From: SMTP%"tillman@swdev.si.com" 23-FEB-1995 14:10:52.55 To: EVERHART CC: Subj: RE: Problem building/starting Ghostscript 3.12 X-Newsgroups: comp.os.vms From: tillman@swdev.si.com (Brian Tillman) Subject: RE: Problem building/starting Ghostscript 3.12 Message-ID: <23FEB95.11305203@swdev.si.com> Sender: news@esseye.si.com Nntp-Posting-Host: terwed.si.com Organization: Smiths Industries Aerospace Date: Thu, 23 Feb 1995 11:30:52 GMT Lines: 157 To: Info-VAX@Mvb.Saic.Com X-Gateway-Source-Info: USENET In a previous article, rccarlo@urc.tue.nl (Carlo den Otter) wrote: -> ->-- ->I want to build a new version, 3.12, of Ghostscript (VMS) on both VAX and ->Alpha. -> ->On the Alpha (with VMS 6.1 and Motif 1.2) I put the sources and fonts in one ->directory, and the JPEG-stuff in [.JPEGP]. ->I applied all mentioned fixes. Everything compiled nicely, and ->linking went fine too (apart from a warning about a missing "UNLINK"). -> ->I defined GS_LIB to point to the directory containing GS. ->When I run GS (except for GS -h) however, the executable leaves me stranded ->with the following error: -> -> Unexpected interpreter error -21. -> Error object: -> Operand stack at 0x1306e0:5c57d8#106)dup -> 0x5a8888: 0x0b int ---------- 0x7fb5 0x00000138 = 312 -> Execution stack at 0x130340: -> 0x5aa1f0: 0x0f oper -------e-- 0x0000 0x0001c1a0 = %interp_exit -> 0x5aa1f8: 0x03 file -----rxe-- 0x0001 0x005c5e78 -> %NONAME-E-NOMSG, Message number 00000002 -> ->Can anybody tell me how to solve this? -> ->I haven't tried to build GS on a VAX yet, but should I expect simmilar ->problems? -> ->Thanks for your help. Where you got Ghostscript should also be another file that describes bugs in the Ghostscript distribution. I'll include it below. Once I followed its suggestions, the above problem disappeared. This is intended to be a short list of fixes to minor problems discovered after the release. If this causes confusion, we will delete it. Each version of this file will include ALL fixes, so you only need to retrieve the latest version. Please note that the fixes are in chronological order. ---------------- v1 ---------------- On VMS systems, the file GCONFIGV.H is not created automatically. You should create this file by hand, with the following 2 lines as contents: #define USE_ASM 0 #define USE_FPU 1 ---------------- v2 ---------------- The installation and build instructions do not tell you that the ..jpeg.tar.gz and sr3.zip file do not include gs3.12/ in the file names. You must cd to gs3.12 before unpacking these files. ---------------- v3 ---------------- The makefile rule for creating gconfig.h has two blank lines in the middle of it (but not next to each other); delete them. gdevdfax.c contains the line state.EncodedEOLAlign = false; which refers to an obsolete structure member; delete this line. ---------------- v4 ---------------- setpagedevice gives an error if presented with an option it doesn't recognize. To work around this, in zdevice.c, change line 211 from code = stack_param_list_read(&list, &o_stack, 0, ppolicy,true); to code = stack_param_list_read(&list, &o_stack, 0, ppolicy, false); and change line 219 from if ( list.results[i] != 1 ) to if ( list.results[i] < 0 ) Note that this is only a workaround; the problem will be fixed in a different way in the next release. ---------------- v5 ---------------- The -dDISKFONTS switch doesn't work; don't use it. The page size of image devices is always set to (0, 0). This has bad effects on many utility programs. To fix this, change lines 374-377 of gsdevice.c from new_dev->width = width; new_dev->height = height; new_dev->x_pixels_per_inch = fabs(x_pixels_per_unit) * 72; new_dev->y_pixels_per_inch = fabs(y_pixels_per_unit) * 72; to new_dev->x_pixels_per_inch = fabs(x_pixels_per_unit) * 72; new_dev->y_pixels_per_inch = fabs(y_pixels_per_unit) * 72; gx_device_set_width_height((gs_device *)new_dev, width, height); ---------------- v6 ---------------- The above v5 fix has a typo in it: the last line should read gx_device_set_width_height((gx_device *)new_dev, width, height); i.e. gx_device * rather than gs_device *. ---------------- v7 ---------------- The VMS script files have two errors that prevent proper compilation. To fix them: 1) Check GSSETDEV.COM, GSSETMOD.COM, and GSADDMOD.COM and remove any blank lines or extraneous characters. 2) Replace the LEVEL1_DEV section of VMS.MAK with the following: $ LEVEL1_DEV: $ GOSUB TYPE1_DEV $ SETMOD level1 $ ADDMOD level1 -include type1 $ ADDMOD level1 -oper zarith zarray zcontrol zdict $ ADDMOD level1 -oper zfile zfileio zfilter zfproc $ ADDMOD level1 -oper zgeneric ziodev zmath zmisc zpacked $ ADDMOD level1 -oper zrelbit zstack zstring $ ADDMOD level1 -oper ztoken ztype zvmem $ ADDMOD level1 -oper zchar zcolor zdevice zfont zfont2 $ ADDMOD level1 -oper zgstate zht zmatrix zpaint zpath zpath2 $ RETURN ---------------- v8 ---------------- doesn't get included properly on some platforms (VMS in particular). To fix this, apply the following context diff patch to time_.h: *************** *** 33,36 **** --- 33,37 ---- # include #else + # include struct timeval { long tv_sec, tv_usec; *************** *** 39,46 **** int tz_minuteswest, tz_dsttime; }; - #endif - - #if defined(AUX) || defined(__EMX__) || defined(_IBMR2) || defined(__MSDOS__) || defined(SVR4) || defined(SYSV) || defined(__WIN32__) - # include #endif -------------------------------- end -------------------------------- -----------------------------+-------------------------------- Brian Tillman | Internet: tillman@swdev.si.com Smiths Industries, Inc. | tillman_brian@si.com 4141 Eastern Ave., MS129 | Hey, I said this stuff myself. Grand Rapids, MI 49518-8727 | My company has no part in it. -----------------------------+--------------------------------