Craig Ringer wrote: > gww at silcom.com wrote: > >> At the Libre Graphics Meeting in Montreal I was asked by someone from Scribus >> (and I can't recall who any longer) to write a tool for checking that a font >> was >> valid (didn't have any intersecting contours, had points at the extrema, >> etc.) >> >> I have now written this tool, called fontlint. >> It is part of the current release of fontforge. >> >> http://fontforge.sf.net/ >> > > Out of >2000 fonts I've run this on, including the full Adobe OTF > collection, the MS TT core fonts, some scary old clipart CD fonts, and > the Vera series, I've found exactly 16 faces that pass (all from the > Adobe OTF collection). > > I realise that, like any `lint'-like tool, it'll identify all sorts of > issues that are routinely worked around in practice or are merely > warnings, but I'm still surprised by the the number of flat out fails. > > Is it possible that these fonts are all really that bad? Are there any > quality collections it'll pass? > > Maybe the man page (and I'm impressed that there is one) needs to note > that fails are extremely common in widely used sets of commercial fonts. > > Do you know on a practical level what things like 'Self Intersecting Glyph' and 'Wrong Direction' mean? Would these be visible? Do they affect exporting to PDF? Are there fonts without problems?
Greg
