dlu@tektools.UUCP (Doug Urner) (03/25/86)
Has anyone out there brought ditroff up on a Gould and lived to tell about it? I would be interested in knowing what you had to do to get it working. A quick look at the nroff sources from Gould make the task look formidable. Doug Urner, dlu@tektronix
trt@rti-sel.UUCP (Tom Truscott) (03/28/86)
> Has anyone out there brought ditroff up on a Gould and lived to tell ...
We have had ditroff (Lou Salkind's revised standard version)
running on our Gould for a couple years now.
The main (only?) problem with troff was code of the form:
unsigned int t;
if (t == -1) ... /* Never true on UTX */
I fixed the N places in the code with
if (t == (unsigned int)-1) /* portability, the futile frontier */
There is also a problem in pic due to a 'bug' in yacc:
objects of type YYSTYPE are not aligned correctly when makeattr is called
(the second arg is correctly declared YYSTYPE in misc.c
but not passed that way from pic.y).
The symptom is that pic core dumps, the fix is to add a middle
dummy argument in calls to makeattr so that the alignment is correct.
------------------
On a vaguely related subject, we are expecting release 2
of the Documenter's Workbench any day now. Is anyone else expecting same,
or are we naive? We have been unable to get any documentation
on it, of any sort, and indeed documentation is claimed to be the
reason it was not out in early March.
It has been six years since Brian Kernighan knocked out ditroff,
and that is plenty of time to improve it to support a zillion typesetters,
fonts, and macro packages, providing better graphics than old pic,
generalized index generation, and other utilities with simplify
document preparation and formatting. Is this a dream?
Tom Truscottgwyn@brl-smoke.UUCP (03/28/86)
In article <857@tektools.UUCP> dlu@tektools.UUCP (Doug Urner) writes: >Has anyone out there brought ditroff up on a Gould and lived to tell >about it? I would be interested in knowing what you had to do to get >it working. A quick look at the nroff sources from Gould make the task >look formidable. If you have UNIX System V Release 2 and DWB Release 1.0 source licenses, you can get the entire DWB package from me along with the BRL UNIX System V emulation for Gould UTX/32. <Gwyn@BRL.ARPA> I didn't have too much trouble getting it to work on the Gould, but of course I cheated and built it under a System V environment. The trickiest part was converting the VAX di10 binary fonts to Gould binary format; one has to account for differences in struct padding, etc. Does anybody know whether DWB Release 2.0 is available yet? I sure would like to have "grap".
gwyn@brl-smoke.UUCP (03/30/86)
In article <741@rti-sel.UUCP> trt@rti-sel.UUCP (Tom Truscott) writes: >There is also a problem in pic due to a 'bug' in yacc: >objects of type YYSTYPE are not aligned correctly when makeattr is called >(the second arg is correctly declared YYSTYPE in misc.c >but not passed that way from pic.y). >The symptom is that pic core dumps, the fix is to add a middle >dummy argument in calls to makeattr so that the alignment is correct. This isn't a proper fix. DWB 1.0 "pic" has a better one. >On a vaguely related subject, we are expecting release 2 >of the Documenter's Workbench any day now. Is anyone else expecting same, >or are we naive? We have been unable to get any documentation >on it, of any sort, and indeed documentation is claimed to be the >reason it was not out in early March. >It has been six years since Brian Kernighan knocked out ditroff, >and that is plenty of time to improve it to support a zillion typesetters, >fonts, and macro packages, providing better graphics than old pic, >generalized index generation, and other utilities with simplify >document preparation and formatting. Is this a dream? The only rumors I have on the contents of DWB 2.0 are that it includes more programmability in "pic", and a new preprocessor "grap" for generating graphs in typeset documents. These features are described in AT&T Bell Labs CSTRs #116 and #114. There are several things I could wish that AT&T would do for DWB 2.0 that I doubt they will, so I fully expect to have to merge my DWB 1.0 enhancements and bug fixes into 2.0. Sigh.