windley@iris (Phil Windley) (11/21/87)
When I compile microemacs 3.9 under Unix (HP-UX) I get a list of undefined symbols out of it that I don't recognize. I thought for a minute that they occured in files specific to non-unix systems and I could solve the problem by not including them in the make (is there a good reason why they are, by the way?). Unfortunately, my hopes were dashed when I noticed that sysint, for instance, is called from main.c. What to do? Has anyone made microemacs under Unix? I would really like to get an emacs on this machine, but I don't have room for GNU. Here is the snapshot of the problem: | # make | make | cc -O ansi.o basic.o bind.o buffer.o crypt.o dg10.o display.o eval.o | exec.o file.o fileio.o hp110.o hp150.o ibmpc.o input.o isearch.o line.o | lock.o main.o random.o region.o search.o spawn.o st520.o tcap.o termio.o | tipc.o vmsvt.o vt52.o window.o word.o z309.o -ltermcap -lc -o emacs | ld: Undefined external - | _aputc | _inportb | _sysint | _outportb | _bdos | _movmem | _segread | ld: output file still contains undefined symbols | ld: (Warning) did not generate an output file | *** Error code 1 Phil Windley Robotics Research Lab University of California, Davis
hartzell@boulder.Colorado.EDU (George Hartzell) (11/21/87)
In article <538@ucdavis.ucdavis.edu> windley@iris.ucdavis.edu (Phil Windley) writes: >| ld: Undefined external - >| _aputc >| _inportb >| _sysint >| _outportb >| _bdos >| _movmem >| _segread >| ld: (Warning) did not generate an output file These look an awfull lot like IBM things. You should check the .h files (don't remember which one at the moment) for the one in which you define the system that you are using and make sure that you have IBM turned off. g. George Hartzell (303) 492-4535 MCD Biology, University of Colorado-Boulder, Boulder, CO 80309 hartzell@Boulder.Colorado.EDU ..!{hao,nbires}!boulder!hartzell
ljz@fxgrp.UUCP (11/22/87)
In article <538@ucdavis.ucdavis.edu> windley@iris.ucdavis.edu (Phil Windley) writes: >When I compile microemacs 3.9 under Unix (HP-UX) I get a list of >undefined symbols out of it that I don't recognize. ... > ... >| ld: Undefined external - >| _aputc >| _inportb >| _sysint >| _outportb >| _bdos >| _movmem >| _segread > ... Look at the estruct.h file. In it are a bunch of #define symbols, some of which need to be reset when compiling under unix. The default is to set up uEmacs for an MSDOS machine, so you will have to at least set the MSDOS symbol to 0, and you might have to change some of the other #define values. It should be fairly easy to figure out which #define's need to be set in which way, as the estruct.h file is documented relatively well. --------------------------------------------------------------------------- Lloyd Zusman Master Byte Software Los Gatos, California UUCP: ...!ames!fxgrp!ljz "We take things well in hand." INTERNET: fxgrp!ljz@ames.arpa ---------------------------------------------------------------------------
tmoody@sjuvax.UUCP (T. Moody) (11/23/87)
In article <538@ucdavis.ucdavis.edu> windley@iris.ucdavis.edu (Phil Windley) writes: >When I compile microemacs 3.9 under Unix (HP-UX) I get a list of >undefined symbols out of it that I don't recognize.[...] >Here is the snapshot of the problem: >[more deleted stuff] >| ld: Undefined external - >| _aputc >| _inportb >| _sysint >| _outportb >| _bdos >| _movmem >| _segread >| ld: output file still contains undefined symbols >| ld: (Warning) did not generate an output file >| *** Error code 1 Thank you, Phil, for posting this. This is pretty much the same result that I got under BSD 4.3 but, since I am not a C programmer (or any kind of a programmer) but a -- how you say -- end user, I assumed that I was doing something seriously wrong. The strange thing (to me) is that as far as I can tell, these things *are* defined. At least, they don't look different from the other definitions in the code that do not appear to be causing problems. A good number of these "undefined" items are linked to MSDOS-specific parts of estruct.h. I tried masking things out, but succeeded only in making a big mess. Under BSD 4.3, an output file was created, but it would not execute. Perhaps others have had the same troubles. [Note: medoc4 also did not survive the trip to this site.] -- Todd Moody * {allegra|astrovax|bpa|burdvax}!sjuvax!tmoody * SJU Phil. Dept. "The wind is not moving. The flag is not moving. Mind is moving."
