[v1.2.9.0][Bug] Some more bugs to squish!

[L.C.]

I just love squishing bugs. I'm running on a 1600x1200 resolution.

 

Operation Aborted --kind of fixed

http://www.hlrse.net/Qwerty/DCME_OperationAbort.gif

 

Runtime --fixed

http://www.hlrse.net/Qwerty/DCME_RuntimeError.gif

 

Default Window Size (Maximum) --> Restored Window Size --fixed

http://www.hlrse.net/Qwerty/DCME_Restore.gif

 

Radar Zooming & Navigation --fixed

http://www.hlrse.net/Qwerty/DCME_RadarZoom.gif

Radar Navigation Suggestion --done

You know how when you click somewhere on the radar to move there, it'll take you there on the grid in an instant frame, right? Have you tried clicking and holding in the radar to dynamically move the red box in the radar around (and dynamically/realtime navigating on the grid)? It kind sucks how I have to click on the area on the radar that I want to view instead of clicking, and dynamically dragging myself across the grid (literally) to wherever I'm dragging the red box. Know what I mean?

 

Tool Options Bar Suggestion --fixed

http://www.hlrse.net/Qwerty/DCME_OptionBarAnnoyance.gif

It's very annoying to have this pop in and out every time I switch tools. It's quite bothersome and distracting. Please make the bar always be there, without any options if you have no tool/a tool selected that doesn't have any (empty bar).

 

eLVL / Regions Suggestion --will be done

Layers? Eh? I'm sure you've used Photoshop to know how its layer system works. You can hide/unhide specific layers as you wish. For DCME under the Radar you should make a list of all regions with a button having a lightbulb picture on it. The lightbulb would be yellow and appear to be on, and when the button (or lightbulb) is clicked, it dims. To note, the button would have the 3D-like shaded borders to make it easy to distinguish that it's actually a button (and not an engraved icon into the background). The space underneath the radar is pretty precious. You should probably put it to good use for some things - maybe the same way the Tool Options Bar works - by default it would be an empty textbox, and when an option with more specific properties is selected, its special options appear in the list.

 

Map Border Visibility Suggestion

Actually it might be more convenient if the user can see maybe like 3 to 5 completely blank (no gridlines here) units outside the maximum map border lines. It would make it a little more comfortable to map around the sides of the map, really. It's very uncomfortable to use the selection tool when selecting patches of tiles adjacent to the border of the map and extends offscreen (towards the center of the map, for example), as I can't really see the border and it makes me feel very uncomfortable for some reason.

 

Rendering Suggestion

If you plan to add in an OpenGL/D3D renderer (for in-editor testing of your map [flying ship around, etc]), might be nice (and half useless) if we can tilt the grid to an angle to get a cheap sideview perception of the map. Might be useless, but I don't think we'd mind it. Maybe we can get some ideas from having this feature (never know unless we have some time with it). An OpenGL/D3D rendering system would be perfect for LVZ Editting and placement.

[Drake7707]

Another bug is when you select something on the pixel level, then move it then apply the selection if you select it again, then if you move it again, it won't be on the correct spot anymore.

 

 

I have no idea why this:

parent.picseltemp.Cls
Call BitBlt(parent.picseltemp.hDC, totalmoveX, totalmoveY, 1024, 1024, parent.pic1024selection.hDC, 0, 0, vbSrcCopy)
parent.pic1024selection.Cls
'parent.picseltemp.refresh
Call BitBlt(parent.pic1024selection.hDC, 0, 0, 1024, 1024, parent.picseltemp.hDC, 0, 0, vbSrcCopy)
'parent.pic1024selection.refresh

is in selection mouseup, but its totally unneccesary, because it was already moved on the fly in in mousemove, if you add the totalmove AGAIN you'll move 2x (its not onpreview in mousemove, it actually moves)

 

DCME window being out of screen when not maximized is fixed

 

DCME can't abort operations because it didn't held the state before doing the operation on selections, you tried selecting everything around the line instead of the line itself

 

 

And some more weiiiiiiird stuff

If Not selectionmoved Then
	selstate = notdrawing
	Exit Sub
End If

in sel.mouseup

 

This effectively screws up anything trying to select & move with magicwant, makes my selection box go nuts, and i completely fail to see for what it's used.

