Book PostScript output different than section PostScript output

[MrVideoSenior]

Here is the first issue that I mentioned in the EPS output issue.

I discovered the issue when I created a PDF of the book I've assembled.  The problem occurs with both FM 7.2 and FM 9.  I was hoping that FM 9 fixed the issue, but it was not to be.

The issue:  text is just plain missing from the PDF.  It is purely random.  There is no common factor in the body of the sections to cause the issue.

The PostScript file that is created has a drastically different character output for some areas, and normal text output in others.  But, the kicker is that if I create PostScript from one of the sections, in this case, the two page Table of Contents, the output is correct.  Same settings for both.

I've cut the good code and bad code from the files:

Good Code:

F /F2 0 /0 T /NewCenturySchlbk-Roman mF

/F2SA6 F2 [166 0 0 -166 0 0 ] mFS

F2SA6 Ji

2700 2044 M (I)S

2768 2044 M (n)S 2869 2044 M (t)S 2934 2044 M (r)S 3008 2044 M (o)S 3091 2044 M (d)S 3187 2044 M (u)S 3289 2044 M (c)S 3363 2044 M (t)S 3428 2044 M (i)S 3480 2044 M (o)S 3563 2044 M (n)S 3665 2044 M ( )S

3751 2044 M (.)S

3797 2044 M (.)S 3843 2044 M (.)S 3890 2044 M (.)S 3936 2044 M (.)S 3982 2044 M (.)S 4029 2044 M (.)S 4075 2044 M (.)S 4121 2044 M (.)S 4168 2044 M (.)S 4214 2044 M (.)S 4260 2044 M (.)S 4307 2044 M (.)S 4353 2044 M (.)S 4399 2044 M (.)S 4445 2044 M (.)S 4492 2044 M (.)S

4538 2044 M (.)S 4584 2044 M (.)S 4631 2044 M (.)S 4677 2044 M (.)S 4723 2044 M (.)S 4770 2044 M (.)S 4816 2044 M (.)S 4862 2044 M (.)S 4909 2044 M (.)S 4955 2044 M (.)S 5001 2044 M (.)S 5048 2044 M (.)S 5094 2044 M (.)S 5140 2044 M (.)S 5186 2044 M (.)S 5233 2044 M (.)S

5279 2044 M (.)S 5325 2044 M (.)S 5372 2044 M (.)S 5418 2044 M (.)S 5464 2044 M (.)S 5511 2044 M (.)S 5557 2044 M (.)S 5603 2044 M (.)S 5650 2044 M (.)S 5696 2044 M (.)S 5742 2044 M (.)S 5788 2044 M (.)S 5835 2044 M (.)S 5881 2044 M (.)S 5927 2044 M (.)S 5974 2044 M (.)S

6020 2044 M (.)S 6066 2044 M (.)S 6113 2044 M (.)S 6159 2044 M (.)S 6205 2044 M (.)S 6252 2044 M (.)S 6298 2044 M (.)S 6344 2044 M (.)S 6390 2044 M (.)S 6437 2044 M (.)S 6483 2044 M (.)S 6529 2044 M (.)S 6576 2044 M (.)S 6622 2044 M (.)S 6668 2044 M (.)S 6715 2044 M (.)S

6761 2044 M (.)S 6807 2044 M (.)S 6854 2044 M (.)S 6900 2044 M (.)S 6946 2044 M (.)S 6993 2044 M (.)S 7039 2044 M (.)S 7085 2044 M (.)S 7131 2044 M (.)S 7178 2044 M (.)S 7224 2044 M (.)S 7270 2 044 M (.)S 7317 2044 M (.)S 7363 2044 M (.)S 7409 2044 M (.)S

8008 2044 M (1)S

30000 VM?

Bad Code:

Pscript_WinNT_Compat dup /suspend get exec

Pscript_WinNT_Incr dup /resume get exec

/DeviceGray dup setcolorspace /colspABC exch def

0 5 scol 27000 VM?

