@PauloCereda though I quite wonder what kind of psychopath those two are, the rabbit wearing the head of a killed goose and the goose standing right next to him as if everything was normal.
@Skillmon it is not a longtable problem, I'm just updating hyperref (but all sort of tests fails because Frank changes latex and Joseph changes line lengths ;-( <-- @DavidCarlisle
If I wanted to write an extension to expl3 (so a programming module that isn't contained in the l3kernel itself) what would be the correct approach? Using LaTeX2e packages? Just providing some file that should be used with \input?
@DavidCarlisle I made a new hyperref and will upload it. Apart from the important fix for longtable, there is the fix for the non-ascii labels and I also removed the break point after whatsits (let's see if I removed to much ...)
@Rmano I should've paid attention on the "RS-FF" label for the dipchip, that could've given it away that this is indeed a flipflop. However, I'll try implementing a pgfutil-only variant of \xdipchip@if@pin.
@Rmano I think mine could end up in the package (if one finds a sane version of the \clist_if_in: code implemented with pgfutil (easy), and a sane implementation of \clist_set:Nn using pgfutil, that also does some error detection (a bit harder))
(pgfutil misses a good reference manual that shows concisely and well ordered, which functions are available, browsing the sources and guessing what is possible is sometimes tiring)
@Skillmon yes --- it would be nice. My solution is much more gross: just do not draw the pins but leave the anchors. The negative thing is that you have to draw the pin you really need, bit it is coherent with the muxdemux thingies that already have it.
what I do with flipflops is to draw the pin only if there is a corresponding label in the symbol --- which normally works well.
Thinking a bit, maybe just the solution I posted is ok for the generic chip. You just have to connect to the bpin and use the pin anchor for the links... hmmm
@StefanKottwitz github forks and CNAME could be easier, it is easy to end up with forks claiming to use the same domain name or (as here) merging back from a fork can easily delete the CNAME file