The text client shall live on in Python (badly written and buggy), but at least you can see the bugs in the source code. No, you can't understand the writers intention (even if they used docstrings).
Just kidding. Carry on (writing it in C). C is not bad, but I have to think too much to grock it these days.
Wed Mar 23 2016 01:53:34 EDT from ax25 @ UncensoredThe text client shall live on in Python (badly written and buggy), but at least you can see the bugs in the source code. No, you can't understand the writers intention (even if they used docstrings).
Just kidding. Carry on (writing it in C). C is not bad, but I have to think too much to grock it these days.
Where can we see the python text client?
Umm ... to the best of my knowledge there is no Python text client. There are Python bindings to the Citadel wire protocol, but no client.
Honestly though, if I were to bind in an extension language for Citadel, it would almost certainly be JavaScript.
It was just some horrible hack I used to allow someone that connected to my packet radio node program to get to Citadel in a text client way (telnet based like one of the old text clients). It was a learning experience in telnet for me - especially the bits on the negotiation between the client and the server.
It is not worth viewing, but here it is:
http://penguinpackets.com/~kelly/kblog/projects
So ... as Ragnar pointed out in the Citadel Support room ... the text client works flawlessly on the new Microsoft/Ubuntu Linux layer for Windows. You need doze10 to run it, though. Still good news for those of us who are penguinistas stuck on Windows machines at work.
Schweet. Glad I am not a Wintendo drone (since 1999!). Good to hear they are catching up though. Also glad to hear that the Red Sea has parted. Just hope the other predictions come true.
Sat Apr 30 2016 11:03:08 AM EDT from IGnatius T Foobar @ Uncensored
So ... as Ragnar pointed out in the Citadel Support room ... the text client works flawlessly on the new Microsoft/Ubuntu Linux layer for Windows. You need doze10 to run it, though. Still good news for those of us who are penguinistas stuck on Windows machines at work.
Sat Apr 30 2016 11:03:08 EDT from IGnatius T Foobar @ Uncensored
So ... as Ragnar pointed out in the Citadel Support room ... the text client works flawlessly on the new Microsoft/Ubuntu Linux layer for Windows. You need doze10 to run it, though. Still good news for those of us who are penguinistas stuck on Windows machines at work.
MS Windows 10, you mean? I mostly use Debian or Ubuntu, but I've been planning to install Windows 10 Pro on one of my PCs; in part so I could check out that Microsoft/Ubuntu Linux layer for Windows. (I spend much of my time working in command line windows, doesn't matter all that much to me which GUI I'm using...) And trying out the citadel text client is another thing I planned to test; glad to hear that it works all right. Does it matter which version of the text client is used.
Yeah, it's the "ubuntu layer" for MS Windows 10, but I did discover that you have to put Windows into "developer mode" and accept beta-quality code (whatever that means in Microsoft land) and they don't recommend you run it "on a production machine" so I may have to wait a little while longer.
Once you've got it running, though, you can get the Citadel client right out of the Ubuntu repository, simply by opening a bash prompt and typing "apt-get install citadel-client"
It looks like the text client needs a bit of saving. All I get when I try to telnet into Uncensored is:
"couldn't stat citadel.config: No such file or directory"
Sorry about that ... telnetd was throwing a "-p" at the client on the command line and I accidentally removed the code to parse that. It's fixed now.
You can also connect with SSH as user "bbs"
This can't be the only room that is not unread when I log in to the web client. Or can it? I am posting here as a test, and will check again with the web client later to see if it un-sets the un-read message count. :-)
I have noticed odd read/unread message counts when switching between the client and the web interface.
For example, if some action takes you to another room without you explicitly selecting a "Goto" option, the behavior can be unpredictable. This is the web client's fault; the text client *always* does the right thing.
Lately I've been getting a lot of "not found" errors on the web client, which usually means some other user agent attached to my session and moved to a different room while I wasn't looking. The existing version of WebCit carries a lot of state with each session, including the room that it thinks it's currently in. This is going away in the new version; no state is kept except for the identity of the user who's logged in -- everything else is RESTful. This will lessen the unexpected behavior while also allowing more user agents to be multiplexed onto fewer sessions.
I have received the not found error once or twice. A few times I see a new message from user "-1" it says something odd like click on a button to edit something.
*sigh*
Every time I look at the source code to the text client, I'm reminded of how bad my coding style was, almost 30 years ago.
That's what's mostly in there, of course, because when our Citadel implementation was split up into client and server in 1995, the structure and form of the monolithic implementation was left in the client, and the server was built as a new program. So the client still has ugly stuff all over the place, lots of goto statements and flow that just doesn't make sense.
But it isn't worth rewriting, not at this moment in the universe anyway.