Pscript_WinNT_Incr begin

%%BeginResource: file Pscript_T3Hdr 5.0 0

{version cvi 2016 ge{32/FontType resourcestatus{pop pop true}{false}ifelse}

{false}ifelse}exec/Is2016andT32? exch def/T32DefSBCMap{/CIDInit/ProcSet

findresource begin 10 dict begin begincmap/CIDSystemInfo 3 dict dup begin

/Registry(Adobe)def/Ordering(Identity1)def/Supplement 0 def end def/CMapType 0

def/WMode 0 def 1 begincodespacerange<00><ff>endcodespacerange 1 begincidrange

<00><ff>0 endcidrange endcmap/DrvSBCMap currentdict/CMap defineresource pop end

end}bind def Is2016andT32?{T32DefSBCMap}def/T32RsrcBegin{Is2016andT32?{

/BitmapFontInit/ProcSet findresource begin}if}bind def/T32RsrcEnd{Is2016andT32?

{end}if}bind def/AddT32Char{6 1 roll 0 get 7 1 roll pop pop 5 1 roll pop

findfont/TT32R get addglyph}bind def/AddT3Char{findfont dup 5 2 roll 1 index

length 0 gt{cvx 1 index exch 4 exch put dup(imagemask)cvx cvn 5 exch put cvx}

{pop cvx}ifelse 3 -1 roll/CharProcs get 3 1 roll put dup/Encoding get 5 -1 roll

4 index put/Metrics get 3 1 roll put}bind def/AddT3T32Char Is2016andT32?{

/AddT32Char}{/AddT3Char}ifelse load def/GreNewFontT32{5 dict begin exch

/FontMatrix exch def exch/FontBBox exch def exch pop exch pop/CIDFontType 4 def

dup currentdict end/CIDFont defineresource 3 -1 roll dup/DrvSBCMap dup/CMap

resourcestatus{pop pop}{T32DefSBCMap}ifelse 5 -1 roll[exch]composefont dup

length dict copy dup/FID undef begin exch/TT32R exch def currentdict end

definefont/BitmapFontInit/ProcSet findresource begin/TT32R get[14 0 0 0 0 0]<>0

4 -1 roll addglyph end}bind def/GreNewFontT3{11 dict begin pop/FontType 3 def

/FontMatrix exch def/FontBBox exch def/Encoding exch def/CharProcs 257 dict def

CharProcs/.notdef{}put/Metrics 257 dict def Metrics/.notdef 3 -1 roll put

AddFontInfoBegin AddFontInfo AddFontInfoEnd/BuildChar{userdict begin/char exch

def dup/charname exch/Encoding get char get def dup/Metrics get charname 2 copy

known{get aload pop}{pop/.notdef get aload pop}ifelse setcachedevice begin

Encoding char get CharProcs exch 2 copy known{get}{pop/.notdef get}ifelse end

exec end}def currentdict end definefont pop}bind def/GreNewFont{Is2016andT32?

{GreNewFontT32}{GreNewFontT3}ifelse}bind def/UDF3{Is2016andT32?{/BitmapFontInit

/ProcSet findresource begin dup/CIDFont findresource removeall/CIDFont

undefineresource undefinefont end}{pop UDF}ifelse}bind def

%%EndResource

end reinitialize

/TTF532b00

[166 0 0 0 0 0 ]

256 array 0 1 255 {1 index exch /.notdef put} for

[0 0 0 0 ]

[1 166 div 0 0 -1 166 div 0 0 ]

/__TTF532b00

GreNewFont

F /F7 0 /0 F /TTF532b00 mF

/F7SA5 F7 [165.625 0 0 -165.625 0 0 ] mFS

F7SA5 Ji

2700 2044 M <01>S

2768 2044 M <02>S 2869 2044 M <03>S 2934 2044 M <04>S 3008 2044 M <05>S 3091 2044 M <06>S 3187 2044 M <07>S 3289 2044 M <08>S 3363 2044 M <03>S 3428 2044 M <09>S 3480 2044 M <05>S 3563 2044 M <02>S 3665 2044 M <0A>S

27000 VM?

/TTF532b00

[166 0 0 0 0 0 ]

256 array 0 1 255 {1 index exch /.notdef put} for

[0 0 0 0 ]

