tkdiff-discuss Mailing List for tkdiff (Page 4)

Hi all,

to make tkdiff even more useful, it should be possible to pass line
numbers (current line or current diff) to the external editor. In
the converse direction, when invoking tkdiff from within an editor,
tkdiff should have a command line option to move the view to a given
line of the text.

Oh, and please specify the tcl/tk version that is required to run
tkdiff.

Best wishes
J=F6rg

Dear Open Source developer

I am doing a research project on "Fun and Software Development" in which I kindly invite you to participate.
You will find the online survey under http://fasd.ethz.ch/qsf/. The questionnaire consists of 53 questions and you will need about 15 minutes to complete it.

With the FASD project (Fun and Software Development) we want to define the motivational significance of fun when software developers decide to engage in Open Source projects. What is special about our research project is that a similar survey is planned with software developers in commercial firms. This procedure allows the immediate comparison between the involved individuals and the conditions of production of these two development models. Thus we hope to obtain substantial new insights to the phenomenon of Open Source Development.


With many thanks for your participation,
Benno Luthiger


PS:
The results of the survey will be published under http://www.isu.unizh.ch/fuehrung/blprojects/FASD/.
We have set up the mailing list fa...@we... for this study. Please see http://fasd.ethz.ch/qsf/mailinglist_en.html for registration to this mailing list.

_______________________________________________________________________

Benno Luthiger
Swiss Federal Institute of Technology Zurich
8092 Zurich

Mail: benno.luthiger(at)id.ethz.ch
_______________________________________________________________________

E-mail is the fastest growing marketing tool.

We offer E-mail Marketing with quality service and the lowest prices.

1. Targeted E-mail Addresses

We can provide targeted e-mail addresses you need, which are compiled 
only on your order. We will customize your customer e-mail addresses.

* We have millions of e-mail addresses in a wide variety of categories.

2. Send out Targeted E-mails for you

We can send your e-mail message to your targeted customers! We will
customize your email addresses and send your message for you.

* We can Bullet Proof your Web Site $ dedicated server.

We also offer Fax Broadcasting Service. For more details, you can
refer to:  Http://www.9206.com

Looking forward to serving you.


Regards!

John Okoh
www.9206.com
Su...@92...

*************************************************************************
To take your address:  Http://213.172.0x1f.16/index.html
*************************************************************************

Hi all,

TkDiff 4.0b2 has been released at the SourceForge site 
<http://sourceforge.net/projects/tkdiff/>.  It has the Windows path 
problems fixed - all of them I could find, anyway.  The browse dialog 
and the tempfile issues are resolved to the best of my somewhat limited 
ability to test. (silly XP machine, very unix-y usage model on the part 
of the owner, which is why I miss things.)

This is _not_ the same as the starkit version Larry mentioned.  I 
downloaded it to see, and what I got seems to be 3.09 with a couple of 
one-line  changes.

Some time ago, someone asked about the 3-way merge functionality, 
thinking it had disappeared.  I examined 3.09 and 4.0x, and the 3-way 
merge code is still there and appears to do the same thing.  It doesn't 
display the parent file, but only takes it into account when deciding 
which of the two files being merged is chosen by default.
If anyone knows of a version that does more, or can tell me if I'm 
missing something, let me know.  There are 10,000 lines of code in one 
file, which can make a person kinda crazy ;-)

Dorothy


Larry W. Virden wrote:
> Just a note - if people are aware of what a Starkit is, the Starkit Archive
> at http://mini.net/sdarchive has a tkdiff.kit that might be worth taking
> a look at.  One reason is that it calls itself 4.0b2 - and there appears
> to be a bug fix in it that isn't in the latest sf.net download...

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Dorothy Robinson
http://www.twobarleycorns.net

I believe 3-way merging maybe got merged out of the current tkdiff =
version.
Look at the earlier versions. I use something based on 3.04, and it has =
3-way
merging.
=20
I've been hoping we can merge that feature back into the current, after =
the next
tkdiff release.
-- Dave Brown

-----Original Message-----
From: e.j...@ph... [mailto:e.j...@ph...]
Sent: Tuesday, June 24, 2003 4:35 AM
To: tkd...@li...
Subject: [Tkdiff-discuss] 3 way merge



I=B4m trying to set up TkDiff as a 3way merge tool.=20
It only passes 2 arguments to the diff tool.=20

 tkdiff myfile yourfile -a oldfile=20

the argument yourfile and oldfile are passed to the diff tool.=20
I=B4m running TkDiff under Cygwin.=20

Eddy
______________________________
E.J. Pronk
Philips Centre for Industrial Technology
Engineering & Test / BM Equipment Design
P.O.Box 218 / SAN-2181
5600 MD Eindhoven

Tel: +31 40 27 32693 /  Fax: +31 40 27 38950
Email: E.J...@Ph...

