Seeing through a wall

[Griogre]

I was building some simple map tiles in D&D 5E and checking them and noticed an error where you can see through an ocular wall. I've attached a screenshot. It's in a simple dev campaign that I have attached too as well as the image's vtt file zipped. The image is a DungeonDraft VTT import and its image and XML are in the campaign image directory.

Edit: I meant to mention FGU imported the VTT file.

[LordEntrails]

I can't reproduce. i.e. when I open your example campaign and the map and turn on player vision preview, there is no bleed for me. Were you seeing this on a second instance or did you do something different?

I'm not sure what to say. Hopefully someone else can try and reproduce the issue. Or maybe one of the devs has more insight.

[SilentRuin]

JarlsmootSeeThrough.png

I've been annoyed by see through walls occasionally also. Most recently this last monday where Rhime of the Frostmaiden Jarlsmoot map had the treasure rooms permanently sealed off from view with giant walls. I removed them (marked in orange squares) by removing key vertices in them. Then my players asked - what are all the chests? Yes, they could see right though the remaining walls seeing what was inside. Super annoying. On the fly during the game I added double walls to the two rooms containing chests but this is not first time something like this has happened. I suspect something crazy is going on that is not evident from LOS graphics walls and vertices but I'm damn if I know what it is.

[Zacchaeus]

I was able to reproduce after moving the token about a bit. I believe that this is a long standing issue which no-one has been able to identify but it seems to be related to when the occluders are drawn to exactly match the grid. I fixed it by redrawing the occluders so that it's not exactly on the grid. This isn't something that can be reliably reproduced. (also I exaggerated the redrawn line - it doesn't need to be that far away - in fact most of the time just redrawing it is enough).

[SilentRuin]

I was able to reproduce after moving the token about a bit. I believe that this is a long standing issue which no-one has been able to identify but it seems to be related to when the occluders are drawn to exactly match the grid. I fixed it by redrawing the occluders so that it's not exactly on the grid. This isn't something that can be reliably reproduced. (also I exaggerated the redrawn line - it doesn't need to be that far away - in fact most of the time just redrawing it is enough).

I think if the dev's know that it can sometimes be reproduced by matching the grid lines they have enough of a clue to rifle through the code and see where that might be an issue.

[Griogre]

After looking at the replies I went back and reloaded the campaign and everything was fine at first.

Then like Zacchaeus, I moved the token around out of the stairwell area up the north passage and back to the stairwell and the token could see through the wall again (I used the arrow keys for what it's worth). I originally found this by moving the token through eight connected tiles because I had made a bunch of tiles and revised some of them and wanted to make sure the LOS still matched the images.

When I noticed this tile had an LOS issue I created another map just using the tile and it still had the same issue so I reported. So it looks like its something that happens when the token moves and causes a re-calculation of the LOS.

[pindercarl]

I was building some simple map tiles in D&D 5E and checking them and noticed an error where you can see through an ocular wall. I've attached a screenshot. It's in a simple dev campaign that I have attached too as well as the image's vtt file zipped. The image is a DungeonDraft VTT import and its image and XML are in the campaign image directory.

Edit: I meant to mention FGU imported the VTT file.

I am able to reproduce this issue with the campaign provided. I'll see if I can find the root cause.

[pindercarl]

The issue is not that the occluder segments are aligned to the grid, though it does appear to be caused by the perpendicular intersection of the occluders. For "reasons," segment 1 test in front of segment 2 (see attached image), so the point of intersection in the middle is being skipped. If you select any of the points on that wall and nudge it up or down (use CTRL+arrow key), it will fix the issue. I'll continue investigating the root cause.LOS_error.jpg

[Griogre]

Thanks for looking into this, Carl and everyone else.

[jharp]

This issue has been reported before (by me) and identified before but has never been fixed at its root.
Jason