Blogs - bjorn98009_91

Time for my weekly update.

This week we've continued to make progress on our branding and now we've got the launch trailer and the entire soundtrack uploaded to YouTube awaiting unveiling. We've also made significant progress in the bug department, and right now we only have two maps that's bugging, none of them very severely.

The maps with issues are 06_HongKong_Helibase and 15_Area51_Final. The problem with Helibase have for a long time been a long standing occlusion issue. This is because in Unreal Engine 1 geometry occludes itself, that is the engine checks if the polygon is in view and if not it doesn't draw the polygon (to save time). The problem with that is that the code that checks if it's in view is not 100% accurate (if it was it would take too long to run and it would defeat the purpose of the check), thus sometimes at the edge of the screen the engine thinks a polygon is out of view when in fact it isn't and you can still see a part of it.

When geometry is built the polygons are cut up in a number of pieces and it's these cuts that the bounds checking code checks. With an unlucky collection of cuts this can result in many polygons flickering out of view and sometimes not just of the edge of the screen but a bit further in as well.

The issue then becomes how to affect how the level is cut up, and there are a number of different tricks one can do. It's a lot of trial and error and one change somewhere in the level can affect how cuts travel across the level and thus issues can either appear or disappear with one change. This of course requires extensive re-testing, even though I as a developer can check the target area to see if it works or not; other issues might have appeared elsewhere and I can't re-test the entire map each time I think it's good. I’m just glad that we still have a tester that is able to very rapidly and thoroughly work through this.

Earlier today I discovered another non-occlusion related issue in Helibase, hopefully I'll be able to fix that soon, but as always one change can undermine the fix for the occlusion issue I worked on before. In Final we have a similar occlusion issue, however this map also have a few other non-occlusion issues as well.

Progress for me has been impeded slightly due to an exam tomorrow, but tomorrow evening I should be able to get on this and try and sort the issues out.

Alright, you requested updates so I'm going to try and make weekly ones from now on.

Over the past few weeks some of the team members have been busy with various real life commitments, and that's just how mods go or any project that depend on people's free time. However this week people have begun to come back and be able to dedicate more time for Revision, and that's awesome.

We've made significant progress about our internal branding and with bugs. Currently we have 3 BSPs left that we are working on (one of them is fairly new). I've also decided implement some small stuff that we've had slated for a post-release patch like some visual improvements to one of the end cutscenes or adding some details to a map that was already acting up and we had to edit anyway.

We are also wrapping up things with the Eidos Montréal team, we've got a fixed release date, and there's a full announcement coming.

Hi!

Was a while ago since I last wrote. I've been busy as always. Drewie wrote to us a while back when some commentators asked about Windows 10 compatibility in the comment section and asked if we wanted the mod tested on Win10. Since I only have one computer, and I don't want to risk screwing things up with my development environment before we are done, I couldn't upgrade myself and test how Windows 10 behaved and I was glad that he could help out (another tester also contacted us but he only did a quick test and then rolled back to an earlier OS version).

So Drewie started a new playthrough and of course found a lot of issue, as always. The more you look the more you find, and also since we rebuild the maps from time to time there is also the risk of creating BSP errors that one might not find on a quick test. He also have loads of suggestions, and while we might not agree on all of them, he has given us plenty of ideas of ways to improve things that we've taken to heart. Larger improvements like doing geometry changes we don't want to do due to the fact that it takes time and also the risk of BSP errors and then extensive re-testing. Fixing a floating pillow of removing a guard for a difficulty setting take about 10 seconds once you have the map loaded.

What's strange is that some bugs we are not even sure how they've arisen as we've not even changed that aspect of the game since the last time we changed the map (like Gunther not showing up for the cutscene in Champs-Élysées). I've played through the game and Harry have played through it five times now I think, it has always worked; except now, despite that we've changed nothing about him.

So of course I'm keeping busy with fixing all the bugs. Someone mentioned if we've given this extra time has decided to take some 1.1 things and implement them into 1.0. For the most part, no. Mainly cause we've been busy with fixing issues/testing and we simply haven't begun work on the 1.1 stuff. However sometimes when bored, a team member or two might make some new thing and we go "hey, why not put it in 1.0, much easier to manage then keeping some separate version of the file and it seems tested enough".

