[Announce] Angband 3.0.8 prerelease 2
Andrew Sidwell
least 28 fewer (see <http://tinyurl.com/33a4j3> for the full list).
With any luck my newsreader shouldn't decide randomly to ignore this
thread like it has the PR1 one.
Downloads:
Source code: http://rephial.org/downloads/angband-3.0.8pr2.tar.gz
Windows: http://rephial.org/downloads/angband-3.0.8pr2-win.zip
RISC OS: http://rephial.org/downloads/angband-3.0.8pr2-ros.zip
Again, no OS X build right off.
If there are any reported bugs not yet fixed in this version (I'm pretty
sure they all are), please report them again on this thread, giving
steps to reproduce. You can send savefiles to bu...@rephial.org if
they're useful to see a bug in action. Any new bug reports or
suggestions for things people think should be in this release (except
ego squelch) are also welcome.
Noticeable changes in this version, compared to PR1:
- Different squelching support.
This can operate in one of two ways. If you prefer your squelchables
hidden, then turn on hide_squelchable, and whenever you find an item
which matches your squelch settings, it will be dropped but not
destroyed. This means that you can unsquelch it by changing your
settings later if you so desire.
If you prefer squelch to just help automate item destruction, then
you can press '!' at the "destroy" prompt to destroy all squelchables
on the floor or in your inventory.
You can set individual squelch settings for only those items which do
not get pseudo-ID'd. For those that do, you can choose to mark things
as "squelch" depending on the feeling you get when you first ID them
in that way. These categories are much broader than those found in
other variants; complain if necessary.
Worthless items and emptied chests are automatically squelched,
saving some micromanagement. All item destruction now takes 0
energy.
When objects are compressed, gold and squelched items are the first
to be destroyed.
- Fewer monsters have the amnesia attack, and CCW cures the effects of
it.
- Defined potions of life, healing, *healing*, augmentation as 'good'
drops, as well as restore mana for spell casters, and stat potions,
when the stat is not maximised.
- Credited Shanoah Alkire for work on the new GTK2 port.
- Over twenty-five other bugs fixed.
Enjoy,
--
Andrew Sidwell
http://rephial.org/ -- the home of Angband
My email address changes monthly, and is the first three letters of the
month (in English), followed by the last two digits of the current year,
@entai.co.uk.
Matthew Vernon
> Again, no OS X build right off.
I have an OSX box; would building be a) easy for me b) useful?
Matthew
--
Rapun.sel - outermost outpost of the Pick Empire
http://www.pick.ucam.org
Andrew Sidwell
please avoid reporting bugs already noted there.
Also, there is an Intel Mac compile available here:
http://www.ualberta.ca/~aphillip/angband-3.0.8pr2-osx.zip
Andrew Sidwell
> Andrew Sidwell <ju...@entai.co.uk> writes:
>
>> Again, no OS X build right off.
>
> I have an OSX box; would building be a) easy for me b) useful?
I doubt it, unless you already have dev tools set up and such. If you
or anyone else wants to make a PPC or PPC-and-Intel build, though,
that'd be much appreciated.
Matthew Vernon
> Matthew Vernon wrote:
> > Andrew Sidwell <ju...@entai.co.uk> writes:
> >
> >> Again, no OS X build right off.
> >
> > I have an OSX box; would building be a) easy for me b) useful?
>
> I doubt it, unless you already have dev tools set up and such. If you
> or anyone else wants to make a PPC or PPC-and-Intel build, though,
> that'd be much appreciated.
I have devtools and suchlike (though I mostly use make/gcc from the
command-line).
Eddie Grove
> The previous prerelease had lots of bugs, it seems. This one has at
> least 28 fewer (see <http://tinyurl.com/33a4j3> for the full list).
> With any luck my newsreader shouldn't decide randomly to ignore this
> thread like it has the PR1 one.
I played a bit, and squelch with ignore was working fine until it
stopped working.
In fact, the initial rings of weakness squelched without me giving any
command,
presumably because they are worthless, but later on I picked them up
and dropped
them and they remained visible. Perhaps it has to do with whether
they are identified?
I went into the menu, squelched by hand, no difference. Pseudo in the
pack worked
fine, at least at the beginning.
I could not sell unknown mushroom of poison to general store, but
could sell it to black market.
I got the "you found something" in the rubble, but since it was
squelched
[I am guessing], there was nothing there.
