Yeah 150mb of video is why I said use Youtube. Not dling all that just to test something.

Took me half an hour to upload it. It should be just a little while to download, not like hours.

heheheheh people are sorta spoiled these days... they don't recall the days when you connected to your BBS over the household phone line and downloaded your file on your spiffy 28.8k modem, and were just glad you didnt have your old 9600 any more. And then, you saved your new file to a floppy disk for safekeeping.

I'm downloading it, should be done in like 5 minutes then I will give it a shot.

Well, crap. I didn't need to download it except to watch the videos (which show the problem quite well). I have already tested this exact cell. And, obviously I can vouch that it indeed works/fails as described.

I was BBSing on a 1200 baud modem son. Trust me, I know the pain.

Havent tried the esp's yet but watched the video. In the geck video the navmesh isn't finalized and no cover edges are made. I think it is misleading in this from.

Yeah it acts just like you sayed it was. If the path between me and fawkes is anything but a straight line he gives up on following me and sits there until I am in a position where there is a straight run to me, then in the fixed one he will come to me regardless of the path required. I will mess with it some more later but I think that is really all you need to know from me.

Since we're on the subject, I have had access to a full T-3 line at 45Mb/sec, now I don't so I am spoiled. But in high school we had a 400 baud modem, and a tape cassette "drive"

Hey, sesom is on to something, I just opened GECK and loaded the broken .esp then went up to the navmesh drop down, did the find cover thing and then finalized my navmesh then saved and it seems to work fine in game now.

yeah, I had deliberately stopped at the point where it was making the problem, I didn't finalize or nothin.

I'm gonna do a quick experiment with re-finalizing the navmesh in my bigger, known-problem cell (I have not touched that beast in a long time). Would be interesting if this was actually fixed in 1.6/1.7 but it requires finalizing the navmesh for the fix to take effect.

Hmm. When i originally tested , I did use the find cover edges and finalize and the bug appeared. hate to beat a dead horse, but did you enter, leave, fast travel, re-enter and then your follower still followed. But I'll try out what you just said to see if I can get it to "work".

Well, the bug manifest for me when I did it with the zipfile. Here is what I do.
I already have Jericho as a follower. I enter the test cell (after using find cover edges and finalize cells in the GECK). I walk around a bit until Jericho moves towards me. I exit the Test Cell; I fast travel to Megaton with Jericho. I walk around a bit until Jericho moves towards me. I fast travel back to Vault 101 and re-enter the test cell. Jericho stands still when I hop over the counter.

In Phalanx' custom cell which is exhibiting the problem, re-finding the cover edges and finalizing the navmesh does not make the cell behave properly when I run the test gambit with Fawkes.

Let's see how the reports come back regarding finalizing the small example cell in the zipfile - - -

Ok, never mind, the first time I only had to exit the cell and then re enter before it manifested but once i went to RC and walked around for a minute then returned so did the bug.

Aite cool - - I tested my copy just now, and the cover edges/finalize didn't fix it.

Sounds like this is a go, then.

Just for the record, you do not have to leave your follower in the cell to get the bug, this can save a few keystrokes when testing if you like.

Ok! I did my tests. I used my own follower for faster testing (its a simple follow package, in a token). He follows, guards or sandboxes switched per dialogue.

I had no troubles with the nonbroken version. The follower was able to follow after my fasttravels and jumping.

After wacthing the video i made the cover edges and finalized the navmesh of the broken version (to be sure) before i tested it. The bug was there (i tried follower sandboxing and guard) after the fasttraveling, follow me and jumping the follower stucked.

Then i edited the nonbroken version in geck and added 6 more triangles. Made cover edges, finalized, closed the navmesh editor and saved in geck (my daily navmesh routine).
Result in game: (i tried follower sandboxing and guard) After the fasttraveling, follow me and jumping the follower stucked.

To make sure i optimized the navmesh of the now broken version in geck (this has helped a lot with strange navmesh errors in my own cells in the past) made cover edges, finalized, closed the navmesh editor and saved.
Result in game: (i tried follower sandboxing and guard) After the fasttraveling, follow me and jumping the followers stucked.

I think with this files it should be possible for a programmer to find the error.

maybe make a new vid with the find cover edges and finalize done just to prove that it isn't an issue??? Your call though.

The zip file is rather slow to download so I'll put the question to those able to easily reproduce the issue: when you type "ToggleNavMesh" (or tnm) into the console is the navmesh missing or is it there and being ignored?

I get CTD when doing this.

I tossed up a version of it without the in-game video (which was the huge one). It's at http://www.finhosting.fi/~fallout/downloads/HowToBreakNavmesh-small.zip. It's 11 megs or so instead of 150. The big file is still on the server too, if anyone wants it.



Debug-related console commands have never worked for me.. I'll go try that one and if it does anything useful I'll post back.

edit: "tnm" at the console is an instant trip back to your desktop and game goes byebye.

tnm is a instant crash to desktop for me. I tried it with a clean save (without any mods, i do all my testings with clean saves) with the testcell loaded and without it. I think "tnm" is not available as console command in the retail version. But i guess the navmesh is there.

When the "break" happens, the mobs seem to act the same as though I had deleted the navmesh out of the cell. I can say this because in one test, I actually did this, and compared the behavior.

I did the same thing with the same results. Without being able to use tnm, I would venture to say that they are missing. Of course, how can one tell the difference between being ignored and missing anyway?

BTW did you notice that was a dev???

I don't think we can tell if its missing vs. being ignored with the tools we have, to be honest. That's why he was suggesting the console command to check.



eh. yes. =p