I=B4m trying to set up TkDiff as a 3way merge tool.
It only passes 2 arguments to the diff tool.

 tkdiff myfile yourfile -a oldfile

the argument yourfile and oldfile are passed to the diff tool.
I=B4m running TkDiff under Cygwin.

Eddy
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=
=5F=5F=5F=5F=5F
E.J. Pronk
Philips Centre for Industrial Technology
Engineering & Test / BM Equipment Design
P.O.Box 218 / SAN-2181
5600 MD Eindhoven

Tel: +31 40 27 32693 /  Fax: +31 40 27 38950
Email: E.J...@Ph...

I noticed a few bugs in the inline diff for cases where the region
returned from diff contains an unequal number of lines.  Here's the
patch against the current 4.0b1 ...

Mike.
 
--- Dorothy Robinson <mo...@ar...> wrote:
> This is great.  I've got it installed at my job, so it will get a
> little use at 
> least.
> 
> Dorothy
> 
> John Klassa wrote:
> > Michael,
> > 
> > Got your patched version into the system as revision 1.11...  I
> went ahead
> > and changed the version string to 4.0b1 (the switch to 4.0 is to
> get away
> > from the strange [IMHO] numbering we've been using :-)).  But
> anyway, I love
> > your changes!  Having a visual indication of the difference,
> inline, is very
> > cool.  Nicely done!
> > 
> > Anybody who wants to take a look can grab the latest via anonymous
> CVS...
> > 
> > I propose that we play around with this version for a while, then
> ship it as
> > 4.0 at some point in the near future.
> > 
> > Thoughts, anyone?
> > 
> > John

__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com

Flashes and thrashes??  I can't get that to happen on my PC, either with Cygwin 
or with a plain old DOS window and the Windows version of Tk.  Both Tk's are 
8.3, though.  I had 8.4 on my PC for awhile and it did some very strange things, 
though flashing and thrashing weren't usually among them.  Or are you running Tk 
under X11 in Cygwin?  I haven't tried that yet.  (Only have a PC to do things 
that _can't_ be done on a reasonable computer ...)

dorothy

John Klassa wrote:
> I applied your patch...  Everything's cool (at least under cygwin), except
> that when you run tkdiff without a filename, it flashes and thrashes and
> generally goes nuts.  If you supply a filename, all is well.  Any ideas?

-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Dorothy Robinson
http://www.twobarleycorns.net

I applied your patch...  Everything's cool (at least under cygwin), =
except
that when you run tkdiff without a filename, it flashes and thrashes and
generally goes nuts.  If you supply a filename, all is well.  Any ideas?

Your version is 1.12 in CVS.

As for your other changes (and David's proposed changes), let's wait =
until
the dust settles on b2 and get 4.0 out the door...  After that, we can
figure out how best to add the feature you're after, in light of David's
work.  Sound okay with everybody?

Thanks, Dorothy!

John

--
John Klassa / jo...@kl...

Please use this address for all personal correspondence.  Thanks!

