cscope-devel

[Cscope-devel] void * casts removed

[Hans-Bernhard B. <br...@ph...>]

As Petr signalled clearance to proceed, I commited the remaining
changes to remove casts of pointers to (void *) when passing them
to/from malloc()-family functions.

I also took the liberty to re-type the 'invterm' function to 'long'.
The return value of that function is completely unused, currently,
but the expression it does is of type 'long', not 'int', so that
seemed a good way to proceed.

Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

Re: [Cscope-devel] BL2 tag on May 5th 4pm EST

[Petr S. <pe...@sc...>]

Hi Joshua,

> > We need a BSD release as well (src and bin). Anybody got access to a BSD
> > machine?
> 
> Which BSD?  OpenBSD?  FreeBSD?  I might know someone who can do
> that for us.  Even then... the autoconf/automake package will be
> invaluable to us. :)
Aparantley NetBSD 1.4.2/i386 is the one to look for. But anyone should
be fine, here is an error list someone sent me with current top of tree
with my comments. I'd rather have someone test these out first though.

> I tried cscope on 1.4.2/i386, with the following results:
> 
> With curses:
> gcc -DHAVE_CONFIG_H -I. -I. -I..     -g -O2 -c input.c
> In file included from input.c:39:
> /usr/include/curses.h:69: warning: `erasechar' redefined
> constants.h:152: warning: this is the location of the previous definition
> /usr/include/curses.h:70: warning: `killchar' redefined
> constants.h:153: warning: this is the location of the previous definition
> input.c: In function `getline':
> input.c:112: invalid use of undefined type `struct termios'
> input.c:112: `VERASE' undeclared (first use in this function)
> input.c:112: (Each undeclared identifier is reported only once
> input.c:112: for each function it appears in.)
> input.c:119: invalid use of undefined type `struct termios'
> input.c:119: `VKILL' undeclared (first use in this function)
> *** Error code 1

The problem can be circumvented by placing an #ifndef __NetBSD__ #endif
around the definition of erasechar() and killchar() in curses.h

> 
> The redefined-warning comes for more than one file.
> 
> After replacing all <curses.h> with <ncurses.h> (didn't change the
> configure arguments):
> 
> gcc -DHAVE_CONFIG_H -I. -I. -I..     -I/usr/pkg/include -c command.c
> In file included from command.c:40:
> /usr/pkg/include/ncurses.h:413: arguments given to macro `cbreak'
> /usr/pkg/include/ncurses.h:441: arguments given to macro `erasechar'
> /usr/pkg/include/ncurses.h:510: arguments given to macro `killchar'
> /usr/pkg/include/ncurses.h:1139: warning: `KEY_BREAK' redefined
> constants.h:135: warning: this is the location of the previous definition
> /usr/pkg/include/ncurses.h:1145: warning: `KEY_BACKSPACE' redefined
> constants.h:141: warning: this is the location of the previous definition
> /usr/pkg/include/ncurses.h:1163: warning: `KEY_ENTER' redefined
> constants.h:138: warning: this is the location of the previous definition

Similar solution, except have also an ifndef __NetBSD__ for cbreak and
the KEY_*

> 
> This is against ncurses-5.0 from NetBSD's pkgsrc.
> 
> Independent of this I also get the following warning:
> gcc -DHAVE_CONFIG_H -I. -I. -I..     -g -O2 -c egrep.c
> egrep.y:162: warning: static declaration for `yylex' follows non-static
> egrep.y:618: warning: conflicting types for built-in function `memset'
Any ideas?

> Autoconf seems also to be missing at least one macro, where can I get
> it?
> autoconf: Undefined macros:
> configure.in:104:      AC_LINK_IFELSE([`cat $LEX_OUTPUT_ROOT.c`], ac_cv_prog_lex_yytext_pointer=yes)

Any ideas here as well?