[1 166 div 0 0 -1 166 div 0 0 ]

/__TTF532b00

GreNewFont

3751 2044 M <0B>S

3797 2044 M <0B>S 3843 2044 M <0B>S 3890 2044 M <0B>S 3936 2044 M <0B>S 3982 2044 M <0B>S 4029 2044 M <0B>S 4075 2044 M <0B>S 4121 2044 M <0B>S 4168 2044 M <0B>S 4214 2044 M <0B>S 4260 2044 M <0B>S 4307 2044 M <0B>S 4353 2044 M <0B>S 4399 2044 M <0B>S 4445 2044 M <0B>S 4492 2044 M <0B>S

4538 2044 M <0B>S 4584 2044 M <0B>S 4631 2044 M <0B>S 4677 2044 M <0B>S 4723 2044 M <0B>S 4770 2044 M <0B>S 4816 2044 M <0B>S 4862 2044 M <0B>S 4909 2044 M <0B>S 4955 2044 M <0B>S 5001 2044 M <0B>S 5048 2044 M <0B>S 5094 2044 M <0B>S 5140 2044 M <0B>S 5186 2044 M <0B>S 5233 2044 M <0B>S

5279 2044 M <0B>S 5325 2044 M <0B>S 5372 2044 M <0B>S 5418 2044 M <0B>S 5464 2044 M <0B>S 5511 2044 M <0B>S 5557 2044 M <0B>S 5603 2044 M <0B>S 5650 2044 M <0B>S 5696 2044 M <0B>S 5742 2044 M <0B>S 5788 2044 M <0B>S 5835 2044 M <0B>S 5881 2044 M <0B>S 5927 2044 M <0B>S 5974 2044 M <0B>S

6020 2044 M <0B>S 6066 2044 M <0B>S 6113 2044 M <0B>S 6159 2044 M <0B>S 6205 2044 M <0B>S 6252 2044 M <0B>S 6298 2044 M <0B>S 6344 2044 M <0B>S 6390 2044 M <0B>S 6437 2044 M <0B>S 6483 2044 M <0B>S 6529 2044 M <0B>S 6576 2044 M <0B>S 6622 2044 M <0B>S 6668 2044 M <0B>S 6715 2044 M <0B>S

6761 2044 M <0B>S 6807 2044 M <0B>S 6854 2044 M <0B>S 6900 2044 M <0B>S 6946 2044 M <0B>S 6993 2044 M <0B>S 7039 2044 M <0B>S 7085 2044 M <0B>S 7131 2044 M <0B>S 7178 2044 M <0B>S 7224 2044 M <0B>S 7270 2044 M <0B>S 7317 2044 M <0B>S 7363 2044 M <0B>S 7409 2044 M <0B>S

27000 VM?

/TTF532b00

[166 0 0 0 0 0 ]

256 array 0 1 255 {1 index exch /.notdef put} for

[0 0 0 0 ]

[1 166 div 0 0 -1 166 div 0 0 ]

/__TTF532b00

GreNewFont

8008 2044 M <0C>S

30000 VM?

Pscript_WinNT_Compat begin /$x mysetup def end

Pscript_WinNT_Incr dup /suspend get exec

Pscript_WinNT_Compat dup /initialize get exec

%%BeginDocument: Pscript_Win_PassThrough

1494 1878/G3.437354 FmPA

Notice that the text characters are now differently encoded inside of the <> characters.  The distillers (I tried three different versions, including Acrobat 9), all printed blanks where the characters were supposed to go.  None of the distillers crashed with an error, but all of them refused to actually render any characters.

The main problem... why did the FM/Adobe PDF printer even do that crap in the first place with the book output, when it did it correctly with the section of the book?  How does one get this crap to not happen in the first place?

[Arnis Gubins]

If you haven't applied the MS hotfix for the PDF drop-out problem, that's a good place to start. For details see:

http://blogs.adobe.com/techcomm/2009/07/repost_hotfix_for_framemaker.html

[MrVideoSenior]

@Arnis: Thanks for responding, but the issue the hotfix fixes is not the issue posted about.  The hotfix removes the problem with files that get corrupted.  As you notice from the code that I posted, the text is not corrupted.  The areas of text that vanish from the PDF file are replaced with all new code that is not interpreted correctly, or at all, by the distillers.