I could not figure out how to squelch magic books or iron spikes.
I could not figure out how to do auto-inscription. OTOH, I don't know
if it
is supposed to be in 3.0.8. I miss it. P.S. I want them to stick
from game to game.
I really really miss the ack when you destroy things, combined with s
or E for squelching.
I'm sure I will make a typo eventually and 'k'ill something I need.
If you squelch properly, you don't 'k'ill so many items, so it is
useful to have a confirm IMO.
[If I was the overlord, I'd change 'k' so that it is ONLY an interface
to squelch.
I know I am deranged about this.]
I get a bunch of asserts because I did not install in /usr/local.
Presumably ./configure should
check where you are and set the local/global install/CC flags
accordingly.
Eddie
Andrew Sidwell
> On Jul 4, 6:56 pm, Andrew Sidwell <j...@entai.co.uk> wrote:
>> The previous prerelease had lots of bugs, it seems. This one has at
>> least 28 fewer (see <http://tinyurl.com/33a4j3> for the full list).
>> With any luck my newsreader shouldn't decide randomly to ignore this
>> thread like it has the PR1 one.
>
> I played a bit, and squelch with ignore was working fine until it
> stopped working.
> In fact, the initial rings of weakness squelched without me giving any
> command,
> presumably because they are worthless, but later on I picked them up
> and dropped
> them and they remained visible. Perhaps it has to do with whether
> they are identified?
> I went into the menu, squelched by hand, no difference. Pseudo in the
> pack worked
> fine, at least at the beginning.
>
> I could not sell unknown mushroom of poison to general store, but
> could sell it to black market.
GS now only buys back ego- or better lights, spikes, ammo, diggers, and
cloaks. I think my original idea was to have the GS not buy anything
back, but maybe I should let it buy food again. It's just that the way
the GS is handled in code now means that it never deletes items, since
it only restocks essentials. I'd rather not change that now.
> I got the "you found something" in the rubble, but since it was
> squelched
> [I am guessing], there was nothing there.
>
> I could not figure out how to squelch magic books or iron spikes.
>
> I could not figure out how to do auto-inscription. OTOH, I don't know
> if it is supposed to be in 3.0.8. I miss it. P.S. I want them to stick
> from game to game.
Autoinscription is done using the object knowledge browser -- try '~',
select the object choice, then browse to the item you want and hit '{'.
To make settings stick from game to game, save your options on the main
options screen.
> I really really miss the ack when you destroy things, combined with s
> or E for squelching.
> I'm sure I will make a typo eventually and 'k'ill something I need.
> If you squelch properly, you don't 'k'ill so many items, so it is
> useful to have a confirm IMO.
> [If I was the overlord, I'd change 'k' so that it is ONLY an interface
> to squelch.
> I know I am deranged about this.]
Removing the prompt for destruction was a mistake. It's fixed now.
> I get a bunch of asserts because I did not install in /usr/local.
> Presumably ./configure should
> check where you are and set the local/global install/CC flags
> accordingly.
asserts? Are you using the GTK port? What were the messages?
Eddie Grove
> > I could not sell unknown mushroom of poison to general store, but
> > could sell it to black market.
>
> GS now only buys back ego- or better lights, spikes, ammo, diggers, and
> cloaks. I think my original idea was to have the GS not buy anything
> back, but maybe I should let it buy food again. It's just that the way
> the GS is handled in code now means that it never deletes items, since
> it only restocks essentials. I'd rather not change that now.
Whatever. I grepped for mushroom in the changes, didn't find anything, but I
guess I didn't look for that.
> Autoinscription is done using the object knowledge browser -- try '~',
Thank you. I saw some post of yours about that, forgot it.
In keeping with new stores, and NPP, I pressed enter, got browsing.
Somewhere, maybe multiple somewheres, there should be a note to use '{' to
autoinscribe. That is only semi-obvious.
> > I get a bunch of asserts because I did not install in /usr/local.
> > Presumably ./configure should
> > check where you are and set the local/global install/CC flags
> > accordingly.
>
> asserts? Are you using the GTK port? What were the messages?
Oops. It was sloppy of me to report a problem and fail to inlude them.
I just did ./configure and then make on SUSE 10 linux. Presumably gtk.
Make install failed due to permissions for /usr/local, so I manually copied
the executable up one level. It runs ok, apparently using ./lib just fine.