Petr

> 
> --
> Joshua Uziel, Senior Linux Consultant, Linuxcare, Inc.
> 415.354.4878 tel, 415.701.7457 fax
> uz...@li..., http://www.linuxcare.com/
> Linuxcare. Support for the revolution.
> 
> _______________________________________________
> Cscope-devel mailing list
> Csc...@li...
> http://lists.sourceforge.net/mailman/listinfo/cscope-devel

-- 
--------------------------------------------------------
Petr Sorfa                  Software Engineer
Santa Cruz Operation (SCO)  
430 Mountain Ave.           http://www.sco.com
Murray Hill 07974
NJ, USA
--------------------------------------------------------
Disclaimer: All my comments are my own and nobody else's
----------------------------------------------------------

Re: [Cscope-devel] (no subject)

[Petr S. <pe...@sc...>]

Hi Hans,

My issue was with conflicting ANSI C documentation and a difference of
opinion between ANSI members which I'm in contact with. The issues have
been resolved, and its ok. So go ahead, but remember that the there is a
4pm EST cutoff today. So if you can't get the changes in by then, rather
have them done for bl3.

Petr
> After having worked on cscope as an outsider, for a while, Petr
> has added me in as a developer. I'd like to present some questions
> to y'all, regarding next steps I'ld like to make.
> 
> Of my patches offered so far, Petr mainly refused one kind: pointer casts
> with malloc() and friends, or their cscope-internal wrappers, mymalloc()
> and friends. Going with the C FAQ recommendations, I'ld like to get rid of
> all these casts. They serve no useful purpose in an ANSI C compiled
> program, for a start, but they tend to clutter the source code. They may
> even hide actual bugs if no proper prototype for the relevant function is
> visible to the compiler. To quote the comp.lang.c FAQ explanation of this:
> 
> --- snip ---
> 7.7:    Why does some code carefully cast the values returned by malloc
>         to the pointer type being allocated?
> 
> A:      Before ANSI/ISO Standard C introduced the void * generic pointer
>         type, these casts were typically required to silence warnings
>         (and perhaps induce conversions) when assigning between
>         incompatible pointer types.
> 
>         Under ANSI/ISO Standard C, these casts are no longer necessary,
>         and in fact modern practice discourages them, since they can
>         camouflage important warnings which would otherwise be generated
>         if malloc() happened not to be declared correctly; see question
>         7.6 above.  (However, the casts are typically seen in C code
>         which for one reason or another is intended to be compatible
>         with C++, where explicit casts from void * are required.)
> --- snip ---
> 
> I.e.: unless we want to make cscope a C++ program, in the near future, we
> might better get rid of those casts.
> 
> Hans-Bernhard Broeker (br...@ph...)
> Even if all the snow were burnt, ashes would remain.
> 
> _______________________________________________
> Cscope-devel mailing list
> Csc...@li...
> http://lists.sourceforge.net/mailman/listinfo/cscope-devel

-- 
--------------------------------------------------------
Petr Sorfa                  Software Engineer
Santa Cruz Operation (SCO) 
430 Mountain Ave.           http://www.sco.com
Murray Hill 07974
NJ, USA
--------------------------------------------------------
Disclaimer: All my comments are my own and nobody else's
----------------------------------------------------------

[Cscope-devel] (no subject)

[Hans-Bernhard B. <br...@ph...>]

Hello, everybody.

After having worked on cscope as an outsider, for a while, Petr
has added me in as a developer. I'd like to present some questions
to y'all, regarding next steps I'ld like to make.

Of my patches offered so far, Petr mainly refused one kind: pointer casts
with malloc() and friends, or their cscope-internal wrappers, mymalloc()
and friends. Going with the C FAQ recommendations, I'ld like to get rid of
all these casts. They serve no useful purpose in an ANSI C compiled
program, for a start, but they tend to clutter the source code. They may
even hide actual bugs if no proper prototype for the relevant function is
visible to the compiler. To quote the comp.lang.c FAQ explanation of this:

--- snip ---
7.7:    Why does some code carefully cast the values returned by malloc
        to the pointer type being allocated?