 

Same goes for selectioncancelled... there is no way that can ever be true, because it's a single threaded process, if you click somewhere, while it is still in some method doing stuff, it gets added to the message queue, vb checks the queue when it goes back to the main loop behind the scenes. You can't abort an operation unless you call DoEvents to give the message a chanche to get sent, so the message can get through

 

 

For the toolbar, i wanted to have an optimal space for the map itself, so i hid the toolbar when it wasn't used for anything (i'll make an option to pin/unpin the toolbar)

 

Rotation bug has been fixed (the newboundaries that were used to redim the tmpobjects array changed afterwards because of out of map checks, tmpobjects are now redimmed after the newboundaries are final)

[Samapico]

Radar Navigation Suggestion

You know how when you click somewhere on the radar to move there, it'll take you there on the grid in an instant frame, right? Have you tried clicking and holding in the radar to dynamically move the red box in the radar around (and dynamically/realtime navigating on the grid)? It kind sucks how I have to click on the area on the radar that I want to view instead of clicking, and dynamically dragging myself across the grid (literally) to wherever I'm dragging the red box. Know what I mean?

I remember trying that... it screwed up everything But I didn't make many experiments... At the moment, the radar view is always a 1pixel:1tile scale view of the map... maybe it should be a scaled down view in which we see the whole map ... or maybe it should be that only when fully zoomed out... hmm

 

as for the 'selection being moved twice' ; I remember making it so the bit array is only moved at the end, I don't remember playing with the bitblt of that selection at pixel level.

 

 

As for the very first bug (selection operation aborted), it's simple... you just didn't click on the tiles... hence selecting the whole black area, and when DCME selects that many tiles, it aborts to avoid freezing the program for a couple of minutes (same applies for Fill Bucket)

And if you zoomed in to see what area was actually selected (the yellow rectangle never shows accurate selection limits for the moment) , you'd notice the selected parts spread out from where you clicked towards the outside.

However, that limit was mostly for the fill bucket, cause filling that many tiles definitly causes huge lag... i'll probably raise the limit for magic wand

[Samapico]

Made the same optimization (sub instead of goto) for magicwand than you did for bucket

Just made the test, the execution time is exactly the same though... but the code is cleaner

[L.C.]

Flipping Bug. --fixed

http://www.hlrse.net/Qwerty/DCME_Misflip.gif

 

Origin & Selection Bug.

http://www.hlrse.net/Qwerty/DCME_OriginCoord.gif

 

Clone & Selection Bug.

http://www.hlrse.net/Qwerty/DCME_MoveBug.gif

[Drake7707]

it must be because of the unusual warm temperature in november... because the bugs seem to keep coming

[Witchie NL]

lol

[Samapico]

flip / mirror bug fixed

[Samapico]

origin & selection bug animation goes way too fast.. cant read a thing... zip every frames individually or something I don't have anything to split frames right now

btw I removed the images / animations of the things that were fixed to make the thread a bit lighter.. I hope you don't mind

Edit: actually, I changed them all to URL tags... it makes reading the thread much more easier. In the future, please post your images with url tags (or attach them directly in the thread).

[Samapico]

As for clone and selection bug... the #2 bug (undo) is a known bug and it was reported 134516 times already... Anything related to the selection will NOT undo correctly. We will eventually redo the whole thing... cause the way it is now, we can't just patch it easily.

 

And I don't really see the problem with the #1 bug in there... You select the whole line, then unselect half of it with CTRL (yes the yellow square stays the same because it only indicates the boundaries of the selection), and when you move your selection it splits the line... that seems normal to me.

It seems to be cloning when you undo because undo doesn't affect what's in your selection... so it restores the tiles that were removed, and what's still in your selection stays there. If you had dropped the selection, then undo, it would have undo'ed correctly.

[L.C.]

And I don't really see the problem with the #1 bug in there... You select the whole line, then unselect half of it with CTRL (yes the yellow square stays the same because it only indicates the boundaries of the selection), and when you move your selection it splits the line... that seems normal to me.

Well if I'm selecting half of a selection, and it ends up being the part that is moved, isn't it more logical that the yellow selected box updates itself to show correctly what it will be really be moving?

 

http://www.hlrse.net/Qwerty/DCME_OriginCoo...SplitFrame1.gif

http://www.hlrse.net/Qwerty/DCME_OriginCoo...SplitFrame2.gif

[Samapico]

Selection doesnt appear correctly at pixel level.

