X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
Date: Mon, 9 Jul 2012 18:46:22 -0700
From: Andrew Poelstra <asp11 AT sfu DOT ca>
To: geda-user AT delorie DOT com
Subject: Re: [geda-user] [PATCH] Allow to create metric Gerber and drill
 files, hopefully final version.
Message-ID: <20120710014622.GE23449@malakian.lan>
References: <20120703140236 DOT GA12646 AT visitor2 DOT iram DOT es>
 <CAKakQcdTgG6E2h+UgWoh-ujP6vFDH=voY-PBJR3OFG2awqL0_A AT mail DOT gmail DOT com>
 <20120705101614 DOT GA19974 AT visitor2 DOT iram DOT es>
 <20120706202633 DOT GA28355 AT malakian DOT lan>
 <20120706213340 DOT GA30622 AT visitor2 DOT iram DOT es>
 <20120707104156 DOT 17641 DOT qmail AT stuge DOT se>
 <20120708042719 DOT GE1637 AT malakian DOT lan>
 <20120709224559 DOT GB4355 AT visitor2 DOT iram DOT es>
 <20120709234535 DOT GD23449 AT malakian DOT lan>
 <20120710004318 DOT GA28800 AT visitor2 DOT iram DOT es>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
In-Reply-To: <20120710004318.GA28800@visitor2.iram.es>
User-Agent: Mutt/1.5.20 (2009-12-10)
X-MIME-Autoconverted: from 8bit to quoted-printable by pobox.sfu.ca id q6A1qZ6s005852
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by delorie.com id q6A1qfEE018781
Reply-To: geda-user AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: geda-user AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

On Tue, Jul 10, 2012 at 02:43:18AM +0200, Gabriel Paubert wrote:
> On Mon, Jul 09, 2012 at 04:45:35PM -0700, Andrew Poelstra wrote:
> > On Tue, Jul 10, 2012 at 12:45:59AM +0200, Gabriel Paubert wrote:
> > > 
> > > Thanks for pointing this out, I completely missed it, but I think 
> > > that they point to the wrong 0 when looking at the source. 
> > >
> > 
> > Oh, good catch!
> >  
> > > Corrected in the following, with a fix for octogonal apertures
> > > in which I did a mistake and ended up with two metric conversion
> > > specifications. Note that I have touched a bit the whitespace at
> > > the beginning of the comments so that they start with a tab like
> > > the surrounding code lines.
> > >
> > 
> > Can you elaborate on this a bit? Specifically, do we need a new
> > test case to check on this (since our existing ones apparently
> > did not catch the bug)?
> 
> Maybe, but I don't know how popular octogonal pads are. It seems 
> to be a very specific feature of PCB, which may be useful for home 
> etching, but subcontracting small PCB runs has become so cheap!
> 
> The only time I tried to use octogonal pads, the photoplotter of my
> PCB manufacturer produced garbage, so I never use them.
> 
> [Testing PCB]
> 
> Ok, it seems the bug is not actually a bug since octogonal pads
> are procuded as polygons in Gerber output. I suspect that the
> OCTOGONAL case in aperture definitions is dead code.
>

This is my suspicion too. I'll investigate.
 
> > 
> > > 
> > > The fact that imperial Gerber have 2 orders or magnitude better
> > > resolution than the associated drill is a problem (25.4µm versus
> > > 0.254µm). At least metric gives micrometer resolution for both. 
> > >
> > 
> > Is this a problem with the spec, or our implementation?
> 
> Our implementation, for losing the last digit. This causes visible
> centering problem in the smallest vias, and is a regression from
> earlier revisions.
>

Sigh.. the problem is that keeping the last digit causes overflows.
We'll have to figure out a smarter solution.

-- 
Andrew Poelstra
Email: asp11 at sfu.ca OR apoelstra at wpsoftware.net
Web:   http://www.wpsoftware.net/andrew
"You shouldn't trust every quote you read on the Internet." -- Socrates