Plus, the problem occurs on W2K (as well as XP with FM 9).

FM, along with the Adobe PDF printer driver, is just plain producing bad PostScript.  Whatever the stuff is that it is doing, is just not renderable by the distillers.

[MrVideoSenior]

I did some further testing and it is not related to the document being a book.

I built up a 14 page file of the first 14 pages of the book, and the problem is withing a single Frame standalone document.

I then took a 32 page section of the book and had it print to PS.  Sure enough, the GreNewFont problem was alive an well in that output.

This is a nasty problem that needs to be fixed.  I can't create a viable PDF as it currently stands.

[Arnis Gubins]

The hotfix is for the drop-out problem in PDFs. The problem is on the MS side of the equation. FM is just using the MS supplied printer drivers to create the postscript.

If you don't want to install the hotfix, a temporary workaround that seems to address the problem is to delete the fntcache.dat file in your System folder and then reboot. The font cache gets rebuilt at start-up and things should be stable for a while, but the problem will creep back in.

[Art Campbell]

To put this even more simply, the MS Hotfix is for all PostScript printers in an XP environment. It addresses a large number of problems in addition to the ones noted.

If you elect not to install it, you are prtty much guaranteeing wonky PS at some level -- which may not always be fatal, but will certainly bite you on the butt when you least need more problems.

Art

[MrVideoSenior]

Excuse me for being blunt, but is anyone reading what I've posted?

The hotfix is for XP and later.  As mentioned, the FM 7.2 PotScript issue I'm having is on W2K, for which there is NO HOTFIX!!!

Also, it isn't corrupted PostScript code.  Please look again at the code samples that I posted.  Some of the text is encoded with something generated that is called GreNewFont.  It doesn't even use the ASCII character set, but some kind on binary representation that the distillers appear not to be able to handle.

Once I get past a small number of pages, the problem starts to take place.  I can take a single page section of the book and create PostScript from that and all is fine.  But when placed into a book as the 5th page, a section of that section is rendered in that GreNewFont garbage.

How do I get FM and the Adobe PDF printer driver to stop doing that and render the text in the desired font?

I'll try deleting the fntcache.dat file tonight after work.

[AxialInfo]

Waaaay (years) before the XP hotfix was released I did a lot of PS detective work on the missing fonts issue and the GreNewFont was always the "indicator" that any text after that point would be missing. The other symptoms you mentioned also occurred (small file sizes good, larger # of pages bad). The badness was also triggered sooner by having pages with multiple font changes (from regular to bold, or using fonts that had a separate file for the "Bold" such as the Univers 45/55 fonts).

This ties in with the workaround to reset the fntcache because there were also VM errors immediately before the problem text, so basically my interpretation is that the "font memory stack" (or whatever it's called) got filled up with each font change and thus triggered the error sooner.

I suspect that you've found a "related" problem in W2K, but I don't hold much hope that there might be a solution forthcoming from Adobe/MS -- if I recall correctly, it took about 2 years of submission from multiple folks without any indication that Adobe were even looking at the problem, much less had a solution, and then it turned out to be Windows related anyhow --  and given that W2K is pretty much long in the tooth I don't see much hope that a W2K fix would be prioritized.

As fellow users we can offer experience or commiseration, but we're not Adobe and the forum isn't a pipeline to Adobe, so if you want to pursue the issue you'll need to do it through the regular Adobe channels.

Sheila

[MrVideoSenior]

@AxialInfo:  Thanks. You are the first to indicate knowledge of the GreNewFont issue.  The book is indeed filled with font changes: size and style (regular, bold, italic).  There are only three font families used though.

More in the next response.

[Art Campbell]

Yes, we're reading....

You said "problem occurs on W2K (as well as XP with FM 9)." which implies that you're working at least part of the time on a XP system, does it not?

***

What version(s) of Acrobat are you using?

Have you verified that your FM and your Acrobat installs are fully patched?