> -----Original Message-----
> From: John Klassa [mailto:kl...@qw...] On Behalf Of Dorothy =
Robinson
> Sent: Wednesday, May 14, 2003 2:33 AM
> To: tkd...@so...
> Subject: Re: [Tkdiff-discuss] tkdiff 4.0b1 available
>=20
> I found a mistake which caused filenames loaded from the Browse dialog =
-
> ie., on
> a PC - to not withstand spaces in the filenames.  I fixed it, along =
with
> the
> setcolor problem, and submitted a patch on Sourceforge.  (I'm sure the
> upload
> "took" this time.)
>=20
> On another subject - I want to allow tkdiff to do a cvs rdiff like
> "cvs -d repository rdiff -r1.1 -r1.2 modulepath/filename" on a file =
that
> isn't
> checked out.  Right now it won't do that, because it doesn't know from
> rdiff and
> because it checks to see if the file given in the argument exists, =
which
> it
> won't in this case.  So I thought of adding a command line option for
> doing an
> rdiff, and removing the check for the existence of the file.  =
Something
> like
> "tkdiff -d repository -rdiff -r1.1 -r1.2 modulepath/filename"
> Objections?  Better ideas?  I can't help thinking there might be an =
easier
> way.
>=20
> Dorothy
>=20
> Daniel H=F6pfl wrote:
> > Hi!
> >
> > In line 153-158 you have:
> >
> > ------------------------------------------------------
> > set w(selcolor) [lindex [.testent configure -selectbackground] 3]
> > set entryfont [lindex [.testent configure -font] 3]
> > destroy .testent
> > # the above results in a nearly undistinguishable darker gray for =
the
> > # selected color (rh8 with tk 8.3.3-74) "#c3c3c3"
> > set w(selcolor) "#b03060"
> > ------------------------------------------------------
> >
> > I think the problem is the index 3 when setting w(selcolor):
> >
> > ------------------------------------------------------
> > set w(selcolor) [lindex [.testent configure -selectbackground] 4]
> > set entryfont [lindex [.testent configure -font] 3]
> > destroy .testent
> > ------------------------------------------------------
> >
> > ... sets the value of "*selectBackground" from your .Xdefaults (or
> > #bfdfff, a light blue, as default) into w(selcolor).
> >
> > An other solution would be:
> >
> > ------------------------------------------------------
> > set w(selcolor) [lindex [.testent configure -selectforeground] 4]
> > set entryfont [lindex [.testent configure -font] 3]
> > destroy .testent
> > ------------------------------------------------------
> >
> > In this case you get the value of "*selectForeground". Better than =
using
> > the background color for the foreground of a checkbox.
> >
> > Is there a way to get the value of "*selector" from the .Xdefaults? =
I
> > think this would be the "correct" value for this case.
> >
> > Greetings,
> >    Daniel
> >
>=20
> --
> =
-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=
=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-
> Dorothy Robinson
> http://www.twobarleycorns.net
>=20
>=20
>=20
> -------------------------------------------------------
> Enterprise Linux Forum Conference & Expo, June 4-6, 2003, Santa Clara
> The only event dedicated to issues related to Linux enterprise =
solutions
> www.enterpriselinuxforum.com
>=20
> _______________________________________________
> Tkdiff-discuss mailing list
> Tkd...@li...
> https://lists.sourceforge.net/lists/listinfo/tkdiff-discuss

I use tkdiff quite a bit, with PVCS. I originally submitted the PVCS =
support for
tkdiff.
Anyway, I've been waiting for the dust to settle, ready to submit some =
stuff:
probably good for inclusion after the next release. I mention this now =
because
Dorothy's idea is similar yet different.

I have added -s [CVS | RCS | Perforce | AccuRev | PVCS | SCCS] for the =
cases
where tkdiff can't guess what source control system is being used. This =
might be
necessary or helpful when specifying the repository. =20

I also added -c <vcs.cfg>. For PVCS, the key to finding the repository =
is the
file "vcs.cfg". If I'm running tkdiff outside of the diretory where =
vcs.cfg
resides, I need to specify -s PVCS and also -c c:/path-to/vcs.cfg.

Perhaps the better approach would be:
-s CVS -d <path-to-repository>
-s PVCS -d <path-to-vcs.cfg>
etc.

My other local modification is the addition of PVCS support for =
promotion groups
and version labels, using the -g and -v prefixes.=20

My main gripe is that the -a ancestor option has to be an actual file. =
I've
tried to fix this, but the limited command-line syntax has gotten in my =
way. If
anyone has a suggestion please let me know. I use -a quite often.
--
Dave Brown <dbr...@ha...>

