ID: | 1158 |
From: | ge...@igor2.repo.hu |
Date: | Sat, 30 Dec 2017 15:53:47 +0100 (CET) |
Subject: | Re: [pcb-rnd] end-of-year stats |
in-reply-to: | 1157 from ka...@aspodata.se |
On Sat, 30 Dec 2017, karl@aspodata.se wrote: > Igor2: >> On Sat, 30 Dec 2017, karl@aspodata.se wrote: > ... >>> Regarding documentation, I like the way http://lilypond.org/ does its >>> documentation. They have texinfo sorce files and produces html, pdf, and >>> info files from that. >> >> Sorry, I don't like that. We use html as source and can generate pdf. We > ... > > Ok, we want different things then. > >>> And, on another note, they use compare(1) for regression tests. >> >> For most tests we use diff(1). > > Yes, what about rendering, compare is a "graphical diff" the imagemagic > suite; just a suggestion nothing more. Sure, thanks; for most formats (including graphical formats like svg, ps, eps) we require byte-by-byte match. If we introduce (or fix) a rounding error, depending on the DPI of rendering for a graphical comparison the difference may not be noticed. The only exception is png, because it use an LZW compression that depends on the version and configuration of the local libpng, so byte-by-byte match is impossible ot achieve. For pngs, we already use compare(1). If you are interested in the details, you can check the methods in trunk/tests/RTT/Export.sh. Best regards, Igor2
Reply subtree:
1158 Re: [pcb-rnd] end-of-year stats from ge...@igor2.repo.hu