Couple of things that went wrong ... somewhere between 5.0 and 5.1.5

[cal-linux]

I’m actually unsure when these “new features” / “changed behaviour” were introduced. But anyway, at the very least I’d like to cast my vote that these, if intentional, were horrific ideas.

(1) Drawing a polygon or a zone (or a hole/cutout in a zone). It used to be not-so-good, in that the highlighted drawing was barely visible. But then, this was changed to a thick bright-white line that makes it almost impossible to see what you’re drawing unless you have a really coarse grid. If you zoom in enough (possibly necessary if your grid is fine enough), the line can end up filling up the whole view. It is essentially impossible to come back to the exact starting point (to complete the drawing). The highlighted cursor (the circle superposed to the crosshairs, that indicates that you’re exactly at the endpoint of another line, and could be the visual hint that we can click to complete the drawing), if activated, it would not be visible because the thick white line will hide it. It is horrific! I’m having the hardest time trying to imagine how someone could have thought this was a good idea!

(2) When trying to make a T connection by snapping the end of a currently-drawing trace to the middle of an existing trace, the connection point causes the existing trace to “bend” (well, it splits it into two, and now the join point drags and moves the three lines. If the line is diagonal or in any case not exactly on-grid, then it is impossible to leave the original trace as a straight line. This screenshot shows the issue. At the bottom, I had a perfectly straight horizontal trace from leftmost via to rightmost via. Then, I draw a trace starting at the top via, and when the cursor touches the horizontal trace at the bottom, it breaks it and bends it:

Now the cursor drags the three traces, and because of the grid, I may not be able to leave the horizontal trace as it was (as a single trace).

Notice: this happens even if the original horizontal line is locked.

Suggestion for (1): I don’t oppose to the white line; in the old way, the zone that one was drawing was barely visible; but the white line should be no more than 1 or 2 pixels wide (well, or 1pt or 2pt, if that addresses any issues with super-high-resolution monitors). I know it’s not a “life or death” kind of misfeature. I can always edit the zone’s vertices. But… c’mon — something bad/undesirable should not be accepted just because it is not a fatal misfeature.

Suggestion for (2): Please remove this abomination! Again, I would like to think that this was an unintended side-effect of some other change / bugfix / added-feature. Because with this, I really, really have a hard time seeing how anyone could have thought this was a good idea. (the fact that the feature does not respect the locked state of the existing trace makes me suspect that it was not an intended feature)

Maybe someone could try to convince me otherwise (for either of the two “misfeatures”)?

These two “misfeatures” have been present for a while; but in any case, they are present for me with KiCAD 5.1.5 on Ubuntu 18.04:

Application: KiCad
Version: 5.1.5-52549c5~86~ubuntu18.04.1, release build
Libraries:
    wxWidgets 3.0.4
    libcurl/7.58.0 OpenSSL/1.1.1 zlib/1.2.11 libidn2/2.0.4 libpsl/0.19.1 (+libidn2/2.0.4) nghttp2/1.30.0 librtmp/2.3
Platform: Linux 5.3.0-42-generic x86_64, 64 bit, Little endian, wxGTK
Build Info:
    wxWidgets: 3.0.4 (wchar_t,wx containers,compatible with 2.8) GTK+ 3.22
    Boost: 1.65.1
    OpenCASCADE Community Edition: 6.9.1
    Curl: 7.58.0
    Compiler: GCC 7.5.0 with C++ ABI 1011

Build settings:
    USE_WX_GRAPHICS_CONTEXT=OFF
    USE_WX_OVERLAY=ON
    KICAD_SCRIPTING=ON
    KICAD_SCRIPTING_MODULES=ON
    KICAD_SCRIPTING_PYTHON3=ON
    KICAD_SCRIPTING_WXPYTHON=ON
    KICAD_SCRIPTING_WXPYTHON_PHOENIX=ON
    KICAD_SCRIPTING_ACTION_MENU=ON
    BUILD_GITHUB_PLUGIN=ON
    KICAD_USE_OCE=ON
    KICAD_USE_OCC=OFF
    KICAD_SPICE=ON

[Rene_Poschl]

Point 2 is a known bug in 5.1.5. I think it is already fixed in the pre release but am not able to check right now.

[fred4u]

I can confirm that Issue#2 is fixed in the Pre-5.1.6

[eelik]

Number 1, see https://gitlab.com/kicad/code/kicad/-/issues/3724 and give a thumb up.

[cal-linux]

Doesn’t let me (I guess I would need a separate account for that system?)

In any case, a suggestion for @Seth_h who seems to be the developer assigned to the issue.

As the main issue is the thickness of the lines, then:

1. You could use a look similar (even identical) to the ratsnest lines — in any case, a fixed thickness (very low)
2. As a related aspect: you coud consider the grid size; for example, if you rather keep it thick, then the grid size is really what defines what “too thick” is. I would perhaps suggest that more than 1/4 of the grid size constitutes “too thick”

[paulvdh]

A (temporary?) workaround for the zones is to always set the grid to very coarse when drawing zones and board outlines. I’m a european guy so I always want boards in whole millimeters anyway.

I also noticed Issue 2 it’s indeed a bit irritating. I was confident though that it would be fixed without me complaining about it. (I already complain too much). So, I’m glad to hear that it is part of the continuous improvement of KiCad.

The old way used to draw 2 or 3 track segments on those junctions which is also sub optimal.

