| Summary: | ps2write (NOT pdfwrite) produces invalid output for some files | ||
|---|---|---|---|
| Product: | Ghostscript | Reporter: | Ken Sharp <ken.sharp> |
| Component: | PDF Writer | Assignee: | Ken Sharp <ken.sharp> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | ||
| Priority: | P4 | ||
| Version: | master | ||
| Hardware: | PC | ||
| OS: | Windows Vista | ||
| Customer: | Word Size: | --- | |
|
Description
Ken Sharp
2010-12-14 14:55:09 UTC
Analysis of the errors shows: FAPI+FreeType issue see Bug #691848 ------------------------------------- 0000626-Q5471Untitled1.ps, 000516-1.pdf, 067a.pdf, 159.pdf, 24-25.ps, 55550j.pdf, 9-11.pdf, Altona.Page_3.2002-09-27.pdf, Altona.pdf, AltonaTech_1v2_pt2com_x3.pdf, Altona_Technical_1v1_x3.pdf, altona_technical_1v2_x3.pdf, Altona_Technical_x3.pdf, Altona_Technical_x3stripped.pdf, ALWA SYSTEM MAPS.pdf, annoncesonder0105m.pdf, apple -- intervention jaussaud 6.pdf, BCL easy.pdf, BFAZHHPD.PDF, bug687603.ps, Bug687724.pdf, Bug687901.ps, Bug688421.pdf, Bug688467.ps, Bug689014.pdf, Bug689083.pdf, Bug689516.pdf, Bug689754.ps, Bug689962.pdf, Bug690348.pdf, Bug690450.pdf, Bug690478.pdf, bug_687457.pdf, C262-PSCutDown.ps, CATX0198.pdf, CATX2376.pdf, CATX2579.pdf, CATX2975.pdf, CATX3740.pdf, CATX4030.pdf, CATX4135.pdf, CATX5185.pdf, CATX6460.pdf, CATX8839.pdf, CATX9201.pdf, endless20040109.ps, err_file.pdf, err_file.ps, err_file0.pdf, Faktura.pdf, FT601DP1.PS, IA3Z0248.pdf, IA3Z0658.pdf, IA3Z1351.pdf, IA3Z3778.pdf, IA3Z4488.pdf, IA3Z4630.pdf, IE553CP1.pdf, IE553CP1.ps, IE554CP1.ps, LP20_020-1.pdf, objlune-notext.pdf, Pages from BFAZHHPD.pdf, PDFEditorSDK2.4.SDP.20040603.pdf, pscharstrings5.ps, Quiberon A0.pdf, raapke.prn, rotate.pdf, S2_Digitalproof-Forum_x3k.pdf, Sal_afrek_100_1_20050201.pdf, SolimarCR3324.ps, source.ps, test_multipage_prob.pdf, TT2029.pdf, type42_glyph_index.ps memory errors caused by transparency see Bug #689805. ------------------------------------------------------- 02 (2).pdf, 02.pdf, 02ok_11h.pdf 200200697.pdf, Accelrys Track Signs A1_UK_PR.pdf, Bug688778.pdf, Bug689369.pdf, Bug689492.pdf, Bug689895.pdf, Bug689918.pdf, Bug689982.pdf, Bug690022.pdf, INTEGER_TEST_5COLOR_PS2_-output.pdf, Murphys_trap_limecmyk2.pdf, PolorLogoCirc_greenV9.pdf, Tran_30-1-unc.pdf, Tran_30-2-unc.pdf, Tran_30-3-unc.pdf, Tran_30-4-unc.pdf, Tran_30-5-unc.pdf, Tran_30.pdf CATX3506.pdf.disabled, uses a CIDFont, which is not embedded, needs a defined replacement, this not a bug. This leaves the following files still needing further investigation --------------------------------------------------------------------- Pages from AC507122.pdf, nocurrentpoint in --run--, works OK with non-FAPI GS. tsts.objects copy 1.ps, tsts.objects copy 11.ps, patterntest.ps rangecheck in --rectfill-- during conversion to ps2write. zweiseit01d4-1.pdf, limitcheck --nostringval-- patterntest.ps contains a pattern which uses a DeviceN colour space. Since DeviceN is a level 3 construct this is why we get a rangecheck. Use of DeviceN also causes the rangecheck errors in tsts.objects copy 1.ps and tsts.objects copy 11.ps. revision 11969 fixes the problem with zweiseit01d4-1.pdf. When rendering a font to bitmaps in order to construct a type 3 bitmap font (in this case as a replacement for a CFF font), glyphs which had already been cached were not updating the show enumerator with the glyph width. However, the width was being used to set the /Widths array for the type 3 font. Using the uninitialised value from the show enumerator could lead to random garbage values in the Widths array, which caused the error. We now update the show enumerator 'wxy' value from the cached character 'wxy' value. The remaining file is a duplicate of the problem raised in bug #691927, we'll continue with that bug and close this one as all the other problems in this report are fixed. |