Also some known hard to track down non-critical bugs that we've previously pushed into 1.1 got fixed given the extra time we've had to test, try and debug. Most of them are vanilla/gameplay mod related issues, so it's not exactly our fault anyway, but they do sometimes detract from the experience so if we have the time and the possibility to fix them we try. Sometimes the hack fixes some mods have implemented to get around certain issues just aren't cutting it and then there is no way to sort the issue, short of rewriting the entire subsystem, like the inventory system; and that's not happening any time soon.

If the bug/suggestion reports we get doesn't contain a lot of BSP issues we'll have most things fixed up in a day. John checks the suggestions, decides what he things are good ones that he considers fit, give feedback and then I implement/fix. We got a fresh report in last night so I'll be looking through that tomorrow to fix stuff. Today has been a talky day with emails going back and forth, real life things (socializing, watching the news, doing food/dishes, and we also had some trouble with some conversation files not being up to spec.

I'm also very excited that we are reaching the end now, or well the start of post-release era. Anyway, cya around!

Alright, so I'm done with the game. Finished tonight with all the endings. As usual there were the little nitpicks here and there, and A51 Bunker and Final had some BSP issues. Tomorrow I'm going to investigate some odds and ends, most of it directly related to Revision like some pathing issue or a small bsp on a mover and some more pure DX bugs. Have to continue to track down the issue with the conversation camera being slightly off, and I'll do some stuff I've put off before. :)

Just thought I should mention, we've wrapped up the issues with Vandenberg_cmd and the everett conversation as previously mentioned and we've managed to create a background to Biomod's challenge menu and implement it.

So I'm on track on continuing my playthrough, just some small things. First need to mess a bit with the RevSoundZone that is playing a sound every minute. Apparently the sound is paused when going into a menu and then restarted when exiting the menu, so that's a bit weird being that it's a PA thing. I also need to do some documentation about the Everett conversation stuff while it's still fresh and need to complete the refactoring of our custom DataVault images from RevsionUI to RevDataVaultImages.utx

This was done to reduce the filesize of RevsionUI that we tend to change quite "often" when we need to include some new graphics. It then makes no sense to have the DV images in when they are not being changed, as they took up around 90% of the size of the package. There are 192 DV images to import so I'm looking for some automated solution for making the import script, as I don't want to do copy/paste that many times (or actually that times 2 since I need to change two strings). I'm hoping han might have some quick cmd tool to use, otherwise I'll just have to bite the bullet.

Oh, also have some personal stuff. Astronomy lab thing due last of July, so I'm operating in a reduced manner for the time being to be able to fix that.

Hi, again sorry for the long absence.

I've been busy playing through the game and fixing stuff. I'm currently at the start of first Area 51 map. Most of the fixes were texture alignments and item positions and those are always fast to do, compared to the other types of fixes. Also recently started to notice that fragtypes on movers are incorrect a lot of times also a very fast fix though.

Things that take time is to explore each path of the level, like go the stealthy route, then the main route and so on, to see that they all work. Also experimental things take time, things you can't see in real-time, like Jaime was sitting weirdly on a chair, so had to adjust the sitpoint coordinates on the chair. Problem is that you need to restart the game each time to test that, so there was a lot of back and forth until I was satisfied.

However the main thing that takes time is if some geometry change is involved, like fixing a BSP issue or something else that requires a rebuild. Cause then you first must fix the issue, then rebuild the map a lot to get it optimized and then test, and since BSP changes are so unpredictable one change can cause another bug to pop up elsewhere and then it's back to square one. Sadly there were a few of these since last time.

Last night we got Paris Catacombs bug free since it had a killer BSP that I just discovered by chance, and currently we are working on Vandenberg_Cmd (that had some bugged brushes and required a rebuild) and on Vandenberg_Sub that had some BSPs. Hopefully we'll have those solved soon.