How, exactly, are you generating the PDFs? SaveAs PDF, or printing to the Logical PDF printer, or generating a PS file and distilling, or???

Have you verified that there are no stray characters in the component files by saving as MIF, opening those files, and saving them as .FM?

Have you tried creating a new book file and copying in the(cleaned)  component files, one at a time?

[MrVideoSenior]

@Art:  As indicated in the above response, the GreNewFont thing I am seeing was verified by AxialInfo.

Here are the answers to your questions:

1) FM 9 on XP was just installed a couple of days ago when I discovered the problem with FM 7.2 to see if it was solved with a newer version of FM.  There has been no previous PostScript, or FrameMaker, usage on this particular computer, as it's main use is for something completely different.

2) Acrobat not installed on any home system.  It was tested on a 2003 server at work. That was before the info about the GreNewFont thing. The distillers and Adobe PDF printer drivers come from the FM versions that were installed on my systems.

3) FM 7.2 is old.  No idea if there are any patches for it.  But, since this appears not to be a FM problem, but a Windblows issue, I doubt that FM 9.0.3 will fix the issue (which I just downloaded yesterday).

4) What is a logical PDF printer? I first went directly to PDF, but when I discovered the missing text, I started outputting PostScript and distilling it. The PostScript output is done by selecting the Adobe PDF printer, selecting the add PDF stuff and printing to a file.  Then I looked at the PostScript output at the location where it screws up and found the GreNewFont thing.

5) I created the MIF files last night.  It was very late (actually early morning), so I haven't gotten to the point of recreating the .fm files and pulling them in to a new book, one file at a time.  Unfortunately the XP box has other priority uses tonight that keeps me from doing Frame stuff at the same time (it is beyond the scope of this discussion).  I do have an older, slower, XP box that I'll turn back on and load up with FM 9 and run tests with it.

There are other comments that AxialInfo made: I understand that you don't work for Adobe.  No idea if Adobe software engineering staff even read these forums, and if they do, are even allowed to respond.

Getting FM 7.2 fixed is obvioously not going to happen.  But, that said, workarounds or OS patches, etc. are welcome.  IIRC what you posted, it took a long time before anyone would believe there was a problem and by the time a fix was found, it was for XP, and later.

Yes, W2K is an older OS.  It has its uses, because I have software that will not work under XP that I still need to use.  Said software will never be updated.  So, I can't upgrade it to XP, as it will break functionality.  I even still have a 98SE computer that has software on it what doesn't work under W2K.  It mostly sits in the off state, but every now and then...

[Arnis Gubins]

Patches for FM 7.2 are available at:

http://www.adobe.com/support/downloads/product.jsp?product=22&platform=Windows

Try deleting the fntcache.dat file in your Windows\System32 folder and reboot the machine.

[MrVideoSenior]

Thanks, I'll grab the latest 7.2 and do plan on deleting the cache when I get home from work, as well as finishing the FM->MIF->FM work.

[MrVideoSenior]

OK, I've done the following items to my system:

1) Installed 7.2p158

2) Deleted fntcache.dat/rebooted

3) FM -> MIF -> FM

4) Created new book from #3 FM files

I have no clue as to which of the above has solved my problem (or combination of above).  I doubt that it was #1, as the README file gave no indication of the issue, plus the issue existed into FM8 and FM9.

Thanks for all the pointers.

All 298 pages appear to be in correct shape.  It is still distilling, but a grep for GreNewFont in the PostScript turned up nothing.

[AxialInfo]

Great news, glad to hear it.

If the issue was resolved by deleting the fntcache.dat and rebooting, then you'll have to adopt some method of doing that "regularly" -- there's no way to know how often would be needed, it's a matter of how often you use FM, how many fonts, etc etc, there's no metric possible. If you're using FM all day, I'd suggest at least once a day.

btw, though, assuming this is a similar issue to that fixed by the hotfix in XP, the "missing font problem" did affect other apps, e.g. Word, it wasn't just FM.

Sheila

[MrVideoSenior]

At this point in time I do not know which piece actually fixed it, but I am leaning toward the cache file.

I don't use Frame all that often, so I don't expect to run across it much.  But I do know what to try the next time.