The Beta Thread

Status
Not open for further replies.
Aura

Aura

Active Member
FC/Active Member
The beta is in the hands of about 5-6 people currently, and pending the results of that (i.e. if no major bugs are found) I hope to release it later this week, or early next week.

In the meantime, this thread is for all beta-related questions, comments, etc from the people who are using it.
 
So far it's beautiful. I turned it on yesterday during Dynamis and it was updating in the "Details" view right before my eyes. Every hit someone else would make would register immediately.

I love the immediate reading I'm getting in terms of ACC.
I love the immediate reading I'm getting in terms of crit rate.

I'm in love with the fact that I don't have to wait a fight to see results.

I'm going to mess around with it a little while running the Spyle's equivalent at the same time and see how different the results are.

I'm going to try to crash it. When/if I finally do, what information do you want?
 
If you crash it, send me debug.log in your main parser directory, as well as Dumps\X, where X is the file displayed in the status bar at the bottom of the parser where it says "Logging to file: X". It should have the extension .dvsr

Actually, that dvsr file in the Dumps directory is a raw dump of the entire chat log during your parsing session. When you want to continue a previous parse to add results to it, that's the file you choose after selecting Continue Parse from the main menu.
 
I'm doing some tests from here, I'll fully test it when I get back to Virginia. So far so good on the laptop... I'll have more ideas and better detailed reports after this weekend. :D
 
Stable so far. I just recorded a few hour long BLU75/WAR37 exp session if you want the parse file for workin on the magic portion. In the case of BLU the results are horribly inaccurate as it doesn't account for magic dmg. It may be a little hokey parsing a spell opening/closing an SC.
I'll record Bahamut 2 tonight as well and pass that along.
 
yea, send me those dvsr for the magic parses, that would be hawt.
 
Tested it in a party last night, and did not encounter any bugs during the duration (about two hours). Will run it on the next HNM I get a chance to fight and will try to put it through some stress tests (dynamis without logs, etc) when I can. So far I am pretty impressed, good job Kalia.
 
Make sure to look at the .dvsr files and see if they're there at the start, during, and after a parse is finished. There seem to be some issues in that little area.
 
Fodder is totally owning everyone in the bug finding department. I just fixed another one that he found, where loading an extremely large parse caused the parser to lockup indefinitely. While trying to reproduce that one, I also found a crash that nobody had yet reported, and I fixed that as well :)

So.... load large parses seems to be a good way to test/break it :) Thanks for all the help so far! It's getting better with each bug fix.
 
I had a problem with the same thing, but was disconnected for about 2 hours last night.

I've been trying to recreate a problem I've had with results after clearing the log once a parse has been initiated.

I had started the parse and fought a few Colibri. I later went to the Ashu Talif and cleared the data before entering (without stopping the parse). When I was done with the fight I didn't have results for ACC or Crit rates, but did have the damage percentages and DMG totals. I was going to try to do it again, but I logged off semi early last night. If I can do it again, I'll try to give you more info.
 
Archain said:
I had a problem with the same thing, but was disconnected for about 2 hours last night.
Click to expand...

The same thing, meaning one of the things I mentioned on the previous page that Fodder found? Make sure you download the newest version tonight, as I think I should have nailed all of those.

Current issue status is:
  1. Lockup while loading a parse (Fixed)
  2. Crash while loading a parse (Fixed)
  3. ReadProcessMemory() error in debug.log (Fixed? Better logging)
  4. Parser not saving dvsr dump file even though it says it is logging to file (New issue)
  5. Inaccurate ACC/Crit results after clearing data during running parse (New issue)
 
Three things I'm kinda looking for (you probably already have the first). I've only really messed with this for a grand total of 25 mins, so I might have missed something so far.

1. A count for WS. I used to use Spyles to get a WS count for something like a trial weapon or one of the 500 point ones. It was extremely useful when getting a good count on WS points. During a party it also lets me know just how much more often that SAM is kicking my ass in putting a WS out there.

2. An ability to edit out certain players or fights prior to export. When I used to export to HTML, I would remove fights that I would have done before parsing the actual party or edit out external players that would skew a party's DMG percentage. I had the ability to use only the items I wanted in the final export. I could take out that bat I fought on the way to camp or remove a player I ran by earlier using a Kraken on a crab outside the party. I know I can filter them out during the parse, I'm looking for something to remove fights where someone may have died prior to a final report prep.