nwd@j.cc.purdue.edu (Daniel Lawrence) (11/23/87)
In article <538@ucdavis.ucdavis.edu> windley@iris.ucdavis.edu (Phil Windley) writes: >When I compile microemacs 3.9 under Unix (HP-UX) I get a list of >undefined symbols out of it that I don't recognize. >| ld: Undefined external - >| _aputc _inportb _sysint _outportb >| _bdos _movmem _segread A lot of people are trying to compile this cold, without first modifying the configuration flags in ESTRUCT.H which specify what system/compiler/features to use. The above undefined symbols are library routines in the MSDOS version of the AZTEC compiler (which is what ESTRUCT.H was set to when I archived it). Please edit ESTRUCT.H before you try to compile MicroEMACS 3.9e. Daniel Lawrence (317) 742-5153 UUCP: {ihnp4!pur-ee!}j.cc.purdue.edu!nwd ARPA: nwd@j.cc.purdue.edu FIDO: 201/2 The Programmer's Room (317) 742-5533
ljz@fxgrp.UUCP (Lloyd Zusman, Master Byte Software) (11/24/87)
In article <999@sjuvax.UUCP> tmoody@sjuvax.UUCP (T. Moody) writes: >In article <538@ucdavis.ucdavis.edu> windley@iris.ucdavis.edu (Phil Windley) writes: > ... >>| ld: Undefined external - >>| _aputc >>| _inportb >>| _sysint >>| _outportb >>| _bdos >>| _movmem >>| _segread >>| ld: output file still contains undefined symbols >>| ld: (Warning) did not generate an output file >>| *** Error code 1 > > ... A >good number of these "undefined" items are linked to MSDOS-specific >parts of estruct.h. I tried masking things out, but succeeded only in >making a big mess. > >Under BSD 4.3, an output file was created, but it would not execute. > ... I got it to build just fine under BSD 4.3 and on a Sun. Re-extract the files from the shar's ... this will make sure you are starting with a clean slate. Then, make sure your "estruct.h" file looks like the following fragment. Copy or rename the "makefile.unx" file to "Makefile" or "makefile", get rid of all the *.o files, and then type 'make'. It should work just fine. /** fragment of estruct.h starts here **/ /* Machine/OS definitions */ #define AMIGA 0 /* AmigaDOS */ #define ST520 0 /* ST520, TOS */ #define MSDOS 0 /* MS-DOS */ #define V7 0 /* V7 UNIX or Coherent or BSD4.2*/ #define BSD 1 /* UNIX BSD 4.2 and ULTRIX */ #define USG 0 /* UNIX system V */ #define VMS 0 /* VAX/VMS */ #define CPM 0 /* CP/M-86 */ #define FINDER 0 /* Macintosh OS */ /* Compiler definitions */ #define UNIX 1 /* a random UNIX compiler */ #define MWC 0 /* Marc Williams C */ #define LATTICE 0 /* Lattice 2.14 thruough 3.0 compilers */ #define AZTEC 0 /* Aztec C 3.20e */ #define MSC 0 /* MicroSoft C compile version 3 & 4 */ #define TURBO 0 /* Turbo C/MSDOS */ /* Debugging options */ #define RAMSIZE 0 /* dynamic RAM memory usage tracking */ #define RAMSHOW 0 /* auto dynamic RAM reporting */ /* Special keyboard definitions */ #define WANGPC 0 /* WangPC - mostly escape sequences */ #define VT100 0 /* Handle VT100 style keypad. */ /* Terminal Output definitions */ #define ANSI 0 /* ANSI escape sequences */ #define HP150 0 /* HP150 screen driver */ #define HP110 0 /* HP110 screen driver */ #define VMSVT 0 /* various VMS terminal entries */ #define VT52 0 /* VT52 terminal (Zenith). */ #define RAINBOW 0 /* Use Rainbow fast video. */ #define TERMCAP 1 /* Use TERMCAP */ #define IBMPC 0 /* IBM-PC CGA/MONO/EGA driver */ #define DG10 0 /* Data General system/10 */ #define TIPC 0 /* TI Profesional PC driver */ #define Z309 0 /* Zenith 100 PC family driver */ #define MAC 0 /* Macintosh */ #define ATARI 0 /* Atari 520/1040ST screen */ /* Configuration options */ #define CVMVAS 1 /* arguments to page forward/back in pages */ #define CLRMSG 0 /* space clears the message line with no insert */ #define CFENCE 1 /* fench matching in CMODE */ #define TYPEAH 1 /* type ahead causes update to be skipped */ #define DEBUGM 1 /* $debug triggers macro debugging */ #define VISMAC 0 /* update display during keyboard macros */ #define CTRLZ 0 /* add a ^Z at end of files under MSDOS only */ #define ADDCR 0 /* ajout d'un CR en fin de chaque ligne (ST520) */ #define NBRACE 1 /* new style brace matching command */ #define REVSTA 1 /* Status line appears in reverse video */ #define COLOR 0 /* color commands and windows */ #define FILOCK 0 /* file locking under unix BSD 4.2 */ #define ISRCH 0 /* Incremental searches like ITS EMACS */ #define WORDPRO 1 /* Advanced word processing features */ #define FLABEL 0 /* function key label code [HP150] */ #define APROP 1 /* Add code for Apropos command */ #define CRYPT 1 /* file encryption enabled? */ #define MAGIC 1 /* include regular expression matching? */ #define AEDIT 1 /* advanced editing options: en/detabbing */ #define PROC 1 /* named procedures */ #define CLEAN 0 /* de-alloc memory on exit */ #define CALLED 0 /* is emacs a called subroutine? or stand alone */ #define ASCII 1 /* always using ASCII char sequences for now */ #define EBCDIC 0 /* later IBM mainfraim versions will use EBCDIC */ /** fragment of estruct.h ends here **/ Good luck. --------------------------------------------------------------------------- Lloyd Zusman Master Byte Software Los Gatos, California UUCP: ...!ames!fxgrp!ljz "We take things well in hand." INTERNET: fxgrp!ljz@ames.arpa --------------------------------------------------------------------------- -- Lloyd Zusman, Master Byte Software, Los Gatos, California "We take things well in hand." ...!ames!fxgrp!ljz
exodus@uop.EDU (G. Onufer) (11/26/87)
I compiled me3.9 on our Sun systems here. We use Televideo 905's for the
most part and there seems to be a small problem with them. me3.9 starts
repainting the screen at column 1 (as opposed to column 0) and when you
start entering text, it is double spaced (ie: type "aaa" get "a a a").
It works fine when the Televideo is emulating something else (I cannot remember
which terminal I used...the emulation is internal to the terminal). And it
works fine on any other terminals in the lab.
Anybody have a similar problem?
Thanks,
Greg
Greg Onufer GEnie: G.ONUFER University of the Pacific
UUCP: **POSTMASTER @ UOP**
{ucbvax, lll-crg}!ucdavis!uop.edu
!{exodus, exodusr, postmaster, root}
{ptsfa!cogent, cepu!retix}!uop.edu [Any of the above names will work]
AT&T: 209-944-7481 USMAIL: #300 Jackson, UoP, Stockton CA, 95211george@rebel.UUCP (George M. Sipe) (11/27/87)
In article <791@uop.EDU> exodus@uop.EDU (G. Onufer) writes: >I compiled me3.9 on our Sun systems here. We use Televideo 905's for the >most part and there seems to be a small problem with them. me3.9 starts >repainting the screen at column 1 (as opposed to column 0) and when you >start entering text, it is double spaced (ie: type "aaa" get "a a a"). >It works fine when the Televideo is emulating something else (I cannot remember >which terminal I used...the emulation is internal to the terminal). And it >works fine on any other terminals in the lab. >Anybody have a similar problem? It sure does. This is the standout problem which has always been in MicroEMACS. Terminals which require embedded attributes are not supported. I have posted patches for 3.8l, 3.8i, and 3.9e which fully correct this problem. The 3.9e patches were just posted to comp.emacs and comp.sources.bugs. Daniel Lawrence, author of MicroEMACS has promised to include them in future releases. I will NOT post the patches for any future versions (three times is more than enough). -- George M. Sipe, Phone: (404) 662-1533 Tolerant Systems, 6961 Peachtree Industrial, Norcross, GA 30071 UUCP: ...!{decvax,hplabs,ihnp4,linus,rutgers,seismo}!gatech!rebel!george
schwartz@gondor.psu.edu (Scott E. Schwartz) (11/30/87)
In article <791@uop.EDU> exodus@uop.EDU (G. Onufer) writes: > >I compiled me3.9 on our Sun systems here. We use Televideo 905's for the >most part and there seems to be a small problem with them. me3.9 starts >repainting the screen at column 1 (as opposed to column 0) and when you >start entering text, it is double spaced (ie: type "aaa" get "a a a"). I have not seed 3.9 yet, but on all previous versions of microemacs (and lots of other programs, for that matter) the author uses the termcap entries "so" and "se" to highlight characters, but chooses to ignore "sg" which tells how much space the magic cookies take up. Uemacs is printing so and se and not realizing that they screw up the output on your terminal. Solutions: you can patch uemacs to never use so/se (easy), you can fix uemacs to do it right (someone posted patches for 3.7 last year), or you can alter your termcap entry to not define them. In any event, someone should let the author know this is going wrong. -- Scott Schwartz schwartz@gondor.psu.edu