add --alias option; catch oddity of Libertine tt

[phi-gamma]

this addresses the feature request: #4

also, verbosity been adapted so that a log level of 2 emulates the behavior of the original mkluatexfontdb script, whereas fontdbutil defaults to 1 (less clutter)

[eroux]

Maybe this could be bugreported to the Libertine authors?

[phi-gamma]

Do you have a sourceforge account? Their tracker doesn’t appear to accept outside submissions: http://sourceforge.net/tracker/?atid=590374&group_id=89513 .
Also, what case do we have? Is there a spec we could cite that prescribes valid subfamily identifiers?

[eroux]

Looking at the bugs, I found http://sourceforge.net/tracker/?func=detail&aid=3556401&group_id=89513&atid=590374 which might be similar... I'll bugreport though.

[phi-gamma]

Looking at the bugs, I found http://sourceforge.net/tracker/?func=detail&aid=3556401&group_id=89513&atid=590374 which might be similar... I'll bugreport though.

This issue appears to be fixed by changing glyph dimensions. It’s the same font, but we don’t test for something like that.

[eroux]

Ok, I'll go ahead and report then

[eroux]

Reported at https://sourceforge.net/tracker/?func=detail&aid=3611833&group_id=89513&atid=590374. There are many weird subfamilies in my otfl-names.lua, example:

{
["familyname"]="EB Garamond 12 All SC",
["filename"]={ "/usr/share/fonts/truetype/pouet/EBGaramond12-AllSC.ttf", false },
["fontname"]="EBGaramond12-AllSC",
["fontstyle_name"]="AllSC",
["fullname"]="EB Garamond 12 Regular All SmallCaps",
["names"]={
["family"]="EB Garamond 12 All SC",
["fullname"]="EB Garamond 12 Regular All SmallCaps",
["psname"]="EBGaramond12-AllSC",
["subfamily"]="AllSC",
},
["size"]={ 120, 1200, 95 },
["slant"]=0,
["weight"]=400,
["width"]=5,
},

...

[phi-gamma]

The issue will occur iff the subfamily string is not covered by the style synonym table, and if the request string equals the family string. With EB Garamond, if we get a

name:EB Garamond 12 All SC

lookup, and with Libertine, if we get

name: Linux Libertine O Mono

-- both could happen more often. The new pull request eliminates an else-branch and fonts with unknown subfamilies should now be matched against their other names as well.