> -----Original Message-----
> From: Dorothy Robinson [mailto:mo...@ar...]
> Sent: Wednesday, May 14, 2003 2:33 AM
> To: tkd...@so...
> Subject: Re: [Tkdiff-discuss] tkdiff 4.0b1 available
>=20
>=20
> I found a mistake which caused filenames loaded from the=20
> Browse dialog - ie., on=20
> a PC - to not withstand spaces in the filenames.  I fixed it,=20
> along with the=20
> setcolor problem, and submitted a patch on Sourceforge.  (I'm=20
> sure the upload=20
> "took" this time.)
>=20
> On another subject - I want to allow tkdiff to do a cvs rdiff like
> "cvs -d repository rdiff -r1.1 -r1.2 modulepath/filename" on=20
> a file that isn't=20
> checked out.  Right now it won't do that, because it doesn't=20
> know from rdiff and=20
> because it checks to see if the file given in the argument=20
> exists, which it=20
> won't in this case.  So I thought of adding a command line=20
> option for doing an=20
> rdiff, and removing the check for the existence of the file. =20
> Something like
> "tkdiff -d repository -rdiff -r1.1 -r1.2 modulepath/filename"
> Objections?  Better ideas?  I can't help thinking there might=20
> be an easier way.
>=20
> Dorothy
>=20
> Daniel H=F6pfl wrote:
> > Hi!
> >=20
> > In line 153-158 you have:
> >=20
> > ------------------------------------------------------
> > set w(selcolor) [lindex [.testent configure -selectbackground] 3]
> > set entryfont [lindex [.testent configure -font] 3]
> > destroy .testent
> > # the above results in a nearly undistinguishable darker=20
> gray for the
> > # selected color (rh8 with tk 8.3.3-74) "#c3c3c3"
> > set w(selcolor) "#b03060"
> > ------------------------------------------------------
> >=20
> > I think the problem is the index 3 when setting w(selcolor):
> >=20
> > ------------------------------------------------------
> > set w(selcolor) [lindex [.testent configure -selectbackground] 4]
> > set entryfont [lindex [.testent configure -font] 3]
> > destroy .testent
> > ------------------------------------------------------
> >=20
> > ... sets the value of "*selectBackground" from your .Xdefaults (or=20
> > #bfdfff, a light blue, as default) into w(selcolor).
> >=20
> > An other solution would be:
> >=20
> > ------------------------------------------------------
> > set w(selcolor) [lindex [.testent configure -selectforeground] 4]
> > set entryfont [lindex [.testent configure -font] 3]
> > destroy .testent
> > ------------------------------------------------------
> >=20
> > In this case you get the value of "*selectForeground".=20
> Better than using=20
> > the background color for the foreground of a checkbox.
> >=20
> > Is there a way to get the value of "*selector" from the=20
> .Xdefaults? I=20
> > think this would be the "correct" value for this case.
> >=20
> > Greetings,
> >    Daniel
> >=20
>=20
> --=20
> =
-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D=
-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-
> Dorothy Robinson
> http://www.twobarleycorns.net
>=20

I found a mistake which caused filenames loaded from the Browse dialog - =
ie., on=20
a PC - to not withstand spaces in the filenames.  I fixed it, along with =
the=20
setcolor problem, and submitted a patch on Sourceforge.  (I'm sure the up=
load=20
"took" this time.)

On another subject - I want to allow tkdiff to do a cvs rdiff like
"cvs -d repository rdiff -r1.1 -r1.2 modulepath/filename" on a file that =
isn't=20
checked out.  Right now it won't do that, because it doesn't know from rd=
iff and=20
because it checks to see if the file given in the argument exists, which =
it=20
won't in this case.  So I thought of adding a command line option for doi=
ng an=20
rdiff, and removing the check for the existence of the file.  Something l=
ike
"tkdiff -d repository -rdiff -r1.1 -r1.2 modulepath/filename"
Objections?  Better ideas?  I can't help thinking there might be an easie=
r way.

Dorothy

Daniel H=F6pfl wrote:
> Hi!
>=20
> In line 153-158 you have:
>=20
> ------------------------------------------------------
> set w(selcolor) [lindex [.testent configure -selectbackground] 3]
> set entryfont [lindex [.testent configure -font] 3]
> destroy .testent
> # the above results in a nearly undistinguishable darker gray for the
> # selected color (rh8 with tk 8.3.3-74) "#c3c3c3"
> set w(selcolor) "#b03060"
> ------------------------------------------------------
>=20
> I think the problem is the index 3 when setting w(selcolor):
>=20
> ------------------------------------------------------
> set w(selcolor) [lindex [.testent configure -selectbackground] 4]
> set entryfont [lindex [.testent configure -font] 3]
> destroy .testent
> ------------------------------------------------------
>=20
> ... sets the value of "*selectBackground" from your .Xdefaults (or=20
> #bfdfff, a light blue, as default) into w(selcolor).
>=20
> An other solution would be:
>=20
> ------------------------------------------------------
> set w(selcolor) [lindex [.testent configure -selectforeground] 4]
> set entryfont [lindex [.testent configure -font] 3]
> destroy .testent
> ------------------------------------------------------
>=20
> In this case you get the value of "*selectForeground". Better than usin=
g=20
> the background color for the foreground of a checkbox.
>=20
> Is there a way to get the value of "*selector" from the .Xdefaults? I=20
> think this would be the "correct" value for this case.
>=20
> Greetings,
>    Daniel
>=20