A:      Before ANSI/ISO Standard C introduced the void * generic pointer
        type, these casts were typically required to silence warnings
        (and perhaps induce conversions) when assigning between
        incompatible pointer types.

        Under ANSI/ISO Standard C, these casts are no longer necessary,
        and in fact modern practice discourages them, since they can
        camouflage important warnings which would otherwise be generated
        if malloc() happened not to be declared correctly; see question
        7.6 above.  (However, the casts are typically seen in C code
        which for one reason or another is intended to be compatible
        with C++, where explicit casts from void * are required.)
--- snip ---

I.e.: unless we want to make cscope a C++ program, in the near future, we
might better get rid of those casts.

Hans-Bernhard Broeker (br...@ph...)
Even if all the snow were burnt, ashes would remain.

[Cscope-devel] Re: autoconf finding flex libs under /usr/local

[Mike Hopkirk(hops) <ho...@sc...>]

> 
> > I'm trying to fix the config for finding flex in  /usr/local
> > but autoconf is proving its usual recalcitrant self.
> > I should have some tweaks for that either today or tomorrow...
> Yeah. I've run into the same problem with UW7. Once you remove flex from
> your $PATH all works fine.

OK - I've put some crowbars in configure.in so that it will now 
discover and use flex libs if they happen to be in /usr/local/lib.
Its a bit more extensive that desired due to some rather 
stupid interdependencies wired into automake macros 

This basically makes AC_CHECK_LIB use  -L/usr/local/lib when its checking
that a prog using yywrap works 
- I don't see this affecting any platform that doesn't have the libs there 
but are instead in  the default system lookup paths but pls let me know 
if theres any  probs.

While I was wallowing in that mire I also added handling for 
configure to  support explicit specification of the lexer and yacc binaries 
to use.
    --with-lex
    --with-flex
    --with-yacc
    --with-byacc
    --with-bison

run/build tested these on both osr5 and UW7.

petr - if you use bison to build cscope on UW7, the egrep source has refs to
alloca in it which then fail the link stage... This is probably 
handlable with an alloca in config.h and some creative macros 
(or we could add alloca.s src to the pkg)


-- 
- hops

Everything disclaimed (including disclaimer)
------<ho...@sc...>--------------------------------------
Mike Hopkirk (hops)  |       Whenever possible steal code.
SCO Inc              |          Tom Duff. (ex) Bell Labs

Re: [Cscope-devel] BL2 tag on May 5th 4pm EST

[Joshua U. <uz...@li...>]

* Petr Sorfa <pe...@sc...> [000504 14:03]:
> Hi Joshua,
> 
> We need a BSD release as well (src and bin). Anybody got access to a BSD
> machine?

Which BSD?  OpenBSD?  FreeBSD?  I might know someone who can do
that for us.  Even then... the autoconf/automake package will be
invaluable to us. :)

-- 
Joshua Uziel, Senior Linux Consultant, Linuxcare, Inc.
415.354.4878 tel, 415.701.7457 fax
uz...@li..., http://www.linuxcare.com/
Linuxcare. Support for the revolution.

Re: [Cscope-devel] BL2 tag on May 5th 4pm EST

[Petr S. <pe...@sc...>]

Hi Joshua,

We need a BSD release as well (src and bin). Anybody got access to a BSD
machine?

Petr
> * Petr Sorfa <pe...@us...> [000504 13:07]:
> > I'm tagging the source as of 4pm EST tomorrow as v15_0_bl2. This
> > means that the current top of tree is bl3.
> >
> > We should also make packages available of bl2 on Monday.
> 
> Great!  If you have any problems making the rpm, lemme know.
> That, or give me a src.rpm and I'll build a sparc rpm for you.
> 
> BTW, you only need to "make dist" and it'll make the .tar.gz
> file for you... :)
> 
> --
> Joshua Uziel, Senior Linux Consultant, Linuxcare, Inc.
> 415.354.4878 tel, 415.701.7457 fax
> uz...@li..., http://www.linuxcare.com/
> Linuxcare. Support for the revolution.