The output is
./angband: Monospace 12
./angband: Width & Height: 9, 19
./angband: Cleaned up.
./angband: ANGBAND_XTRA path = ./lib/xtra/angband.glade
(angband:1988): Gtk-CRITICAL **: gtk_file_system_path_is_local: assertion `path != NULL' failed
That assert appears a lot. I also get the following, once I think.
(angband:1988): libglade-WARNING **: could not find signal handler 'font_selector'.
Eddie
Andrew Sidwell
>
> Thank you. I saw some post of yours about that, forgot it.
>
> In keeping with new stores, and NPP, I pressed enter, got browsing.
> Somewhere, maybe multiple somewheres, there should be a note to use '{' to
> autoinscribe. That is only semi-obvious.
Fixed. Also added a link to the object browser from the squelch menu.
>>> I get a bunch of asserts because I did not install in /usr/local.
>>> Presumably ./configure should
>>> check where you are and set the local/global install/CC flags
>>> accordingly.
>> asserts? Are you using the GTK port? What were the messages?
>
> Oops. It was sloppy of me to report a problem and fail to inlude them.
>
> I just did ./configure and then make on SUSE 10 linux. Presumably gtk.
> Make install failed due to permissions for /usr/local, so I manually copied
> the executable up one level. It runs ok, apparently using ./lib just fine.
Odd. I can't reproduce that on Debian -- "./configure && make && make
install" just moves the executable up a directory here. If this
continues in the final version on a fresh install, then I'll get a
VMWare image of SuSE and see if I can figure it out. (Well, I'm
downloading one as we speak, but it looks like it won't be finished for
a few days, and I don't want to delay 3.0.8 that much.)
> The output is
>
> ./angband: Monospace 12
> ./angband: Width & Height: 9, 19
> ./angband: Cleaned up.
> ./angband: ANGBAND_XTRA path = ./lib/xtra/angband.glade
>
> (angband:1988): Gtk-CRITICAL **: gtk_file_system_path_is_local: assertion `path != NULL' failed
>
> That assert appears a lot. I also get the following, once I think.
I don't know the GTK code well enough to fix it, but it can be safely
ignored, I believe. The "./angband: " bits were just left-in debugging
code, they're removed.
> (angband:1988): libglade-WARNING **: could not find signal handler 'font_selector'.
That's fixed now.
Eddie Grove
> Autoinscription is done using the object knowledge browser -- try '~',
One more thing about this. The knowledge browser puts weapons and armor
first, but who cares about them? All that does is to force one to scroll down
further before being able to examine and/or inscribe flavored objects.
IMO melee weapons and armor should come dead last.
Too late for 3.0.8, but maybe worth keeping in mind for the future.
Eddie
Nick
> Eddie Grove wrote:
> > I just did ./configure and then make on SUSE 10 linux. Presumably gtk.
> > Make install failed due to permissions for /usr/local, so I manually copied
> > the executable up one level. It runs ok, apparently using ./lib just fine.
>
> Odd. I can't reproduce that on Debian -- "./configure && make && make
> install" just moves the executable up a directory here. If this
> continues in the final version on a fresh install, then I'll get a
> VMWare image of SuSE and see if I can figure it out. (Well, I'm
> downloading one as we speak, but it looks like it won't be finished for
> a few days, and I don't want to delay 3.0.8 that much.)
I seem to get the same thing in Ubuntu. I get this:
mk/objective.mk:267: .depend: No such file or directory
as first and last output lines from 'make', then make install gives the same 4
times plus:
/usr/bin/install: cannot create directory `/usr/local/include/angband':
Permission denied
I can mail you the configure output if you want it :)
HTH, Nick.
--
"There is no safety, and there is no end. The word must be heard in silence;
there must be darkness to see the stars. The dance is always danced above the
hollow place, above the terrible abyss."
- The Farthest Shore, Ursula Le Guin
Andrew Sidwell
> On 2007-07-07 22:23:16, Andrew Sidwell <ju...@entai.co.uk> wrote:
>
> I seem to get the same thing in Ubuntu. I get this:
>
> mk/objective.mk:267: .depend: No such file or directory
>
> as first and last output lines from 'make', then make install gives the same 4
> times plus:
>
> /usr/bin/install: cannot create directory `/usr/local/include/angband':
> Permission denied
>
> I can mail you the configure output if you want it :)
Typical that this gets through after the release. :) I think I have a
vague idea of what's going on, but it still confuses me that I'm not
getting errors anywhere. (Not even on Ubuntu.)