The reason it happens this way is very likely because it is missed during testing, and this is also the workaround: As long as all your tracks are on the grid it does not happen.
So if you’re working with V5. It helps if you route the whole board on the same grid. The Interactive router does not care about your grid, it just makes sure clearance values are met. It may help to ensure that:
trackwdith + clearance == grid.

Also, if you ignore the ugly lines (I know, it’s hard) then, as soon as the Interactive Router bumps something else into the ugly slanted tracks and pulls them straight again. I use this trick as it’s often quicker to drag a via or track segment around to clean up neighboring tracks than to straighten the track itself.

[cal-linux]

@paulvdh

A (temporary?) workaround for the zones is to always set the grid to very coarse when drawing zones and board outlines. I’m a european guy so I always want boards in whole millimeters anyway.

I normally see it more like “I’m from Planet Earth, so of course I want things in whole milimeters”

Kidding/ranting aside … Setting a thick grid as a workaround is only useful for certain situations. If you see the link above, posted by eelik, the author there points out a situation where the misfeature truly gets in the way.

The thing is: you cannot always rely on zones being on a coarse grid. Even for the board’s outline, your board may need to mechanically fit on something that does not fit in a coarse grid.

Regarding item (2):

@paulvdh

As long as all your tracks are on the grid it does not happen

That’s not correct. It still can happen. As long as the trace is thicker than twice the grid size, your cursor can fall within a point inside the trace that is not the centerline of the trace.

[der.ule]

1. You can set the transparency of that line to make the things behind it easier to see when dragging the zone (I KiCAD!)

Cheers!

[cal-linux]

@der.ule

You can set the transparency of that line to make the things behind it easier to see

This is indeed an interesting workaround! Thanks for pointing that out (already set opacity to 50% for Aux items!!!).

However, it doesn’t work that well (please point out if I’m missing something). For example, if I want to make a zone that covers exactly two adjacent pins in an SMT chip where the pads are rounded rectangles, it is hard-ish to click at the exact place; when the “translucent” line overlaps with the solder mask (at least with the default color), the line becomes practically white, and completely hides the crosshairs that help you place the polygon’s vertex at the right location).

Also, one of the issues caused by this “misfeature” is that when you want to complete the zone or pollygon, you need to click exactly on the first vertex. But then, the line that is already drawn is not translucent (as your second image above shows), so again you cannot see the marker that tells you that you are exactly at the position of the first vertex.

Still, thanks so much for pointing out this workaround!! I also love KiCAD, of course, but I certainly love this forum!!!

[der.ule]

2. If you right click on your last point you can select “Close zone outline” to place the last line, no fishing for that last spot anymore

1. I do not use the default colors unfortunately so I cannot test this, but as you see in the screnshot, it looks quite ok, maybe you don’t need to make such an exact placement if you can play with the parameters of the zone itself (clearance, min. width , etc. etc.)

[cal-linux]

@der.ule

If you right click on your last point you can select “Close zone outline” to place the last line,

Man I love this forum!!!

I was not familiar with this feature — thanks for pointing it out!

Let’s say that I’m sold on the low-urgency to fix this “misfeature” (the thick lines when drawing the zone). I still claim that it is still a misfeature and would be good if it gets fixed — not all users know of this workaround, and won’t necessarily stumble into this thread to learn about it.

Again, my suggestion for a fix is: just make the white line thin (e.g., exactly like the ratsnest lines look).

[eelik]

@der.ule

If you right click on your last point you can select “Close zone outline” to place the last line

Or you can just double click.

[cal-linux]

@eelik

Or you can just double click.

I seem to recall double-clicking (not sure for creating zones; more like when drawing rectangles, either on the silk or edge layers, or on Fab or CrtYrd when creating footprints). But I thought that the cursor needed to be exactly at the position of the initial point when creating a closed polygon. (in all those cases, the “circled crosshair” appears when the cursor is at the correct position)

[eelik]

You can double click on the last point, there’s no need to go back to the first point. See for example the der.ule’s screenshot - instead of opening the context menu you can double click.

[paulvdh]

Does not compute.
On my KiCad V 5.1.5-52549c5~86~ubuntu18.04.1 Zones are drawn as polygons, which are always closed.

On the Edge.Cuts layer I can apparently choose between lines or a polygon.

[cal-linux]

Sure. When I said "I thought the cursor needed to be … " I didn’t mean to argue or disbelieve your statement. It was more like a “historical fact” about my understanding of how the thing works…

[paulvdh]

Just before I read this I wrote a quite long part on opacity settings today, and some ways in which ergonomics of viewing stuff in Pcbnew could be improved. It’s in:

Two questions about filled zones on PCB

PCB’s are crowded places, and coming up with a good algorithm to clearly show what you want to see, but also see other stuff to keep orientation can be an extensive study in GUI design. I had some trouble adjusting from the old “xor” drawing to the way it works now. Back then I thought the “viewability” of the old way was better, maybe I just thought it was better because I was used to it. The “xor” way was dropped because it had to be handled by the CPU, while transparency is done by you video…

[eelik]

We don’t argue in this forum After der.ule’s and your messages I thought you didn’t know that a polygon can be closed by double clicking on the last point (without going back to the first point) which is identical to Close Zone Outline action.

BTW, there’s no Close Zone Outline for non-zone polygons, but double click works.

Please move on, folks, there’s nothing (arguing) to see here.

[cal-linux]

@eelik

We don’t argue in this forum

No, of course. I completely agree with you! :‒)

[der.ule]

That is too easy, an takes away from the excitement of finding “the secret menu” !

(still not arguing … )