Hi All, Just a few comment (great report btw John). 2. TwitterTwitter has dropped that old user/password authentication thing and moved to oauth. The procedure to get oauth up and running with a client is somewhat complicated.
I guess you have all been through that by now but to sum it up. 1. Initiate an authentication from the client (jpskmail in this case)2. That request should point to twitters web site and present an html form where the user can log in using his/her user and password (this is all through a server during a connected session then).
3. When logged in twitter will display an authentication number series.4. That unique oauth number is entered at the local client and is used to authenticate all interaction with twitter.
We can build around this but with our loosely connected architecture its a challenge. Is it worth it?
Another way is to go for google buzz instead of twitter. 3. Outbox null pointerI have not seen that. If I can replicate that behaviour then I will fix it (just need to know how to cause it). I'll look for it. Is it when a mouse click is done somewhere in the table? On the header?
73 de Per, sm0rwo 2010-12-21 12:35, Rein Couperus skrev:
Hi John, tnx for the tests, very valuable information. Client->Server: 1.: This has to be tested further, I have not seen that behaviour here... 2.: Twitter has broken our app by changing their API., This is on the todo list, the server code has to be changed. (volunteers?) 3.: Maybe Per can look into this, I have seen the error message as well. Does not break the application... TTY: 1.: You guessed it, the CQ button is for calling CQ for a client to client chat. It is used unproto, and a responder just has to push the CONNECT button to be in business... 2.: Has to be investigated... 3.: Good idea, I have changed it so compressed content is only shown in the Files window. There it is valuable, as you can see immediately when one of the lines becomes carapped (never happens, but this is a reassurance for the OP :) 4.: Good idea, I will change it. 5.: Hmmm... I just tested all 4 quadrants, and no error... Maybe something todo with numbers> 90 degrees, will do more tests. I will make installers for the client including the changes, so we can all test over Xmas holidays :) 73, Rein PA0R Rein and Per, Here are the results of testing the alpha version of Pskmail both server and client: Versions were: 1.0.25 for the server and revision 924 from the subversion repo running under WinXP SP3. 1. As per the commented log attached, I had to repeat one email send 3 times (without deleting the temp file). It seems there are issues with the client not sending the FO block before issuing an FM (not always but very often, even after close and restart). 2. I could not get to send twitter updates. Not sure if it is a password issue or not. 3. Outbox often reports a null pointer error when empty. Seems fixed on latest svn revision so not sure if this was specifically addressed since then or not. Here are the results of the TTY to TTY exchanges bases on revision 929 of the client running on Windows XP SP3 both sides. 1. What is the CQ button for? Is it for initiating a TTY session between two clients? 2. In TTY session when selecting "pending transactions" not all pending transactions are shown. 3. Improvement: since the terminal window is the main exchange window in TTY sessions, I propose not to display the content of compressed data exchanges as this creates a confusing mix of text and data. I would propose to display the data exchange in the file window so that it can be referred to if required, but would leave the terminal window uncluttered. In fact I would prefer to have this behavior also for server connections since the data displayed in the terminal windows does not add value but makes it more messy. We already have the progress bar and if necessary could display the data exchange in the file window. 4. Improvement: In the file window we could duplicate the display of the file exchanges (sent file xyz, received file xyz etc...) to avoid having to switch between the terminal and file window to check that the transfer was done correctly. Also, pretty important: The client's igate is not reporting the received position properly (maybe due to negative numbers): -33.7846, 150.7111 becomes 22'47.08S (correct), 06042.55E. The beacons seems to be reported properly through the server. That's all at this time. I don't have much time for coding these days. It should get better after Christmas. All the best and merry Christmas and Happy new year (with not too much snow for you guys up there I hope). 73's, John