user accounts and we run out of space to create more. I'll try to fix this situation ASAP, however if somebody has worked with mediawiki before and knows how can we solve this please give me a hand.