« first day (374 days earlier)      last day (3202 days later) » 
00:00 - 19:0019:00 - 00:00

@Duga Thanks @Marc-Andre!
 
[Cardshifter/Cardshifter] build for commit 749bf3ae on ???: The Travis CI build passed
 
@jacwah No problem!
 
7:56 PM
@Phrancis I'm a little confused by your question. Are you interesting the structure of a card?
@jacwah Is the client working now?
 
Monking
 
Monking @Dan
 
Hello, @DanPantry.
 
@SirPython The web client? I didn't change anything relating to that so I guess not.
I must be doing something wrong as it works fine for you, right?
 
Yes, it works fine for me.
What are the exact steps that you take to run it?
 
8:02 PM
1. Start server with IDEA->Run
2. Start node server with `npm run develop`
3. Go to `localhost:8080/webpack-dev-server` in Firefox
4. Choose Local host as server
5. Type user name `jacwah`
6. Press "Log in"
 
Have you installed the dependencies for the client?
 
I have run npm i
 
That is so weird.
I do the exact same thing (except for running the server from the IDEA)
 
I know, it doesn't make sense at all
OK I'll try to run the server the same way
How do you run it?
 
I run the server by navigating to the release that I downloaded from Duga, then typing in java -jar <server.jar>
But I don't think that makes a difference; I just don't know how to use the IDEA.
 
8:05 PM
I'll try
If the same thing still happens, we've ruled that out
 
Ah....
IntelliJ might be using a different main class than the jar?
 
Might be, let's see
It worked!
 
bows
 
Haha WTF
@skiwi It's not a different class, I checked the manifest. It might be a different source version though
 
@jacwah Weeeeeeird
 
8:11 PM
The commit hash should be distributed with snapshots...
 
Does Cardshifter have a logo?
 
8:25 PM
OK so I checked out the commit tagged 0.6 and built a dist package - but it failed
There is something magical about that dist... I also tried to make a dist from master (which supposedly the latest snapshot is based on). That failed to.
 
is this web client or html cliennt?
 
The HTML client - although when I've said web client, that's what I've meant
 
8:47 PM
@SirPython we do have a logo, it's pretty low res but I'll hook you up when I'm back on my pc
 
Thanks. It's just that the HTML client could use a favicon.
 
9:06 PM
Alright, I think that I don't really understand how Cardshifter works: are different mods entirely different versions of the game? I say this because, as I am reading server logs from testing out both current mods, I am getting entirely different messages at different times.
 
@SirPython Define "entirely different"
 
I haven't looked entirely into the messages being sent, but Cyborg Chronicles sends me 18 messages, and Mythos sends me 31-32 messages upon the start of the game.
@jacwah Maybe different in general: is card information sent at different times? Are there different options for actions? Etc.
 
TBH I don't know anything about this, but browsing the code of the mods (specifically Game.groovy) I see that CC has ~60 LOC setup(), while Mythos has ~100 LOC
 
I guess I need to look more into what a mod actually is.
 
Look inside extra-resources/mods
 
9:15 PM
Ah, thanks. I didn't see that before.
Maybe the better question is: "How much and what variety can mods add to Cardshifter?"
 
9:30 PM
> This happens _only when not_ using http://www.cardshifter.com/releases/cardshifter-0.6.1-SNAPSHOT-20150729-0036.zip as far as I can tell.