--=20
-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=
=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-
Dorothy Robinson
http://www.twobarleycorns.net

Hmm.  On all the wm's I have at hand, indices 3 and 4 are the same.  Howe=
ver, I=20
see I was grabbing the color from the wrong place anyway, and I should ha=
ve got=20
it from the test menu:

*** 138,143 ****
--- 138,144 ----
   . configure -menu .native
   menu .native
   set menufont [lindex [.native configure -font] 3]
+ set w(selcolor) [lindex [.native configure -selectcolor] 3]
   destroy .native

   # Find out what the tk default is
***************
*** 150,161 ****
   destroy .testtext

   entry .testent
- set w(selcolor) [lindex [.testent configure -selectbackground] 3]
   set entryfont [lindex [.testent configure -font] 3]
   destroy .testent
- # the above results in a nearly undistinguishable darker gray for the
- # selected color (rh8 with tk 8.3.3-74) "#c3c3c3"
- set w(selcolor) "#b03060"

   # Make sure they're specified in points not pixels.  AquaTk doesn't.
   if {! [regexp -- {-} $labelfont]} {
--- 151,158 ----

Dorothy

Daniel H=F6pfl wrote:
> Hi!
>=20
> In line 153-158 you have:
>=20
> ------------------------------------------------------
> set w(selcolor) [lindex [.testent configure -selectbackground] 3]
> set entryfont [lindex [.testent configure -font] 3]
> destroy .testent
> # the above results in a nearly undistinguishable darker gray for the
> # selected color (rh8 with tk 8.3.3-74) "#c3c3c3"
> set w(selcolor) "#b03060"
> ------------------------------------------------------
>=20
> I think the problem is the index 3 when setting w(selcolor):
>=20
> ------------------------------------------------------
> set w(selcolor) [lindex [.testent configure -selectbackground] 4]
> set entryfont [lindex [.testent configure -font] 3]
> destroy .testent
> ------------------------------------------------------
>=20
> ... sets the value of "*selectBackground" from your .Xdefaults (or=20
> #bfdfff, a light blue, as default) into w(selcolor).
>=20
> An other solution would be:
>=20
> ------------------------------------------------------
> set w(selcolor) [lindex [.testent configure -selectforeground] 4]
> set entryfont [lindex [.testent configure -font] 3]
> destroy .testent
> ------------------------------------------------------
>=20
> In this case you get the value of "*selectForeground". Better than usin=
g=20
> the background color for the foreground of a checkbox.
>=20
> Is there a way to get the value of "*selector" from the .Xdefaults? I=20
> think this would be the "correct" value for this case.
>=20
> Greetings,
>    Daniel
>=20

--=20
-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=
=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-=3D-
Dorothy Robinson
http://www.twobarleycorns.net

Hi!

In line 153-158 you have:

------------------------------------------------------
set w(selcolor) [lindex [.testent configure -selectbackground] 3]
set entryfont [lindex [.testent configure -font] 3]
destroy .testent
# the above results in a nearly undistinguishable darker gray for the
# selected color (rh8 with tk 8.3.3-74) "#c3c3c3"
set w(selcolor) "#b03060"
------------------------------------------------------

I think the problem is the index 3 when setting w(selcolor):

------------------------------------------------------
set w(selcolor) [lindex [.testent configure -selectbackground] 4]
set entryfont [lindex [.testent configure -font] 3]
destroy .testent
------------------------------------------------------

=2E.. sets the value of "*selectBackground" from your .Xdefaults (or #bfd=
fff,=20
a light blue, as default) into w(selcolor).

An other solution would be:

------------------------------------------------------
set w(selcolor) [lindex [.testent configure -selectforeground] 4]
set entryfont [lindex [.testent configure -font] 3]
destroy .testent
------------------------------------------------------

In this case you get the value of "*selectForeground". Better than using =
the=20
background color for the foreground of a checkbox.

