RE: [Devel] Re: Distiller produced PFA problems

[Blinn, Eric W]

Tom Kacvinsky wrote:
>
> It might help me a *lot* if I saw one of the PDF files you 
> guys are trying to rasterize.

Ask and ye shall receive.  I'll send you the PDF that I've been using as a
testcase while troubleshooting this stuff.  It is a contrived example that
is mostly just a series of font names, in their font, one per line. It is
only about 182k, but I'll just send it directly unless somebody requests it
sent to the mailing list.

Thanks,
Eric


> -----Original Message-----
> From: Tom Kacvinsky [mailto:address@hidden
> Sent: Tuesday, January 23, 2001 8:50 PM
> To: address@hidden
> Subject: [Devel] Re: Distiller produced PFA problems
> 
> 
> Ugh...  That wasn't it.  I had changed the FontMatrix to be a 
> 1000 units per em
> scale instead of 2048 units per em scale (the reason why I 
> chose the latter
> number is because the font was clearly autoconverted from TT 
> to Type 1 by some
> Windows print driver, and the resulting Type 1 font was 
> distilled into a PDF
> file).
> 
> Any way, once I changed the font matrix back, things looked fine...
> 
> It might help me a *lot* if I saw one of the PDF files you 
> guys are trying to
> rasterize.  There is more font info in a PDF file than just 
> the FontFile,
> FontFile2, or FontFile3 streams...
> 
> Tom
> 
> On Tue, 23 Jan 2001, Tom Kacvinsky wrote:
> 
> > I suspect this might have to do with the fact that the 
> FontBBox in the font you
> > sent me is {0 0 0 0}.  Does this over magnification happen 
> if the FontBBox is
> > non-zero (by zero bbox, I mean all entries are zero.  a 
> non-zero bbox then means
> > that not all entries are zero.  this is the mathematician 
> in me speaking. :)
> >
> > Tom
> >
> > On Tue, 23 Jan 2001, Blinn, Eric W wrote:
> >
> > >
> > > Hi Tom,
> > >
> > > Thanks.  With your patch we are almost there with 
> Distiller produced PFAs,
> > > but there is still one problem.  We now have glyphs and 
> metrics to work with
> > > and they look good, but the glyphs themselves need to be 
> scaled down by 50%.
> > >
> > > I have a local fix... okay hack really, where I'm keying 
> off some of the
> > > missing FontInfo data to detect that we are dealing with 
> a Distiller
> > > produced PFA and then scaling the font matrix xx and yy 
> values back by half.
> > > Not a fix I'll want to stick with mind you, just "proof 
> of concept" in the
> > > course of troubleshooting.  I'll leave it to you and the 
> other FreeType
> > > gurus to determine how and where to handle this in the library.
> > >
> > > Thanks,
> > > Eric
> > >
> > >
> > > > -----Original Message-----
> > > > From: Tom Kacvinsky [mailto:address@hidden
> > > > Sent: Monday, January 22, 2001 11:34 PM
> > > > To: address@hidden
> > > > Subject: [Devel] Re: Distiller produced PFA problems
> > > >
> > > >
> > > > Well, here is the patch.  I have yet to add dummy 
> FontInfo support (I
> > > > am still thinking about how best to approach it).  I 
> send this patch
> > > > along for review before I commit it.  The thing to note 
> is that the
> > > > test I removed prevented one from parsing the 
> dictionaries if the
> > > > FontInfo keyword was never seen.  Now that test is 
> gone, and things
> > > > work fine.  I tested with both MM and regular Type 1 
> fonts, and had no
> > > > problems.
> > > >
> > > > Oh yeah, I rechecked my Type 1 notes: it is OtherSubrs 
> that can be
> > > > missing from the font, and ATM and its ilk is 
> *supposed* to provide
> > > > defaults.  If any given char string uses a subroutine, 
> all subroutines
> > > > must be present (whereas I had previously thought that 
> subrs 0 - 3
> > > > could be missing).  What I will do is add code that 
> will return an
> > > > error if the font uses subroutines and the Subrs array 
> is missing.
> > > > This, I think, is supposed to go into t1decode.c (part 
> of the psaux
> > > > module).
> > > >
> > > > Tom
> > > >
> > > > On Mon, 22 Jan 2001, Tom Kacvinsky wrote:
> > > >
> > > > > Hi,
> > > > >
> > > > > I will get the patch out tonight.  I know what to change, I
> > > > just need to make
> > > > > sure that there aren't any ill side affects.
> > > > >
> > > > > Tom
> > > > >
> > > > > On Mon, 22 Jan 2001, Blinn, Eric W wrote:
> > > > >
> > > > > >
> > > > > > Tom Kacvinsky wrote:
> > > > > > >
> > > > > > > I found the problem, finally.  There is bug in 
> FreeType that
> > > > > > > reports there is no CharStrings dictionary when 
> there is one.
> > > > > > > Adding a FontInfo dictionary causes the false report to go
> > > > > > > away.  I'll see what I can come up with.
> > > > > >
> > > > > > Thanks for all the effort Tom.  Since you've found 
> the problem
> > > > > > this information probably has little value, but you 
> had asked
> > > > > > in another message about whether I had added a 
> dummy FontInfo
> > > > > > dictionary when testing your fix on my platforms.  
> I had not,
> > > > > > but the analysis you gave pretty well explains how the dummy
> > > > > > FontInfo data was masking the CharStrings dictionary bug at
> > > > > > your end.
> > > > > >
> > > > > >
> > > > > > > Eric, I hope this isn't causing you too much stress, what
> > > > > > > with the 11th hour show stopper you were referring to...
> > > > > >
> > > > > > Well, maybe only a couple days of stress!  :-)
> > > > > > It looked like it could hold up our release, but since then
> > > > > > we have done a risk analysis.  With the data we're handling,
> > > > > > we don't think it will be a problem in the immediate future
> > > > > > but it could be (likely will be) down the road.  Now that
> > > > > > you have a handle on the problem, I'll just roll whatever
> > > > > > fix you settle on into our next release (or perhaps 
> even this
> > > > > > one if circumstances permit).
> > > > > >
> > > > >
> > > > >
> > > > > _______________________________________________
> > > > > Devel mailing list
> > > > > address@hidden
> > > > > http://www.freetype.org/mailman/listinfo/devel
> > > > >
> > > >
> > >
> > > _______________________________________________
> > > Devel mailing list
> > > address@hidden
> > > http://www.freetype.org/mailman/listinfo/devel
> > >
> >
> >
> > _______________________________________________
> > Devel mailing list
> > address@hidden
> > http://www.freetype.org/mailman/listinfo/devel
> >
> 
> 
> _______________________________________________
> Devel mailing list
> address@hidden
> http://www.freetype.org/mailman/listinfo/devel
>