3. Enemy kill count. I was looking for something like this yesterday when we were doing the Sergeant Assault with the 40 Chigoes. I ended up attempting to look in the loot tables for this, but still had Colibri results in there.



If I 'tard out and the stuff is already there, allow me some time to delete what I've got above, lol.
 
1. FFXI is showing DMG for all players in my PT. I am the only one showing up in parse. I had dmg for pt members unfiltered for both the BLU records I sent you.

2. Large parses not saved. Now that I had some time to sleep on it, the last line in the Bahamut 2 parse that didn't take was around 65,535. I say around because it was 65k and change on the page I was looking at, but I didn't think to check the last line number. Worth checking at least.
 
I don't think any of those fields are there.

1. In the next version I'm planning detailed weapon skill reports. Right now when you see skill average, and skill average and skill high/low it combines all skills with each other. so if you do evisceration once and dancing edge once both of those factor into your average. detailed weaponskill reports will allow me to distinguish between these two, and provide high/low/averages for each one separately, as well as more information in general about skill usage, including how many were used overall, how many of each type, etc. I guess this is one of those things that has to wait, but I'll make a note of it and put it in the next version.

2. This also isn't there. In a future version I'm planning to try to determine this automatically. My thought is that if a player doesn't do anything for more than like 10 seconds, consider him "inactive" until such time that he does something again. This would exclude all those players who were dead or afk without your intervention at all, although I might still be able to add a feature to exclude certain people.

3. For now the loot table is the best way to see this, but as you probably noticed it doesn't show anything if out of all the mobs with a particular name, none of them ever dropped anything. I may be able to change this in the beta so that it shows an emptytable for enemies that never dropped anything, that would probably have to suffice until the next version, when I can do a better job.
 
Fodder said:
1. FFXI is showing DMG for all players in my PT. I am the only one showing up in parse. I had dmg for pt members unfiltered for both the BLU records I sent you.
Click to expand...
This is weird, I'm still skeptical about this one. I'd like to see if it's reproducible. In the chat log you sent me, I saw things like
Line X - Spundundee readies Spiral Hell.
Line X+1 - blahblahblah
Line X+2 - blahblahblah
Line X+3 - blahblahblah
Line X+4 - blahblahblah
Line X+5 - blahblahblah
Line X+6 - blahblahblah
Line X+7 - blahblahblah
Line X+8 - blahblahblah

and never anywhere anything about the damage he produced from Spiral Hell. That alone tells me that every single line from FFXI was written to that file. Otherwise it would have skipped some lines somewhere, like I would have seen it go from X+1 to X+3. If that really happened, then idk it goes against everything I know about how FFXI does chat logs :( the line number I write out to the file is exactly the same as the line number that FFXI thinks it is. So if FFXI had known about a line where another party member was doing melee damage, either it would be in that file, or it would be skipping lines in the log file. idk, hopefully this can get reproduced somehow.

Fodder said:
2. Large parses not saved. Now that I had some time to sleep on it, the last line in the Bahamut 2 parse that didn't take was around 65,535. I say around because it was 65k and change on the page I was looking at, but I didn't think to check the last line number. Worth checking at least.
Click to expand...

That actually makes sense, off the top of my head I use an unsigned short (16 bits) for the line number field, and the largest number that can fit in that is 65,535. I can fix that without too much trouble.
 
Couple of notes:

I had a couple of random crashes for unknown reasons.... I'm kind of think that was due to other programs running and high-loads. I'll have some time tomorrow to run through that again.

Because of time, I have not been able to do EXP or any other group events... (I'm still in Denver. :( ) --- I'll get more done before I fly out on Sunday...

Everything else, seems ok.
 
Whenever you get a crash, make sure to check debug.log (Especially using the newest version, it prints more logging).
 
Kalia said:
That actually makes sense, off the top of my head I use an unsigned short (16 bits) for the line number field, and the largest number that can fit in that is 65,535. I can fix that without too much trouble.
Click to expand...

Bah, that actually isn't the problem after all. I use a 32 bit integer, and I even searched the entire code for all occurences of 16 bit integer use, and none of the search results have anything to do with how many lines. Maybe I'll try parsing unfiltered in Besieged to see if I can reproduce this.
 
Status
Not open for further replies.
Back
Top