-- 
--------------------------------------------------------
Petr Sorfa                  Software Engineer
Santa Cruz Operation (SCO) 
430 Mountain Ave.           http://www.sco.com
Murray Hill 07974
NJ, USA
--------------------------------------------------------
Disclaimer: All my comments are my own and nobody else's
----------------------------------------------------------

Re: [Cscope-devel] BL2 tag on May 5th 4pm EST

[Joshua U. <uz...@li...>]

* Petr Sorfa <pe...@us...> [000504 13:07]:
> I'm tagging the source as of 4pm EST tomorrow as v15_0_bl2. This
> means that the current top of tree is bl3.
> 
> We should also make packages available of bl2 on Monday.

Great!  If you have any problems making the rpm, lemme know.
That, or give me a src.rpm and I'll build a sparc rpm for you.

BTW, you only need to "make dist" and it'll make the .tar.gz
file for you... :)

-- 
Joshua Uziel, Senior Linux Consultant, Linuxcare, Inc.
415.354.4878 tel, 415.701.7457 fax
uz...@li..., http://www.linuxcare.com/
Linuxcare. Support for the revolution.

[Cscope-devel] Welcome Hans-Bernhard Broeker

[Petr S. <pe...@us...>]

Hi All,

Hans-Bernhard Broeker has been added to the developer list.

Petr

[Cscope-devel] BL2 tag on May 5th 4pm EST

[Petr S. <pe...@us...>]

Hi Folks,

I'm tagging the source as of 4pm EST tomorrow as v15_0_bl2. This
means that the current top of tree is bl3.

We should also make packages available of bl2 on Monday.

Petr

Re: [Cscope-devel] uppercase letters for more than 9 lines

[Petr S. <pe...@sc...>]

Hi Mike,

> It seems to handle more than that (up to no of avail lines in window)
> e.g in a 80 line xterm it displays 60 lines of selections each
> selectable with a single char (0-9a-zA-Z) 0 varies on window size
> 
> It also makes the display dual windowed, using TAB to switch between the
> hits area and the command area.
> The selection area is paged so that
> SPC goes fwd and BS takes you to prev page (there probably more
> paging capability)
Ok. It sounds like a better implementation.

> all of those file systems are NFS exported ( or if you prefer I can
> generate a tarball and put in my FTP area).
Could you just bundle everything into a tarball?

> I'm trying to fix the config for finding flex in  /usr/local
> but autoconf is proving its usual recalcitrant self.
> I should have some tweaks for that either today or tomorrow...
Yeah. I've run into the same problem with UW7. Once you remove flex from
your $PATH all works fine.

Let me know when the tarball is available.

Petr
> 
> 
> --
> - hops
> 
> Everything disclaimed (including disclaimer)
> ------<ho...@sc...>--------------------------------------
> Mike Hopkirk (hops)  |       Whenever possible steal code.
> SCO Inc              |          Tom Duff. (ex) Bell Labs
> 
> _______________________________________________
> Cscope-devel mailing list
> Csc...@li...
> http://lists.sourceforge.net/mailman/listinfo/cscope-devel

-- 
--------------------------------------------------------
Petr Sorfa                  Software Engineer
Santa Cruz Operation (SCO)  
430 Mountain Ave.           http://www.sco.com
Murray Hill 07974
NJ, USA
--------------------------------------------------------
Disclaimer: All my comments are my own and nobody else's
----------------------------------------------------------

Re: [Cscope-devel] uppercase letters for more than 9 lines

[Mike Hopkirk(hops) <ho...@sc...>]

