yes I am aware of this problem. Someone posted a bugfix of bbox calculation (very long time ago). When I was digging my mail box containing pdftex bug reports, I found that post and applied without careful testing. My apology for that. 1.20a-rc3 no longer has this bug. Regards, Thanh On Wed, May 12, 2004 at 01:50:41AM +0200, Hartmut Henkel wrote:

Hi,

in pdftex as of teTeX 2.96.2 there seems to be a problem with reading pk fonts, e. g.

\font\aaa xbmc10 at 10pt \aaa hi \bye

(xbmc10.mf see pdfetex.web. BTW, this xbmc10.mf is missing in teTeX)

I get "bad pk file, expected pre" from readchar() of pkin.c, but the xbmc10.600pk is ok, starting with preamble 247. The problem is somehow, that the check_preamble is true even after the 1st byte read, so that later in pk file read suddenly a preamble check is done. This has somehow to do with t3_glyph_num in writepk() of writet3.c. The kludge, which seems to work, is to have some

boolean checkpreamble=true;

at the begin of writepk(), and then have the while loop like:

while (readchar(checkpreamble, &cd) != 0) { checkpreamble=false;

so that the header is only checked once at start of file read. But this might cover some problem with t3_glyph_num...

Regards Hartmut _______________________________________________ ntg-pdftex mailing list ntg-pdftex@ntg.nl http://www.ntg.nl/mailman/listinfo/ntg-pdftex