Arun Persaud | 1 Sep 07:01 2009

beta2

Hi

I tagged beta2 and am uploading to alpha.gnu.org at the moment. The plan would be to have the beta2 up for a few
days and release 4.4.0 on the weekend (I should have some more time on the weekend). Let me know if
everything in beta2 works for you.

One change I will do is to move the installer directory out of the master branch in git and into it's own branch
(one without a parent, similar to the fselch stuff). Including it made the tar.gz quite large and we don't
really need to have it in master.

Other things I will do include updating the AUTHORS file, NEWS, CHANGELOG etc. although most of them should
be up to date...

ARUN

Arun Persaud | 1 Sep 07:12 2009

Re: Polyglot 2.0,

Hi

my two cents:

the best solution would be, if we could avoid the fork and keep just one version around... I have to say that I
also don't like the idea of calling it 2.xx, if we fork we should change the name, not the version number,
polyglot-x or something like that sounds better... but the best would be if you two can figure out a way to
avoid forking :) 

(Un)fortunately, I have no idea about what you guys are talking about (never done anything with engines),
so I can't comment on the issues...

ARUN

h.g. muller | 1 Sep 12:06 2009
Picon

Re: Polyglot 2.0,

Well, it seems calling it 1.4.38x is  good solution. It is a small
patch, after all. Michel's branch has in the mean time advanced
to 1.4.42b, and before we release I will have a look at this new
version to see if I should patch it with an even smaller patch to
get a 1.4.42x and include that. Or even include both 1.4.38x and
1.4.42b to give the user an opportunity to decide what he likes
best.

Arun Persaud | 2 Sep 05:37 2009

Re: Polyglot 2.0,

Hi

> Well, it seems calling it 1.4.38x is  good solution. It is a small
> patch, after all. Michel's branch has in the mean time advanced
> to 1.4.42b, and before we release I will have a look at this new
> version to see if I should patch it with an even smaller patch to
> get a 1.4.42x and include that. Or even include both 1.4.38x and
> 1.4.42b to give the user an opportunity to decide what he likes
> best.

so no chance of getting our patch into polyglot or adapting things in xboard and polyglot?

ARUN

Arun Persaud | 3 Sep 08:17 2009

upcoming release 4.4.0

Hi

planing on releasing 4.4.0 this weekend (either Saturday or Sunday 2009-09-05), after two beta versions.
Let me know if you find any bugs that need to be fixed beforehand...  

ARUN

h.g. muller | 3 Sep 08:39 2009
Picon

Re: upcoming release 4.4.0

At 23:17 2-9-2009 -0700, Arun Persaud wrote:
>Hi
>
>planing on releasing 4.4.0 this weekend (either Saturday or Sunday 
>2009-09-05), after two beta versions. Let me know if you find any bugs 
>that need to be fixed beforehand...
>
>ARUN
>

OK, great!

Yes, these latest patches can still go into 4.4.0. They are quite safe;
just some corrections in menu texts (and doc files). The most involved
was the error message XBoard prints on receiving a non-understood
option feature, but I tested that. (There are currently no engines that
send option features anyway, except Polyglot, and the latter was
not supposed to send any erroneous option definitions, but I made
a special version that does, and even that works perfectly with the
new code.)

Actually I was a bit weary on making a change like the latter so close
before release, but OTOH it would be very silly to violate our own
protocol specs. Plus that the old version seemed to handle erroneous
option definitions in a suspect manner anyway, which I could not
guarantee as safe. (It accepted any option feature, which might
have lead to problems in later processing.) So in the worst case
I replaced one bug by another. But it tests OK.

If more bug reports come in, I will let you know immediately.
(Continue reading)

h.g. muller | 3 Sep 19:46 2009
Picon

Re: upcoming release 4.4.0


>
>planing on releasing 4.4.0 this weekend (either Saturday or Sunday 
>2009-09-05), after two beta versions. Let me know if you find any bugs 
>that need to be fixed beforehand...

Currently there are no known bugs. I have made a new Fairy-Max version,
which implements a number of engine-defined options. (For setting a resign
threshold, altering the file with game definitions, a button to clear the 
hash table,
and some dummy options so that every option type is used once.) Now that we
extended WinBoard protocol, we must at least supply one engine that actually
uses that new protocol! So this Fairy-Max can go with 4.4.0.

I am still considering the Polyglot situation. I will let you know before 
Saturday
when the final installer package is on my website.

Arun Persaud | 4 Sep 08:35 2009

getting closer

Hi

I moved the installer files to their own branch and removed them from the master branch (this way they won't
show up in the automatically created tar-ball). Also added a patch from HGM to fix mostly typos. Still
planing on releasing on the weekend.

ARUN

Arun Persaud | 4 Sep 08:49 2009

GTK version

Hi

merged all the changes from master into the GTK version. Also converted more X-stuff into GTK and I'm able to
play against a computer now on the GTK version. This should only work on linux at the moment, since there is
some code that needs an #ifdef for windows (handling open files/connections). The GTK version also still
crashes, for example when promoting a pawn, since the promotion-window is not been ported to GTK and the
old X-code will now create a crash...

However, people can start playing around with the GTK version.

NB I might have deleted some features (although I tried to comment them out rather than deleting them),
which we might have to add again... other features such as Mono-mode might not make sense in a GTK
environment anyway (I still need to read up on this).

If people want to work on the GTK version, here is what you should probably use:

glade-3 version>=2.6.0
gtk  2.16

Drawing should be done via cairo AFAIK, see for example

xboard.c:drawHighlight()
xboard.c:drawSquare()

all the GTK stuff is in:

xboard.c
callback.[ch]
gtk-interface.xml (the gtk builder file, created with glade)
interface.[ch] (perhaps move all the gtk stuff out of xboard into here)
(Continue reading)

h.g. muller | 4 Sep 10:20 2009
Picon

GTK version, promotion popup

At 23:49 3-9-2009 -0700, Arun Persaud wrote:
>The GTK version also still crashes, for example when promoting a pawn, 
>since the promotion-window is not been ported to GTK and the old X-code 
>will now create a crash...

Getting rid of the promotion popup altogether is on my to-do list,
so perhaps you should not bother porting the old dialog. One of
the very annoying things about the XBoard promotion popup is
that it is shown even for moves that are already known to be illegal.
Try for instance to play e2d8 from the opening setup. When legality
testing is on, it first shows te promotion popup to make you select
a piece, and no matter what you choose, the move is then rejected
as illegal later. In WinBoard I fixed this, but not yet in XBoard.
Especially in Shogi, where moving any piece type in or out of a rather
deep promotion zone could cause a promotion, this behavior was very
annoying. (For normal Chess you are of course not very likely to draw
Pawns to the last rank by accident.)

But what is even more annoying is that the promotion popup is not
variant-proof. Not all of the 22 piece types can appear in the promotion
menu, although it is patched a little to work in the most common variants
(e.g. in giveaway you can promote to King, and in Gothic to Archbishop
and Chancelor, while in Shogi you can select Yes/No.)

Another problem is that currently internationalization is broken in the
promotion popup: It uses the first letter of the name on the button as
a piece indicator (except for Knight), so if yo change the button labels,
you would get invalid choices unless you would also adapt the
pieceToChar table to the language. (Which is in general undesirable,
as it would preclude you from reading international PGN.)
(Continue reading)


Gmane