Bad line junctions sometimes

[Rmano]

Consider the following snippet:

\documentclass[border=10pt]{standalone}
\usepackage[RPvoltages]{circuitikz}
\begin{document}
\begin{tikzpicture}
    \coordinate(o) at (2,0);
    \draw (0,0) -- ++(0,1.5)coordinate(tmp) to[D,i=$i_{D}$,v=$u_{D}$]
    % problem here
    (tmp-|o)
       to[short,-*]
    (o)
    ;
\end{tikzpicture}
\end{document}

If you look at the result, you have:

Now, reverting this patch: 6efc2ee

seems to fix it. My problem is that I do not understand what the patch is doing... I will try to dig into it. Meanwhile, if @sistlind is here, could you shed some light? I have seen the bug report but I can't really understand it. Doing

  \draw (node) to[...] (whatever)

is expected to fail if the (node) has no border anchor anyway... or I am wrong?

I am trying to compile the manual with it reverted... let's see.

[Rmano]

So the original problem can be reduced to this thing:

\documentclass[border=10pt]{standalone}
\usepackage[RPvoltages]{circuitikz}
\begin{document}

\begin{circuitikz}
	\draw
	(0,0) node[npn] (npn6) {}
	(npn6.C)++(0,1) coordinate[](P4)
	(npn6.B) to[short,l=LO-,-o] ++(-0.5,0)
	(P4) to[R,l_=$R_2$,-] ++(0,2) node[vcc]{}
	;
	\draw
	(3,0) node[npn] (npn6) {}
	(npn6.C)++(0,1) node[draw=red](P4){}
	(npn6.B) to[short,l=LO-,-o] ++(-0.5,0)
	(P4) to[R,l_=$R_2$,-] ++(0,2) node[vcc]{}
	;
	\draw
	(6,0) node[npn] (npn6) {}
	(npn6.C)++(0,1) node[draw=red](P4){}
	(npn6.B) to[short,l=LO-,-o] ++(-0.5,0)
	(P4) to[] ++(0,2) node[vcc]{}
	;
\end{circuitikz}
\end{document}

which produces:

Now, the code is slightly incorrect --- (P4) is a node and not a coordinate, and as you can see, using the coordinate is ok. But effectively the behavior is strange, and it's due to circuitikz, because a normal to does not show it.

The patch fixes it but it disrupts the path, unfortunately. Not sure why because I do not understand the code ;-)

[Rmano]

Now, this is a different patch (applied over the revert):

diff --git a/tex/pgfcircpath.tex b/tex/pgfcircpath.tex
index a20f8b7..c6a855a 100644
--- a/tex/pgfcircpath.tex
+++ b/tex/pgfcircpath.tex
@@ -107,6 +107,7 @@
 
     (\tikztostart) coordinate (\ctikzvalof{bipole/name}start)%necessary to get correct coordinates in the case of relative start/end or constructions like ((node1)-|(node2))
     (\tikztotarget) coordinate (\ctikzvalof{bipole/name}end)
+    (\tikztostart) % go back to start
     \pgfextra{
         \pgfmathanglebetweenpoints{\pgfpointanchor{\ctikzvalof{bipole/name}start}{center}}
         {\pgfpointanchor{\ctikzvalof{bipole/name}end}{center}}

Seems to fix both things:

and

So... any comment? @sistlind @mredaelli @der-stefan

[Rmano]

Notice that now the original example in issue #76 compiles like this:

Which is correct --- P3 and P4 are nodes and so they have a non-zero sizes. You can see what's happening if you add [draw=red] to them:

The behavior of forcing the start point at the center of the node is not standard and creates a lot of problems.