292 replies to this topic

[Deathlike]

I'm not going to break his post down, but I have somewhat of an idea what he's talking about. Here's part of the post that is technically relevent.

Quote

CryScaleform layer attempted to lookup a structure which did not yet exist in their cache layer, this triggers a load and bind operation. This lookup operation involved passing a string key value as a const char * to the map operator[] function; however the operator[] function was templated in such a way that the key value provided actually expanded to a const char * at compile time rather than a CryName instance.

What followed was a terrible failure case, where a CryName instance would be constructed, inserted into the global name table, and compared against a given node within the map, and then destroyed and completely removed from the global name table; only to happen all over again at the next level down in the red-black tree.

There some "oddities" when you look at what manifested itself. In some of Koniving's videos, you could see his own name as the name of the mech. I personally have seen my name as part of the topmost scoreboard (Deathlike to random score #).

So, what he's saying that the game kinda tries to get data that isn't there yet (for whatever the reason), and kinda gets other data from other areas (where it shouldn't be getting in the first place)... which pretty much results in the HUD going out of whack.

Imagine if this bug continued further and did more than just HUD corruption (like the current file corruption that some of us are still suffering from).. it would be a disaster. Now only if PGI gets onto that magical tour of fun.

On the surface, it looks like PGI's code is "working as intended"... it's the foundation that it was built on was not.

Edited by Deathlike, 14 May 2013 - 01:15 PM.

[Multitallented]

I'd like to think I understood all of that. I think I got lost at the part where the function in question was trying to use undefined stuff in the library and thus causing incorrect/unpredicable HUD bugs. At any rate, I look forward to my friends being able to play this game reliably as it has been a huge problem.

I'm also curious if this has had any effect on fps? I've always felt that MWO has had serious inefficiencies with fps.

Anyway, thanks again!

[dyndragon]

Thank you for the technical write up. Keep doing these--they're kinda fun and keeps the community engaged, as silly as that sounds.

[Kyle Lawrence]

Chemie, on 14 May 2013 - 12:56 PM, said:

Does Karl work for PGI or is there truth to the rumor than PGI hired an outside compnay to fix bugs (generically vs HUD)

Karl's been on the MWO team since the beginning. He just normally works on a separate system. His attention was simply diverted temporarily for this hunt.

[DCLXVI]

I actually havnt had hud bugs in 2 days now anyway ^^ i get flashes of tearing during the start up sequence now, not that i care at all amd6870 13.5 beta win8 64

[El Moosechacho]

Hammerfinn, on 14 May 2013 - 11:19 AM, said:

Can we post this in every thread where people complain about laziness or incompetence at PGI?

Yes indeed.

What a colossal pain in the *** for such a (relatively) small issue. Props to PGI.

Edited by El Moosechacho, 14 May 2013 - 01:22 PM.

[Syric Malvayis]

A very interesting read....

...thank you for taking the time to share with us the teams woes, and victory, over there! It's posts like this that remind me GREAT games take time, patience, and a whole hell-of-a-lot of caffeine.

A personal, "thank you," from me to all of you guys n' girls over there that are making our dream MWO a reality! I sincerely appreciate ALL THE HARD EFFORT going into making this game RIGHT.

[Belorion]

Thontor, on 14 May 2013 - 01:12 PM, said:

Reminds me why I quit "computer science" after one semester

Going though code line by line sounds like absolute hell

Thanks Karl and the rest of the PGI team for the hard work and long hours tracking this down.

Its really not that hard most of the time. There are many tools that let you pin point trouble areas of the code and find what is broken. Finding a third party issue really is a pain though, and you can't always fix the problem if you don't have access to the source code that is causing the issue, and instead have to find a work around.

Edited by Belorion, 14 May 2013 - 01:27 PM.

[WolvesX]

just two words...

AWESOME JOB!

AWESOME JOB!

AWESOME JOB!

This is just great!