Is there a way to get the value of "*selector" from the .Xdefaults? I thi=
nk=20
this would be the "correct" value for this case.

Greetings,
    Daniel

--=20
Daniel H=F6pfl                                            targit GmbH
                                                R=FCdesheimer Stra=DFe 7
Tel: +49 (89) 517 397 - 0                           D-80686 M=FCnchen
Fax: +49 (89) 517 397 - 77                    http://www.targit.de/

The first public beta of the upcoming release 4.0 is available on
tkdiff.sf.net...  Go to the "files" section and grab a copy.  The current
release version (3.09, or 3.0.9, depending on how you look at it) is there
as well.

Thanks,
John

--
John Klassa / jo...@kl...

Please use this address for all personal correspondence.  Thanks!

This is great.  I've got it installed at my job, so it will get a little use at 
least.

Dorothy

John Klassa wrote:
> Michael,
> 
> Got your patched version into the system as revision 1.11...  I went ahead
> and changed the version string to 4.0b1 (the switch to 4.0 is to get away
> from the strange [IMHO] numbering we've been using :-)).  But anyway, I love
> your changes!  Having a visual indication of the difference, inline, is very
> cool.  Nicely done!
> 
> Anybody who wants to take a look can grab the latest via anonymous CVS...
> 
> I propose that we play around with this version for a while, then ship it as
> 4.0 at some point in the near future.
> 
> Thoughts, anyone?
> 
> John
> 

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Dorothy Robinson
http://www.twobarleycorns.net

Michael,

Got your patched version into the system as revision 1.11...  I went =
ahead
and changed the version string to 4.0b1 (the switch to 4.0 is to get =
away
from the strange [IMHO] numbering we've been using :-)).  But anyway, I =
love
your changes!  Having a visual indication of the difference, inline, is =
very
cool.  Nicely done!

Anybody who wants to take a look can grab the latest via anonymous =
CVS...

I propose that we play around with this version for a while, then ship =
it as
4.0 at some point in the near future.

Thoughts, anyone?

John

--
John Klassa / jo...@kl...

Please use this address for all personal correspondence.  Thanks!


> -----Original Message-----
> From: Michael D. Beynon [mailto:be...@ya...]
> Sent: Monday, May 12, 2003 8:33 AM
> To: John Klassa
> Subject: Re: redo your patch?
>=20
> I reapplied my inline diff changes to cvs 1.10.  I am not sure how you
> want me to send the revised version, so I am simply attaching it to
> this message.  I don't think I can directly update the version in cvs.
>=20
> I did find a problem on my machine (rh8.0 with tcl/tk 8.3.3-74) where
> there is a new test to determine what color a selection is (selcolor)
> by creating a dummy entry widget, and selecting the text in it, and
> using configure to determine the color.  The problem is I get a very
> dark gray, which is nearly indistinguishable from the unselected case.
> This means all the checkboxes in the menus look like they are
> unselected.  I simply hardcoded this to the default color I get for a
> checkbox.
>=20
> Mike.
>=20
> --- John Klassa <jo...@kl...> wrote:
> > Hi Michael,
> >
> > Not sure if the sf patch mechanism feeds back to you when you submit
> > a
> > patch, and there's a subsequent comment, but I thought I'd contact
> > you
> > directly just in case not.
> >
> > I applied your patch on top of the previous one in the queue
> > (Dorothy's),
> > and got failures on 5 of 30 hunks.  Rather than have me hose
> > something,
> > could I ask you to re-do your changes, on top of the latest that's =
in
> > CVS
> > (1.10, I believe) -- which is basically just the old 3.0.9 with
> > Dorothy's
> > patch thrown in?
> >
> > Thanks,
> > John
> >
> > --
> > John Klassa / jo...@kl...
> >
> > Please use this address for all personal correspondence.  Thanks!
> >
> >
>=20
> __________________________________
> Do you Yahoo!?
> The New Yahoo! Search - Faster. Easier. Bingo.
> http://search.yahoo.com

Hi Michael,

Not sure if the sf patch mechanism feeds back to you when you submit a
patch, and there's a subsequent comment, but I thought I'd contact you
directly just in case not.

