[texhax] visualize font expansion

[Arno Trautmann]

Paul Isambert wrote:
> 
> 
> Le 22/06/2011 10:21, Arno Trautmann a écrit :
>> Paul Isambert wrote:
>>>    function (head)
>>>      local f = font.getfont(font.current()).characters
>>>      for l in node.traverse_id(0, head) do
>>>        local g = l.head
>>>        while not(g.id == 37) do
>>>          g = g.next
>>>        end
>>>        texio.write_nl("Expansion: " .. g.width / f[g.char].width)
>>>      end
>>>      return head
>>>    end
>> I did:
>>
>>    for line in node.traverse_id(HLIST,head) do
>>      g = line.head
>>      while not(g.next == nil) do
>>        g = g.next
>>        if g.id == 37 then
>>          act_width = g.width
>>          normal_width =
>> font.getfont(font.current()).characters[g.char].width
>>          exp_factor = act_width / normal_width
>>        end
>>      end
>>    end
>>
>> So, as far as I can see, the difference is that I load the font at each
>> comparison which seems to slow down things very much, especially for
>> some 30-line-paragraph. With your code, I don't see any speed loss, too.
> 
> Oh, no, what makes your code so slow is that you inspect all glyphs in a
> line. So obviously that takes some time. And actually it takes time to
> print the values to the terminal, not to process the code; if I call
> LuaTeX with your code without the command line, then it isn't slow.

Ok …

> I suppose your inspecting every glyph is unintentional: your "while not
> (g.next == nil)" loop is true as long as your're not on the last node of
> the line. I suppose you meant to retrieve only the first glyph of a
> line, then you should have done:
> 
> local g = line.head
> while not (g.id == 37) then
>   g = g.next
> end
> ... then retrieve the width...
> 
> Note that this code will crash in the unlikely case where there is no
> glyph in a line. I let you fix it.

Yes, that's the reason for the while not(g.next == nil) do. However,
that was sloppy and yes, it should have stopped after the first id ==
37. However, I'll skip this test, as it indeed is very unlikely.

cheers
Arno

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 262 bytes
Desc: OpenPGP digital signature
URL: <http://tug.org/pipermail/texhax/attachments/20110622/3336064f/attachment.bin>