Another thing that's been holding me up is that a conversation with Morgan Everett about his mechanic was sort of broken in vanilla. I've been hard at work setting it up so it does indeed work properly like I envision it should work. When I jumped into the conversation file it was a mess, and it still is, but I think I've managed to fix it. A lot of logic going around and it's very easy to make a mistake and break everything. I'll do some testing later today to see if it's good.

Hey!

Sorry for not writing for a long time. During the last week we wrapped up that nasty BSP issue in HK Compound and did some optimization work on the Market level. Fixed a few typos relating to the new computer logos that prevented them from showing up. We added some new texts to a newspaper, a datacube and a book that a tester noticed were empty.

Continued playing through the 09 levels, that's Dockyard, ShipFan, Ship, Shipbelow and Graveyard. Mostly no issues, just the minor polishing treatment and some pathing adjustments. For some reason changes I made to ShipBelow previously had been lost, so was a bit frustrated about that and had to redo those.

Realized that the "Old sword" that we added way back in 2009 didn't glow as the NanoSword even though it should, so fixed that too. Sometimes it's old bugs you notice ^^ Also fixed a vanilla conversation issue at Tong's place, cleaned up some code that had "accessed none" warnings, and updated some in-game graphics with our new logo and font.

Continued on to Paris, and completed the Catacombs maps, that's Denfert Rochereau and adjacent maps. And did the tunnels part. Next up is Paris_Metro aka Champs-Élysées.

Yesterday I did a lot of playing, completed all of Hong Kong and fixed all of the bugs there, the usual mostly. Texture alignment/scaling issues and floating items or otherwise small tweakes that were made to items or incorrect fragtypes on some movers. Some bugged brushes and actually a small BSP in the compound map. The bsp is all that remains now, and it has proved to be a bit stubborn, fixing it prouced two new ones and fixing those resulted in a forth, and that then resulted in two new again. Well, that's how it is normally with BSP, we'll have them solved soon though.
I also updated the Chinese textures with the fixed one that use the correct chinese characters (thanks to the community here on moddb).

Today I wrapped up the 8th chapter, that is the return to NYC. Pretty bug free there, but that's mostly because they share a lot from the other instances of the maps, so if I find a bug in 02 or 04 I open up 08 to check if it's there too and fix it. There was some small things like always and worked on some other stuff not strictly related to the mod, like troubleshooting and fixing so HK infolinks works if DeusExConversations are rebuilt, if we would ever need that in the future.

So, stopped at 09 Dockyard. Will be away over the weekend, but I'm sure Harry gets the BSP fixed until I get back.

Alright, so yesterday I added code support for 2x computer logos and added in the new 1x and 2x ones. I fixed the issue with DXOgg (I was very stupid there before I realized). and added cheat prevention for rmode without cheats on.

Continued playing from 04 NYC, fixed a few texture and lighting issues in the bar levels including swapping to a custom curtain instead of the ugly low-res one.

Today I continued on to UnatcoMJ12Lab where I found some texture alignment issues, some placement issues and a few pathing problems. Fixed them all and carried on. Also fixed an issue with the wrong audio playing in InfoLinks in that level, not sure why that was. Also updated the datavaults with accurate chinese letters according to some help we've been getting.

Fixed a few floating items in 05 HQ, Sam carter now no longer fears gunfire and for some reason Anna had Gunther's killphrase in her computer. Carried on to 06 helibase where I fixed some more texture alignment issues :)

Stopped for the day as I arrived at 06 HongKong WanChai Market. So far this mod is very enjoyable, except for the times I stop and notice something. Hopefully when I'm done with all the maps nitpickers like me will find it constatly enjoyable too :D

Today I was busy for the most of the day, hang out with a friend and did stuff.

Finalized Ford's convo in 08 smuggler and then I continued my playthrough.

Did the 04 maps, found a few different things to fix in NSFHQ. Floating things, brushes whose surfaces have been "merged" due to a optimization issue in the editor, movers with incorrect pivot points, floating coronas. Of course I fixed all of them and continued on. Nothing else to report in the other levels, ultimately stopped at 05 UnatcoMJ12Lab.

Tomorrow I'll look into some code issues like the previously mentioned dxogg thing, the killing of paul if you bail out of the window, and the need to code up 2x support for the computer logos.