I applied your patch on top of the previous one in the queue (Dorothy's),
and got failures on 5 of 30 hunks.  Rather than have me hose something,
could I ask you to re-do your changes, on top of the latest that's in CVS
(1.10, I believe) -- which is basically just the old 3.0.9 with Dorothy's
patch thrown in?

Thanks,
John

--
John Klassa / jo...@kl...

Please use this address for all personal correspondence.  Thanks!

I made some patches to support inline diff highlighting.
I moved the patches I made to the 3.09 version from
accurev.com to the latest version in cvs (3.09 + the
macosx aqua changes?). Attached is my patched version.

I implemented two algorithms for performing inline diffs:

(1) byte-by-byte - good for large text files with diffs
corresponding to column numbers. I use this for
detailed comparison of images represented as text files.

(2) recursive pattern matching based on
Ratcliff/Obershelp. This finds the longest common
substring, assumes this is an anchor, and recursively
repeats for the remaining left and right substrings. I
revised it a couple of times to make it more efficient,
and it is not too bad.

This required adding to, and changing, the color/font
scheme in the text windows. I also added some gui
changes to support the enabling of either inline diff
algorithm.

I posted the patch to the sourceforge tkdiff patches area ...
http://sourceforge.net/tracker/index.php?func=detail&aid=723135&group_id=64960&atid=509274

Feedback?

Mike. 

__________________________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
http://search.yahoo.com

Hi Tony,

There used to be a turned-into-an-exe version out there somewhere, but =
I'm
not sure where.

Basically, yes, you should install Tcl/Tk.  The easiest way (IMHO) is to =
go
to ActiveState.com and download theirs.  Others may have other =
suggestions.
You'll also need a "diff" executable, as tkdiff is really just a wrapper
around the command-line version of "diff".  For that, I'd just install
cygwin (personally)...  I live in cygwin.  There are other, standalone
implementations of diff for Windows out there, though, and they should =
also
work.

Hope this helps!

John

--
John Klassa / jo...@kl...

Please use this address for all personal correspondence.  Thanks!



> -----Original Message-----
> From: Tony Hughes [mailto:ant...@ve...]
> Sent: Wednesday, March 05, 2003 4:05 PM
> To: jo...@kl...
> Cc: ant...@ve...
> Subject: tkdiff
>=20
> John,
>=20
>      I found your name thru google looking for tkdiff.
> I am runing Windows XP. I tried downloading tkdiff
> but all I got was a zip file that unzipped to a tcl script.
>=20
>      How do I get tkdiff for Windows XP ? Is there a set
> of instructions somewhere ? Am I first supposed to
> install tcl/tk on my Windows XP ? If so, how do I do that ?
>=20
> Sincerely,
> Tony Hughes
>=20
>=20
> =
-------------------------------------------------------------------------=
-
> ----
> Tony Hughes
> 296 Payson Road
> Belmont, MA 02478
> 617-489-3560
> http://members.bellatlantic.net/~vze3n779/resume.html

[ Copying Daniel, the author of the patch, and also the new =
tkdiff-discuss
list at sourceforge. ]

Hi Dorothy,

Good to hear from you!

> Glad to hear you're taking on tkdiff again.  Here's a patch someone
> sent me for the version I included with tkcvs 7.1.  The first part
> fixes a boneheaded mistake I made,  and the rest of the patch is
> something that sounds like a good idea, although I haven't really
> studied it yet.  I'm in the middle of moving all my stuff onto a
> new computer.

Hmmmmmmm.  Not sure what to do...  Reasons:

1.  Your version is a heavily-modified 3.09.

2.  There's already a 3.10-in-progress in CVS (that is, a
differently-modified 3.09).

Any chance you'd be willing to work with us to get the relevant portions =
of
your version into ours, where they're appropriate? :-)  What exactly did =
you
do to create yours?

Anyway, for ignoring whitespace, the generally-accepted approach is to =
use
"diff" options when you invoke tkdiff.  Anything tkdiff doesn't =
recognize is
passed on to diff, and so one usually accomplishes this particular thing =
by
doing something like:

	tkdiff -t -w firstfile secondfile

The -t and -w are passed to diff, and tell it to ignore whitespace.

We made a conscious decision, way back, to forgo adding new built-in =
options
when there already exist diff options (that can be passed through) to
accomplish the same thing.  Note further that under Preferences > =
General,
you can actually set the "-t -w" persistently, so that access to a =
command
line isn't required.

Good to hear from you...  Hope all is well!

John