Re: sgml-lex

"Richard A. O'Keefe" wrote:
> 
> I just picked sgml-lex up and tried to build it.
> Problem 1:  it assumes that $(CC) = gcc, which is not true.
> Fixed the easy way (by defining CC in Makefile).

Really? I intended it to be pure ANSI C. Details?

> Problem 2:  it assumes that characters are unsigned.  gcc -Wall
> -Werror doesn't like using signed characters as array subscripts,
> and rightly so.  Fixed by adding (unsigned char) in three places.

ok.

> Problem 3:  suddenly everything broke.  It turns out that this
> thing depends on details of the code generated by flex, and those
> details have _changed_.  Fixed by patching excerpt.pl to recognise
> the new pattern and by patching sgml_lex.c to use the new pattern
> itself.

guilty.

> Problem 4:  `make' tries to make filter, which depends on filter.o,
> but there is no filter.c, and no rule for making filter.o from
> anything else.  This is a real killer and I am completely stuck.

Sorry... but...

note the README file says

|The "home" of this project on the web is:
|
|        http://www.w3.org/pub/WWW/MarkUp/SGML/#sgml-lex

which links "missing filter.c"
to

http://www.w3.org/MarkUp/SGML/sgml-lex/filter.c


> I would be delighted to send you the patches I have so far if they
> are of any interest.

Please do, and please copy www-html.

> You might also like to know that the sgml-lex web page has a number
> of out-of-date links:  the SGML stuff at SIL has moved to Oasis.

Yes, and they didn't put HTTP redirects at the old address.
Grumble.
http://www.w3.org/Provider/Style/URI.html

-- 
Dan Connolly, W3C
http://www.w3.org/People/Connolly/

Received on Friday, 28 May 1999 18:36:31 UTC