Nick
[snip install probs]
Further to this - 3.0.8 final is the same, plus gives this warning in make:
z-file.c: In function ‘safe_setuid_drop’:
z-file.c:40: warning: implicit declaration of function ‘setresgid’
Eddie Grove
> On 2007-07-08 00:08:23, Nick <nckmc...@yahoo.com.au> wrote:
>
> [snip install probs]
>
> Further to this - 3.0.8 final is the same, plus gives this warning in make:
>
> z-file.c: In function ‘safe_setuid_drop’:
> z-file.c:40: warning: implicit declaration of function ‘setresgid’
>
> Nick.
I should have mentioned that I have seen all of the things Nick is talking
about, the .depend error, and something about setresgid.
Curiouser and curiouser. In the non prerelease version, I just got an error
from -lSM. For whatever reason, /usr/X11R6/lib is not in the default library
path. I suppose that's another thing for .configure to check.
I changed the file mk/rules.mk to
LDFLAGS = -lncurses -lSM -lICE -lX11 -L/usr/X11R6/lib
[adding the -L/usr/X11R6/lib] and it successfully built the executable.
I noticed a small bug right away [spellbook does not start in object knowledge
for autoinscribe at turn 0], but it seems to run fine, and fonts are better.
Eddie
Nick
> Typical that this gets through after the release. :)
Yeah, but on an importance scale of 1 to 10, this is a -3.
Nick
> Curiouser and curiouser. In the non prerelease version, I just got an error
> from -lSM. For whatever reason, /usr/X11R6/lib is not in the default library
> path. I suppose that's another thing for .configure to check.
Mine all built fine, but without SDL or GTK modules unless I gave --enable-gtk
--enable-sdl as arguments to configure.
Andrew Sidwell
> On 2007-07-08 00:20:52, Eddie Grove <eddie...@hotmail.com> wrote:
>
>> Curiouser and curiouser. In the non prerelease version, I just got an error
>> from -lSM. For whatever reason, /usr/X11R6/lib is not in the default library
>> path. I suppose that's another thing for .configure to check.
>
> Mine all built fine, but without SDL or GTK modules unless I gave --enable-gtk
> --enable-sdl as arguments to configure.
Yeah, that's intentional. I got some rather bizarre behaviour whilst
doing last-minute testing of the SDL port last night, and the GTK port
doesn't have any method to do font settings or anything yet. Otherwise
they'd be enabled by default, too.
Andrew Sidwell
> Nick <nckmc...@yahoo.com.au> writes:
>
>> On 2007-07-08 00:08:23, Nick <nckmc...@yahoo.com.au> wrote:
>>
>> [snip install probs]
>>
>> Further to this - 3.0.8 final is the same, plus gives this warning in make:
>>
>> z-file.c: In function ‘safe_setuid_drop’:
>> z-file.c:40: warning: implicit declaration of function ‘setresgid’
Looks like I need to add:
#define _GNU_SOURCE
before #including <unistd.h> in h-basic.h to get that warning to go away.
> I should have mentioned that I have seen all of the things Nick is talking
> about, the .depend error, and something about setresgid.
I'm going to investigate the .depend stuff soonish; I get it on fresh
compiles, but after a while it seems to just go away.
> Curiouser and curiouser. In the non prerelease version, I just got an error
> from -lSM. For whatever reason, /usr/X11R6/lib is not in the default library
> path. I suppose that's another thing for .configure to check.
./configure already does check, using the best macros available to me,
how to link and compile X11 code. I can't really fix this -- it's an
autoconf thing.
> I changed the file mk/rules.mk to
> LDFLAGS = -lncurses -lSM -lICE -lX11 -L/usr/X11R6/lib
> [adding the -L/usr/X11R6/lib] and it successfully built the executable.
Actually, the X11 port doesn't even need -lXM or -lICE. Apparently some
X11 apps do, though, so autoconf includes it.
You may notice a running theme with everything being autoconf's fault,
not mine. :)
> I noticed a small bug right away [spellbook does not start in object knowledge
> for autoinscribe at turn 0], but it seems to run fine, and fonts are better.
Noted.