On Wed May 03,2000 (03:55:56PM -0400), pe...@sc...(Petr Sorfa) wrote:
> Hi Mike,
> 
> > > Following Joshua's suggestion, the selection is display in uppercase if
> > > it is a letter.
> > 
> > well its displaying a 9 lines with a 1 char selection then the remainder
> > with a 2 character one: - 0[A-Z].
> > ( source update from about10.00 this a.m.)
> Yes, that is correct.
> 
> > This and the >9 line display work is probably unnecessary effort since
> > the osr5 version of cscope already does this:
> >     i.e
> >  - Uses all the spare lines available for selection display (9-??)
> > and displays a  single letter for the selection
> > [0-9][a-z][A-Z] in that order.
> The way I understand it, the OSR5 version allows you to view 35
> selection lines?

It seems to handle more than that (up to no of avail lines in window)
e.g in a 80 line xterm it displays 60 lines of selections each
selectable with a single char (0-9a-zA-Z) 0 varies on window size 

It also makes the display dual windowed, using TAB to switch between the
hits area and the command area. 
The selection area is paged so that
SPC goes fwd and BS takes you to prev page (there probably more 
paging capability)


 The way its currently implemented gives you 45
> selection lines. Granted you need to type two characters to access
> certain lines. However, I've found this to be no hinderance. 
> Am I understanding this correctly?

some.
 
> > I'd suggest holding off on any further effort in either of these areas
> > till the osr5 differences are categorised and applied ...

> Mike, we need to get this code rolling in asap. Maybe you can pass the
> source to me, and I'll start working with it.

OK. Thats probably goanna be faster.

My machine charmstr.pdev has the requisite sources on it:
osrc tree and UW7 sources at 
/u1/hops/tmp/src/uw7/cscope
                cscope-UW7/     contains UW7 sources
                filesDiffFmOsr5     file has description of file differences
                        between osr5 cscope srcs and uw7 srcs
    osr5 cscope tree is at 
        /u2/osd/unixds/src/usr/src/cmd/cscope

all of those file systems are NFS exported ( or if you prefer I can
generate a tarball and put in my FTP area).

 
> I have another major patch coming along, and its going to make things
> more difficult as development progresses.
> 
> I'm thinking of tagging the current code as bl2 before applying the new
> patch. We'll probably need to make another tag for the OSR5 stuff.

I'm trying to fix the config for finding flex in  /usr/local
but autoconf is proving its usual recalcitrant self.
I should have some tweaks for that either today or tomorrow...
 

-- 
- hops

Everything disclaimed (including disclaimer)
------<ho...@sc...>--------------------------------------
Mike Hopkirk (hops)  |       Whenever possible steal code.
SCO Inc              |          Tom Duff. (ex) Bell Labs

Re: [Cscope-devel] Important message

[Joshua U. <uz...@li...>]

* Petr Sorfa <pe...@us...> [000503 15:51]:
> Since a lot of changes went in today, I would really appreciate if
> folks would test the current top of tree. Compilation and basic touch
> testing is ok.
> 
> I've tested on UW7 and Redhat 6.2 (i386) with ok results.

Works fine here for me as well (in RH 6.1)... I'll try Solaris when
I get home... and with Debian...

> Also I have extended and offer to Hans-Bernhard to be added as a developer.
> He has added a large amount of work that indicates a dedication and interest.
> If anybody has an issue with this please raise it soon.

I haven't had a look at the changes specifically... but either way,
he has my blessings. :)

-- 
Joshua Uziel, Senior Linux Consultant, Linuxcare, Inc.
415.354.4878 tel, 415.701.7457 fax
uz...@li..., http://www.linuxcare.com/
Linuxcare. Support for the revolution.

[Cscope-devel] more code cleanup

[Petr S. <pe...@us...>]

Hi,

More code cleanup and -Wall warnings removal, thanks to 
Hans-Bernhard Broeker.