[CompproB237]

I just wanted to thank you for explaining the process behind fixing this. It really gave me a good idea of how much work had to be put in.

I'd love if you could continue to post explanations like this. I really enjoyed the read even if some of it I didn't quite understand.

[Rebas Kradd]

Coolant, on 14 May 2013 - 01:09 PM, said:

PGI has created a fun game, cares about the IP, and trying hard to do it justice, but they are far from perfect, and your post (2nd post in the thread, you were ready for retribution), makes it seem like they do no wrong.

No, it doesn't. It simply protests the over-the-top accusations of "laziness and incompetence". PGI even did what people asked and devoted extra manpower from other areas to this problem to fix it.

[Mike Forst]

Kyle Lawrence, on 14 May 2013 - 01:17 PM, said:

Karl's been on the MWO team since the beginning. He just normally works on a separate system. His attention was simply diverted temporarily for this hunt.

Yep. Karl has been on MWO from the start.

[Deathlike]

Here's some stuff to think about for the "Debbie Downers":

I have no love for the "balancing" some of the higher up do when it comes to this game. It is mind-boggling and no PR can fix that disaster.

However, when it comes to bug fixing, there is no PR involved. The devs that respond to these issues are usually ones that are hacking at the problem, without much PR or a filter to them. They provide meaningful and insightful info... of which not all of us are able to make sense of, and that's OK. It is easy to whine and moan about the bugs, but they did actually fix some serious bugs that have occurred in recently memory... like the 4 FPS bug (with their memory "cleanup") and the major netcode update (and I don't mean HSR, which is still a work in progress). It is easy to "forget" how much we whine and bash them, and yet when they put in the time to address it, we complain it's not good enough? Seriously?

Anyways, I appreciate any info that comes from the mouths of the devs that know what the hell they are talking about and we should be supporting them. If you want to whine about something, there's are many threads for them (like ECM, 3rd person, weapon balancing). This is not the place for the other BS.

[Rawyn]

I dink ma bwain asploded...

But thank you for the interesting read and detailed post. And for nailing down that bug.
Unreproducable errors and race conditions. You just have to love them.

How many table boards were harmed (i.e. bitten, smashed, ... ) on that bug hunt ?

[FireDog]

Thank you so much for all the extraordinary efforts... I learned a long ago it takes a special knack to troubleshoot thousands of lines of code and the results they collectively produce. From that write up I think your team has the knack. Looking forward to the fruits of your labors. Take Care...

[Shumabot]

Rebas Kradd, on 14 May 2013 - 01:30 PM, said:

No, it doesn't. It simply protests the over-the-top accusations of "laziness and incompetence". PGI even did what people asked and devoted extra manpower from other areas to this problem to fix it.

These bugs have been existent for the games entire lifespan. They pulled people from other areas to fix it recently, and because the people who were working on it couldn't get it done. Uninformed accusations of sloth and incompetence are as easily supported now as they were before. Such accusations are rarely fair to the people doing the actual work of building up the game, it takes a lot of people to assemble a project like this, but this game is not exactly doing well for bugs. Even after this fix (9-12 months after the bugs became "just how mwo is") there are still many, many optimization issues and glaring bugs.

[INSEkT L0GIC]

I'm picturing Devs rejoicing with lots of mouth kissing followed by an awkward letter from HR about inappropriate touching.

[Darius Deadeye]

More feedback like this, and you'll have the community licking the sweat out of your palms.

Thanks for awesome!

[Zynk]

Next time you give an explanation for a bug please use the K.I.S.S. method and give me the CliffsNotes in the first paragraph with the engineers version after.

[GroovYChickeN]

Holy crap!

Dude from one programmer to another. Kudos! If I had to deal with all of that I would have broken a number of mice and most likely put my fist through my monitor.

Hammerfinn, on 14 May 2013 - 11:19 AM, said:

Can we post this in every thread where people complain about laziness or incompetence at PGI?

^^^ Also This ^^^