These are the versions I've tried:
- master
- develop
- 0.6 tag

These are the methods I've tried
- IDEA->Run
- `./gradlew dist && java -jar`

When a HTML client connects through WebSocket this is what is logged:
```
[2015-08-11 22:16:27,944] INFO ServerWeb [WebSocketWorker-11] ( ServerWeb.java: 46) - Connection opened: org.java_websoc
 
@jacwah Are you there? I could use some help testing something.
 
@SirPython Sure
 
Thanks! Please start the HTML client and connect to dwarftowers.com -- there is no need to start up a local server.
 
I can't... Looks like the same issue as above
The "Log in" button is shaded like it's being pressed and nothing happens
 
Thanks. I just wanted to make sure it wasn't only me.
I don't think I can even connect with my Java client..
 
9:39 PM
@SirPython I can
Someone involved in the HTML client should try to check what commit caused this to happend. git bisect is useful
 
I'll try it now.
Hmm. I don't see how this can possibly work. How is git supposed to test if a bug was introduced?
 
You test it. Git helps you check out commits to test. You specify a know good and know bad commit and it starts a binary search. First it checks out the commit in the middle. If it's good -> check out commit between middle and bad, if it's bad -> check out commit between middle and good.
 
Ah, I see.
 
setup debugging tools

This commit is to store the current code in the case that I break something when trying to use bisect
 
@Duga There's always git reflog :)
 
9:58 PM
Hmm. So I have to find the last working commit? As in, the commit before this server problem started happening?
 
@SirPython If you know the latest working commit then we know the commit that broke it. For the "good" commit specify any commit that works. Try something a few days ago or whenever you remember connecting to dwardtowers
 
Ah, that sounds better than what I was going to do.
I'm using this as a guide.
Will any working commit suffice?
 
Yes
What's wrong with that guide?
 
Nothing. Sorry.
 
The closer the good and bad commits are, the less you have to search, but don't spend a lot of time trying to find the perfect ones - that's what bisect will help you do
 
10:25 PM
Finally: I found a working commit.
Hmm. git bisect brought me down to a single commit, but it doesn't make any sense that this would be the commit that messed things up.
 
@SirPython Weird. Have you tried with the commit before and after to ensure it's right?
 
10:41 PM
I'm about to. I've tried the commit before and it works, now I'm going to try the commit it gave me. Then, I'll try the one after.
I may even run the git bisect again to make sure I get the same answer.
 
Good luck
 
21 hours ago, by Phrancis
I hope there's not a "You'll need it!" implied ;)
 
:)
 
Ran git bisect again and was brought back to the same commit.
I still don't get how the client could only connect to some servers and not others.
And, the commit it brought me to is a commit that changes the way a message is sent.
This change would apply to all messages sent, not to ones on specific servers.
@bazola Is it possible that dwarftowers is running an old version of the server?
 
11:12 PM
Post your findings on the issue I opened github.com/Cardshifter/Cardshifter/issues/353
Great work so far!
 
Monkernoon
 
Hello @Phrancis!
 
Lots to read back on!
 
Contradicting 40k users is scary...
 
> I did a little of [bisect](file:///C:/Program%20Files%20(x86)/Git/doc/git/html/git-bisect.html)ing and was brought back to [this commit](https://github.com/Cardshifter/HTML-Client/commit/6b885c8914698561a961fabb225980071fcb6664).

I can't see how this could've possibly broken the code: in fact, this commit was fixing a bug where the `.command` attribute of the message was not being sent.

However, that confuses me even more: how did the previous commit work when the `.command` attribute wa
 
11:27 PM
@jacwah I have posted them ^^
 
@Duga
 
Fixed a typo in the comment.
 
> Aha! I was wrong about the previous commit working; the client would navigate to the lobby page regardless of what the server said.

To test, I opened up a local server and tried connecting to that instead; there was an error for there being no `.command`.

I will now try to bisect the project again, now knowing that I should test each commit on two servers: dwarftowers and a local host server.
 
@jacwah ^^
 
11:41 PM
@Marc-Andre @SirPython @jacwah @DanPantry @bazola @skiwi ^^
Hope this helps! Please let me know if you see anything that should be reworded/corrected.
 
Is it called the "Web Client" or the "HTML Client"?
 
@Phrancis Great!
 
@SirPython I prefer Web Client personally; the GWT client is technically a HTML client as well, although it is rather "shoehorned" into working for web, rather than being made for it
 
200_success vs. rolfl: 3224 diff. Year: +2478. Quarter: +2812. Month: +233. Week: +63. Day: +1.
 
(GWT "translates" Java into JavaScript, which in our case has shown to be a rather clunky implementation)
 
11:45 PM
Mat's Mug vs. Simon André Forsberg: 646 diff. Year: +1536. Quarter: +32. Month: +310. Week: +166. Day: +59.
Loki Astari vs. Simon André Forsberg: 2920 diff. Year: -2856. Quarter: -1260. Month: +40. Week: +95. Day: +20.
200_success vs. janos: 15293 diff. Year: -4307. Quarter: -1386. Month: -572. Week: -90. Day: -39.
 
@Duga Fixed small stuff
 
> Great job @jacwah :+1:
Refuse login if user name is already taken

The server doesn't handle failed logins correctly yet, see #347.

Resolves #345
[Cardshifter/Cardshifter] jacwah pushed commit a44a9cf5 to develop: Add test for user name already in use
Refactor validation of user names

Validation is now done inside Server. A client's name should always be
set with trySetClientName.

You can no longer use an empty user name.
Refactor user name exceptions

As per suggestion by Marc-André.
Merge pull request #348 from jacwah/develop-names

Refuse login if user name is already taken
 
Thanks @Phrancis! I just need to work on getting the server to handle failed logins correctly now...
 
@Phrancis It does? That's pretty cool.
 
11:48 PM
@SirPython ... for some values of "cool" =;)
 
@Phrancis Should I open a separate issue for user name validation since you closed #345?
 
[Cardshifter/Cardshifter] build for commit ed3a80ba on develop: The Travis CI build passed
 
Oh right...
 
@jacwah Wait; can you clarify that name validation you want to add?
@SirPython I need to get a better-looking one from my friend who made this
 
11:52 PM
I posted a suggestion here
A user name...
    is unique. Two clients can not connect with the same name simultaneously.
    is at least 1 character and at most 20 characters long.
    contains only ASCII letters, spaces, underscores and decimal numbers.
    does not begin or end with whitespace.
It's just a starting point for discussion - feel free to comment on any of the points
 
@jacwah Why don't we go ahead and address that separately if you don't mind. I think there are other points related specifically to validation that I think would benefit from their own work thread
 
@Phrancis Work thread == issue?
 
Right
 
Ok I'll open a new one
 
@SirPython See this for example:
{
  "MAX_HEALTH" : 10,
  "SICKNESS" : 1,
  "MANA_COST" : 30,
  "ATTACK" : 3,
  "HEALTH" : 10,
  "ATTACK_AVAILABLE" : 1,
  "flavor" : "The Great Grandfather, Emperor of all Deities, Vanquisher of Evil.",
  "name" : "JADE EMPEROR",
  "description" : "Give creatures of type Chinese owned by you on Battlefield 3 HEALTH\nGive creatures on Battlefield -1 ATTACK\n",
  "type" : "Chinese God"
}
What I need to be able to do is access those fields to display in the web client.
7 hours ago, by Phrancis
@SirPython I have a question; I saw in deck builder you used {{card.properties.name}} in one place, and {{card.max}} another place. How do you differentiate which properties are part of which set? I'd like to document that as I will need to be able to query these fields in several areas of the client
Well; I'm mostly wondering how I need to go about it
 
11:59 PM
I still don't entirely understand the question. What do you mean by "differentiate which properties are part of which set"?
What is the "set"?
 
00:00 - 19:0019:00 - 00:00

« first day (374 days earlier)      last day (3202 days later) »