Petr

[Cscope-devel] Important message

[Petr S. <pe...@us...>]

Hi,

Since a lot of changes went in today, I would really appreciate if
folks would test the current top of tree. Compilation and basic touch
testing is ok.

I've tested on UW7 and Redhat 6.2 (i386) with ok results.

Also I have extended and offer to Hans-Bernhard to be added as a developer.
He has added a large amount of work that indicates a dedication and interest.
If anybody has an issue with this please raise it soon.

Thanks,

Petr

Re: [Cscope-devel] uppercase letters for more than 9 lines

[Mike Hopkirk(hops) <ho...@sc...>]

On Wed May 03,2000 (07:21:03AM -0700), pe...@us...(Petr Sorfa) wrote:
> Hi,
> 
> Following Joshua's suggestion, the selection is display in uppercase if
> it is a letter.


well its displaying a 9 lines with a 1 char selection then the remainder 
with a 2 character one: - 0[A-Z].
( source update from about10.00 this a.m.)

This and the >9 line display work is probably unnecessary effort since
the osr5 version of cscope already does this:
    i.e 
 - Uses all the spare lines available for selection display (9-??) 
and displays a  single letter for the selection 
[0-9][a-z][A-Z] in that order.

I'd suggest holding off on any further effort in either of these areas 
till the osr5 differences are categorised and applied ...

-- 
- hops

Everything disclaimed (including disclaimer)
------<ho...@sc...>--------------------------------------
Mike Hopkirk (hops)  |       Whenever possible steal code.
SCO Inc              |          Tom Duff. (ex) Bell Labs

[Cscope-devel] Major code cleanup and memory leak fixes

[Petr S. <pe...@us...>]

Hi,

Thanks to Hans-Bernhard Broeker cscope has gone through a major code
cleanup and a memory leak fixes. I provided some minor error correction
in to code.

Petr

[Cscope-devel] autoconf now checks for vsnprintf

[Petr S. <pe...@us...>]

Hi,

Thanks to Hans-Bernhard Broeker autoconf and support files now check
for vsnprintf

Petr

[Cscope-devel] uppercase letters for more than 9 lines

[Petr S. <pe...@us...>]

Hi,

Following Joshua's suggestion, the selection is display in uppercase if
it is a letter.

Petr

Re: [Cscope-devel] Support for more than 9 select lines

[Joshua U. <uz...@li...>]

* Petr Sorfa <pe...@us...> [000502 15:15]:
> After a bit of slog, I've managed to allow more than 9 select lines, all the
> way up to 45. As its in test mode, its enabled through the option -t.
> 
> Select lines greater than 9 are selected with 0[0-9,a-z] (case insensitive).

Interesting.  One thing I would change is this:

=============================================================================
diff -u -r1.2 display.c
--- display.c   2000/05/02 21:41:48     1.2
+++ display.c   2000/05/02 22:27:12
@@ -254,7 +254,7 @@
                                        }
                                        else {
                                                printw("0%c",
-                                                       disprefs - 19 + 'a');
+                                                       disprefs - 19 + 'A');
                                        }
                                }
                        }
=============================================================================

I think it looks a little better with uppercase than lowercase.  I
also think there's something a little kludgy about the whole 0[0-9a-z]
method... but it's a decent solution (at least for now).  Some of my
friends were complaining for this feature... I'll have them try it
and see what they think.
 
> Seems to work ok. Please play around so we can get any glitches out. I
> made sure that all the display stuff is nicely aligned and that the
> "Change this text string" works nicely as well.

Yep... seems to work fine over here as well.  One thing we should
look at is handling the screen being resized.  That comes later. :)

-- 
Joshua Uziel, Senior Linux Consultant, Linuxcare, Inc.
415.354.4878 tel, 415.701.7457 fax
uz...@li..., http://www.linuxcare.com/
Linuxcare. Support for the revolution.

[Cscope-devel] Support for more than 9 select lines

[Petr S. <pe...@us...>]

Hi,

After a bit of slog, I've managed to allow more than 9 select lines, all the
way up to 45. As its in test mode, its enabled through the option -t.

