kicad-developers team mailing list archive

[Lorenzo Marcantonio <l.marcantonio@xxxxxxxxxxxx>]

On Fri, Jun 22, 2012 at 05:39:02PM -0500, Dick Hollenbeck wrote:
> Was that a request for help?

That was a bug report. Last time I encountered it Jean-Pierre asked for the board so I attached it.

> You call me a broken record and a tyrant within in a span of two days and then don't know
> how to ask for help.  Put yourself in a box did you?

"Tyrant" was for fun. Broken record was for real...

> Please help test the attached patch on Linux.
> The "double" was not good enough, "long double" is working better so far.

Uhm... the suspect was something like that in fact... shouldn't be better a solution not involving floating point? We could simply format as an integer and then insert the decimal point in the right place, that shouldn't be too complex.

> If it works, I will work on the Mingw solution tomorrow, but I gotta run now.

Why are there issues on mingw for that? Isn't long double 80 bit on both or its printf doesn't support long doubles?

> My testing strategy was to simply compile for nanometers, load you nanometer board, and
> save it under another name, then use diff on the two files.

Seems reasonable.

-- 
Lorenzo Marcantonio
Logos Srl