Bug#440478: gs-esp: hangs at ~98% CPU with DEVICE=bbox reading a .ps from tiff2ps

[Jonathan Nieder <jrnieder@gmail.com>]

reassign 440478 ghostscript
found 440478 gs-esp/8.15.3.dfsg.1-1
found 440478 ghostscript/9.00~dfsg-2
tags 440478 + confirmed upstream
quit

Hi Paolo,

Paolo wrote:

> let PS file be fax.ps:
>
> - gv fax.ps              -> shows up on screen correctly
> - ps2pdf fax.ps fax.pdf  -> xpdf shows just a tiny white square
> - ps2ps fax.ps fax2.ps   -> bbox is 0 0 0 0, gv fax2.ps -> blank
> - gs -sDEVICE=lj4dith    -> prints just ~70 char of pjl cmd - nothing gets
>                             printed on printer
> - gs -sDEVICE=bbox       -> gs hangs ~98% CPU forever, strace -f shows:

Interesting.  Yes, I can reproduce this:

 -   gs -dSAFER fax.ps;	# just shows a white screen.
 -   evince fax.ps;	# shows the fax.
 - $ gs -dSAFER -sDEVICE=bbox -sOutputFile=/dev/null fax.ps
   GPL Ghostscript  9.00 (2010-09-14)
   Copyright (C) 2010 Artifex Software, Inc.  All rights reserved.
   This software comes with NO WARRANTY: see the file PUBLIC for details.
   [hangs here]

Backtrace at a random moment:

| #0  zoom_y (st=0xb2697ac, pr=0xbfc24734, pw=0xbfc24728, last=0) at ./base/siscale.c:388
| #1  s_IScale_process (st=0xb2697ac, pr=0xbfc24734, pw=0xbfc24728, last=0) at ./base/siscale.c:594
| #2  image_render_interpolate_icc (penum=0xb108214, buffer=0xb2690d0 "\377\377[... etc ...]\377"..., data_x=0, iw=1728, h=1, dev=0xb269fb4) at ./base/gxiscale.c:1019
| #3  gx_image1_plane_data (info=0xb108214, planes=0xb26a75c, height=9, rows_used=0xbfc25d6c) at ./base/gxidata.c:209
| #4  bbox_image_plane_data (info=0xb20c1cc, planes=0xb26a75c, height=9, rows_used=0xbfc25d6c) at ./base/gdevbbox.c:1063
| #5  bbox_image_plane_data (info=0xb20c0f8, planes=0xb26a75c, height=9, rows_used=0xbfc25d6c) at ./base/gdevbbox.c:1063
| #6  gs_image_next_planes (penum=0xb26a5ec, plane_data=0xbfc25ddc, used=0xbfc25e54) at ./base/gsimage.c:600
| #7  image_file_continue (i_ctx_p=0xb033ef8) at ./psi/zimage.c:549
| #8  in interp (pi_ctx_p=0xb00ba14, pref=<value optimized out>, perror_object=0xbfc26bf4) at ./psi/interp.c:1150
| #9  gs_call_interp (pi_ctx_p=0xb00ba14, pref=0xbfc26b78, user_errors=1, pexit_code=0xbfc26bfc, perror_object=0xbfc26bf4) at ./psi/interp.c:484
| #10 gs_interpret (pi_ctx_p=0xb00ba14, pref=0xbfc26b78, user_errors=1, pexit_code=0xbfc26bfc, perror_object=0xbfc26bf4) at ./psi/interp.c:442
| #11 gs_main_interpret (minst=0xb00b9c0, user_errors=1, pexit_code=0xbfc26bfc, perror_object=0xbfc26bf4) at ./psi/imain.c:240
| #12 gs_main_run_string_end (minst=0xb00b9c0, user_errors=1, pexit_code=0xbfc26bfc, perror_object=0xbfc26bf4) at ./psi/imain.c:556
| #13 gs_main_run_string (minst=0xb00b9c0, str=0xb130be8 "<2f746d702f6661782e7073>.runfile", user_errors=1, pexit_code=0xbfc26bfc, perror_object=0xbfc26bf4) at ./psi/imain.c:496
| #14 run_string (minst=0xb00b9c0, str=<value optimized out>, options=<value optimized out>) at ./psi/imainarg.c:814
| #15 runarg (minst=0xb00b9c0, pre=<value optimized out>, arg=<value optimized out>, post=0x84ba3e5 ".runfile", options=3) at ./psi/imainarg.c:805
| #16 argproc (minst=0xb00b9c0, arg=0xbfc27cc4 "/tmp/fax.ps") at ./psi/imainarg.c:738
| #17 gs_main_init_with_args (minst=0xb00b9c0, argc=5, argv=0xbfc27694) at ./psi/imainarg.c:215
| #18 main (argc=5, argv=0xbfc27694) at ./psi/gs.c:96

Do you know of a ghostscript version that did not have this problem?
(That could make life very easy.)

> fax.ps (~10kB) pasted at bottom - hope it gets through

Attached to this message in more wasteful MIME format for the
convenience of others. ;-)

Thanks for a clear report.