KiCad 5.1.2.
Track clearance during routing hides from me the working area. It really disturb me as I don’t know where I precisily am. The track alone is small enough that I know where I am but with this clearance - not.
When I hide it by setting in Display Options Track Clearance to ‘Do not show’ then also disappears the clearance info around the H10 via which I would like to have when I will be positioning it a moment later.
The best would be to have this clearance information but also see what is behind.
Is it possible to set it to outline mode or to make it transparent? I didn’t found where I can set its colour (transparency should also be there). When I set tracks to outline mode it works only for existing ones and not for just routing and its clearance.
I know that more KiCad like way would be to be not precision and eventually use later Push, Shove. But it would be possible if I can define all clearances as I would like, but I see I can’t. To allow me routing I had to set clearance smaller than I would like so to put tracks with clearance as I wont I have to do it manually.
I noticed that stable versions are generated at weekends (I don’t know if it is rule, but it is my observation). Since two months I check each Monday if it is that great day of 5.1.3
I spend one hour reading at wiki what github is and main what I understood is that I should first know what Git is. The next position “to understand” but where from to get a time for all waiting "to understand"s
If ‘master’ from github means that you get KiCad 5.1 nightlies installed automagically then it can’t reach my working computer as it is not connected to internet (I fill myself not enough good to be sure I will never suddenly see “Pay $$$$ if you wont to get your work back”).
I am not using nightlies as in download. I pull the most recent state of master and compile locally.
All you need to do is a --> git pull git+ssh://jos+@git.launchpad.net/kicad <-- inside a terminal/console window. Then follow the processing instructions.
You may have to resolve some dependencies. Otherwise it’s a straight forward procedure.
Alternatively you can use things like git-cola. Might be in your list of software sources. Presuming you run some *nix version.
You are speaking different language
I know that there are somewhere the console window but one minute spend on searching was not enough for me to find how to get to it.
DOS times when I felt I understand what my computer was doing are gone.
Since 1989 if I have to program anything (may be a week per year I am a ‘programer’ ) I use Borland environment (starting from Turbe C++ 1.0) which effectively isolates me from all such things like console.
Master is the name of a git branch. (Simplified: git branches are how different states of the repo can live at the same place.) For KiCad the master branch is the one used for developing the next major release (Right now version 6)
Meaning if you “download” the master branch then you do not get the nightlies for v5.1 but for v6. Github is just a mirror of the source code repo not the place where compiled packages are found. (You would need to compile them yourself after “downloading” the master branch.) To find out where to get the compiled nightlies for your operating system check the download page of the kicad website.
Some time ago I found how to get to 5.1 nightlies starting from kicad website. The kay is to notice that when you select to go to nightlies you get not into list but into directory structure and you can travel to 5.1. First time being there I just didn’t noticed that.
Reinstalling (in my case at two computers - just to have at internet connected one a version info when I report a bug) takes time and as I expect 5.1.3 soon I decided to wait. 5.1.2 bugs disturb me a little in PCB so I am considering today - may be instead of finishing my first PCb I will spend time on makeing schematics for next 3 PCBs (some new symbols and footprints will be needed) - when I will finish may be 5.1.3 will be ready.
For one week working per year may be not economy reasonable.
The main program I am conserving is our production program - to load a software into each of our products. I have it under Borland C++ Builder 5 and it is the tool I fill myself the best in.
I didn’t upgraded to Builder 6 because they changed Standard version into Personal (not allowed for commercial use).
I also have Builder 2010 but don’t like it mainly because of its feature “Next run only after computer reset”. It is known bug under Windows 7, and there are some third source work-arounds, but didn’t tried to install it. I just remember if I run it I don’t close it. I use it only for few programs which have to communicate via ethernet and builder 5 don’t have the libraries which could be needed for it.
I have never learned myself how to use third source libraries. If I need something I prefer to invent the wheel once more. For example when I needed crypt functions I have written them based on NIST documents and I got very short (but probably not optimal) source (AES128 + AES256 + CTR + CMAC - 30 lines in .h and 160 lines in .cpp).
Especially bad idea for crypto as it is very unlikely that you do not introduce possible attack vectors. (There are full books written about how bad of an idea it is to implement crypto functions for yourself without properly testing it against all known attacks. Main problem are sidechanbel attacks like timing and current consumption based ones that are non trivial to prevent.)
I have not read them, but suppose they are mainly about inventing your own super-hiper top-secret crypto algorithms (see mifare-classic) but not about just implementing well tested algorithms.
Timming - in AES there are no if()s so whatever are the bits values you do the same set of calculations. I think timming is rather important in protocol and not encryption algorithm.
Current consumption - I don’t believe in such attack on my program working at PC. The other subject is microcontroller. But there in opposite I don’t believe you can do a lot to protect yourself against current consumption attack (I am speaking about implementing cryptology by software and not hardware).
No they are literally about how to implement AES such that it is even secure against the most basic attacks. (Just implementing the mathematical function does not suffice as there are a lot of things to consider.)
I literally did such an attack in one of the first courses at university. So yes it is more than just possible! this is how one attacks any embedded device. For programs on a real pc one attacks the caching stuff. (Even harder to protect against and is shown to be attackable from java script running in your browser! And this is another form of timing attack so timing is not just violated by use of if statements but also by things like cache hit/miss, read and write to different places in the memory, writing ones can also take longer or shorter than zeros depending on memory technology, … In other words there are a ■■■■■■■ of places where timing can be used to find out which value has been written.)
And we have not even started to think about the normal attack vectors like buffer overflows, format string attacks, integer overflows, … (You are kind of save if you use a very modern high language, c and c++ are quite bad in this regard.)