Select lines greater than 9 are selected with 0[0-9,a-z] (case insensitive).

Seems to work ok. Please play around so we can get any glitches out. I
made sure that all the display stuff is nicely aligned and that the
"Change this text string" works nicely as well.

No documentation as yet. I'll wait till we are happy with it before I make it
default. The old 9 line select will be available through some option (yet
to be determined).

Petr

Re: [Cscope-devel] configure.in updated

[Joshua U. <uz...@li...>]

* Petr Sorfa <pe...@us...> [000502 11:05]:
> Updated configure.in to generate a workable version of configure for UW7.

Great. :)
 
> However, I have decided to keep configure in the source tree. Makes
> things easier for folks to type ./configure rather than autoconf and
> ./configure (I hope).

Fair enough... that door can swing either way.

-- 
Joshua Uziel, Senior Linux Consultant, Linuxcare, Inc.
415.354.4878 tel, 415.701.7457 fax
uz...@li..., http://www.linuxcare.com/
Linuxcare. Support for the revolution.

Re: [Cscope-devel] autoconf for UW7

[Mike Hopkirk(hops) <ho...@sc...>]

On Tue May 02,2000 (12:52:41AM -0700), uz...@li...(Joshua Uziel) wrote:
> 
> I might even suggest taking the "configure" file out of the
> tree, since it is a generated file...

Don't do this.

The normal entry point to bootstrapping osrc src dists is to find 
and run the configure script., then the make salutations

Besides which You need to leave it in for systems that don't have the entire 
autoconf/automake/m4/.. house of cards installed .....


-- 
- hops

Everything disclaimed (including disclaimer)
------<ho...@sc...>--------------------------------------
Mike Hopkirk (hops)  |       Whenever possible steal code.
SCO Inc              |          Tom Duff. (ex) Bell Labs

[Cscope-devel] configure.in updated

[Petr S. <pe...@us...>]

Hi,

Updated configure.in to generate a workable version of configure for UW7.

However, I have decided to keep configure in the source tree. Makes
things easier for folks to type ./configure rather than autoconf and
./configure (I hope).

Petr

Re: [Cscope-devel] autoconf for UW7

[Petr S. <pe...@sc...>]

Hi Joshua,

Thanks, will fix it later today.

Petr
> * Petr Sorfa <pe...@us...> [000501 10:29]:
> > I've updated ./configure to work properly for UW7 to pick up -lcurses.
> >
> > I'm not too sure if the fix is valid, so please test.
> >
> > Basically there was a line (1700) with #include</usr/include/curses.h> which
> > was causing a test failure. By changing it to #include<curses.h> the
> > test passed and all is rosey.
> >
> > I've tested it on UnixWare 7.0 and RedHat 6.2 with no problem.
> 
> Umm... no, the fix is not valid.  The change must be done in the
> configure.in (or related) file... not in the configure script
> itself.  The second you run "autoconf" to regenerate the configure
> script, that change becomes moot.  This is what makes it so
> challenging to deal with.
> 
> I might even suggest taking the "configure" file out of the
> tree, since it is a generated file...
> 
> --
> Joshua Uziel, Senior Linux Consultant, Linuxcare, Inc.
> 415.354.4878 tel, 415.701.7457 fax
> uz...@li..., http://www.linuxcare.com/
> Linuxcare. Support for the revolution.
> 
> _______________________________________________
> Cscope-devel mailing list
> Csc...@li...
> http://lists.sourceforge.net/mailman/listinfo/cscope-devel

-- 
--------------------------------------------------------
Petr Sorfa                  Software Engineer
Santa Cruz Operation (SCO) 
430 Mountain Ave.           http://www.sco.com
Murray Hill 07974
NJ, USA
--------------------------------------------------------
Disclaimer: All my comments are my own and nobody else's
----------------------------------------------------------