Yellow box only indicates the boundaries of the selection. If you remove parts of it with Control key, it will not show it correctly. For now, the only way to know what is in and what is not in the selection within these boundaries is to hover the mouse over it with the selection tool. If the cursor becomes the "move" symbol, you're over selection. If it's the standard cross, these tiles aren't selected.

 

I know it sucks... but fixing that will require some major tweaking.. If I remember right, there was a problem with drawing the selection limits like its done at normal zoom level. Drake did that part though, I'm not sure how it could be fixed...

[Samapico]

Not sure to understand the first thing mentionned in there:

Imagine it this way: I create a new lvl file. So far I've done nothing but sit there with a new map open. I look at size and it sisplays a huge number [......]

Size isn't displayed on a new map... you can only the position... I'm not sure what you mean there...

 

And for the 2nd part... the only thing I notice is that the "From:" values aren't updated when moving an existing selection.

 

The "Size:" values are hidden as soon as you change tool. And when clicking back in the selection area, the "Distance:" is shown and updated as 0 - 0, even without moving... i don't see any problem...

[L.C.]

It's hard to explain that bug ^^. That's the best I could do. Maybe someone else will understand me and report back the same.

[Samapico]

I think I understood right.. but it just don't happen to me when I follow the same steps

[L.C.]

Well the bug with where the origin of this unusual invisible selection of which you can locate the coordinates of when "Size:" displays 1x1 instead of some huge number* - I don't know how I did that.

 

*This would be that you suddenly notice this odd number still being display and continuously updated as you work about, and !@#$%^&*uming that you are a few hundred units away from where the origin coordinates of this invisible selection is.

 

As according to that one screenshot, my mouse is in the top right corner hovering over the grid. So imagine as if you had your mouse there too with the same tool selected as I. You would look down to see "Size:" displaying the same (or similar dimensions) and start wondering where "1x1" would be. So you move your mouse around to see which way is closer to this origin, and eventually you find it a bunch of tiles southwest of where you have your mouse over the grid.

 

It's like a 4-Quadrant grid with (0,0) being the origin of course. We will use this as another example to help us imagine what I mean. Color the horizontal and vertical origin lines with green. Now we will relocate this origin somewhere else on the grid (while leaving behind a ghost copy of a green color origin 'crosshair' on the grid; this is our original origin location that we started this imagination with), perhaps 63 (random number) units southwest. We will color this one red. Now green doesn't exist any more, but we can see where the origin originally was. This is basically what it's doing.

 

However in DCME, I know I originally had an active and visible selection on the grid (green origin). After the bug would mysteriously go into action, at this point I would no longer have a visible and active selection display on the grid anywhere (I'm thinking, "Yup! No selection in sight, move on..").

 

Combine the last 2 paragraphs into one.

[Samapico]

I'm currently working on the radar navigation problem...

That's how it would look like when fully zoomed out:

 

I was thinking keeping it as it is for all zooms except pixel level... But seeing how nice it appears, there might be eventually some radio button under the radar for : Full map view or Focus View (or whatever it could be called)

 

Only problem now is that alot of functions need to be redone... we need a SetFocusAt(tileX, tileY) function because now all 'goto' functions (bookmarks and such) simulate a click on the picradar and move to the correct coordinate that way.... So all that needs to be redone.. Drake maybe you could help me with that, it shouldn't be too hard, but you're the one who coded the picradar_mousedown, so I'm not sure what I can reuse in there ...

[Samapico]

Radar is now improved and much cooler

[L.C.]

Found two more bugs, one of them I can easily explain. Actually, I can't seem to retrigger the other bug. :\

 

First Tile Bug --fixed

Either select the Filled Rectangle tool or plain Rectangle tool. Anywhere besides the first square or unit (Coordinate: 0,0) click and hold your mouse. Then drag all the way over to coordinate (0,0). Can you see the render problem there? There's also another problem I've experienced with this rendering bug; I don't know how, but for the two maze arena's I made for SSCI Dragonball Z, sometime after their early releases I saw that the squad (0,0) was never filled in, but everything around it was. Doesn't appear to do that now. :S

[Samapico]

rofl.. nice one... fixed

 

we had:

If endlinex = curtilex And endliney = curtiley Then
	Exit Sub
End If

but guess what... endlinex and endliney aren't even used ANYWHERE in the program... so they were initialised at 0... heh.

[L.C.]

You ought to make a scoreboard chart of bug reporters and listed by top number of unique bugs reported (meaning they can be retriggered). Perhaps I may be one of the first 5.