[Announce] NPPAngband 0.5.1 WIP

[nppangband]

Today a preliminary version of NPPAngband 0.5.1 was released (WIP is a "work in progress"). A more detailed announcement is found here:

http://forum.nppangband.org/viewtopic.php?id=929

Source code:

http://download.nppangband.org/npp-0.5.1-src-WIP1.zip

Windows executable:

http://download.nppangband.org/npp-0.5.1-win-WIP1.zip

This is essentially NPP 0.5.0, but with an updated codebase, and a couple dozen minor features added as well. NPP050 was based on Angband 3.0.6. This release incorporates the backbone of Angband 3.1.2v2, and most of the UI improvements. I still need to incorporate the Angband ui improvements to the knowledge menus, the options menus, and the squelch menus. I will be doing this over the next couple weeks. And I will be releasing updates probably every couple days as I fix things and add other features & improvements from Angband 3.1.2v2.

Even though I have tested this as best as one person can, I expect it will be moderately buggy. At this point it is for harcore NPP fans, people willing to playtest, or for anyone kind enough to attempt to compile this for other operating systems (this version has all of the makefiles and main-xxx files from the current Angband, so it should compile about the same). I expect in about a month or so the code will be pretty clean, and I will have most of the recent Angband changes worked into NPP, and then I will do an official release of 0.5.1.

[ekolis]

Cool!

I did find that 'l'ooking at shop services displays info on items stocked by the shop, not services...

And I'm kinda surprised that there's no detailed spell-browser...

[ekolis]

Oh, yeah, is it intentional that items in shops have their prices in all sorts of fancy colors (matching the item name colors)? It's kinda neat, but on the other hand, I do like seeing the price of the selected item highlighted in cyan.... Maybe if you used a lighter version of the item color? And maybe use that lighter version for the item name, too? Then the whole row would be highlighted

[Pete Mack]

Bugs:
From Makefile.win:
Makefile.src lines 74, 76: missing '\' terminator
Also, some instances of leading <blank> rather than <tab>
attack.c:
line 77: declaration after code not legal in ansi pedantic
dungeon.c:
make: *** No rule to make target `option.h', needed by `dungeon.o'. Stop.

[Daniel Fishman]

...
I still need to incorporate the Angband ui improvements to ... the squelch menus.
...

I have to wonder, in what way is V squelch ui better than NPP squelch ui?

[nppangband]

Cool!

I did find that 'l'ooking at shop services displays info on items stocked by the shop, not services...

I fixed this, thanks.

And I'm kinda surprised that there's no detailed spell-browser...

I had already done something similar in NPP. All spells give a detailed description when the spellbook is browsed. I will double check Angband to see if I missed some additional features.

[nppangband]

I have to wonder, in what way is V squelch ui better than NPP squelch ui?

I assume it has bigscreen support, and you can input with the mouse. I am going to keep the NPP squelch system, though. I haven't looked at it too closely yet, though. The Angband dev team has been working hard the last couple years. I have *ALOT* of catching up to do.

[nppangband]

Oh, yeah, is it intentional that items in shops have their prices in all sorts of fancy colors (matching the item name colors)? It's kinda neat, but on the other hand, I do like seeing the price of the selected item highlighted in cyan.... Maybe if you used a lighter version of the item color? And maybe use that lighter version for the item name, too? Then the whole row would be highlighted

I didn't notice that the selected item's colors were inconsistent. The whole line now shows up as light blue (cyan?). Thanks again.

[nppangband]

Bugs:
From Makefile.win:
Makefile.src lines 74, 76: missing '\' terminator.

Fixed. Thanks.

Also, some instances of leading <blank> rather than <tab>
attack.c:
line 77: declaration after code not legal in ansi pedantic.

Whoops. The object_flags line just needed to be moved down. I used to do a test compile in cygwin before any release, because that was good at picking up those "mixed declarations and code" problems, and LCC-WIN doesn't notice those. I should take up that practice again.

dungeon.c:
make: *** No rule to make target `option.h', needed by `dungeon.o'. Stop.

Pasting error. NPP doesn't have an option.h (yet). The entry a little further down for quest.o has the same problem.

Thanks for trying Pete, I cleaned all that up. Are there any more?

[Tatami]

I just rolled a high elf ranger and the character description is missing a few spaces between the different aspects of the char description. I know I know, stop the presses the game doesn't function..

[nppangband]

I just rolled a high elf ranger and the character description is missing a few spaces between the different aspects of the char description. I know I know, stop the presses the game doesn't function..

That was a tough one, but I think I came up with something that will hopefully work. Thanks.

[Tatami]

That was a tough one, but I think I came up with something that will hopefully work. Thanks.

You guys are miracle workers.

Seriously: Will give NPP another hard player test. Big cheers for the hard wrok! (dont't wrok the grass por favore..)

[Pete Mack]

OK, I got a bit further by removing:
player/.*.h
object/.*.h
monster/.*.h
from Makefile.inc

However, I am still seeing:

C:\Projects\Exp\npp-0.5.1-src-WIP\src>make -f Makefile.win
gcc -DWINDOWS -Wall -Wextra -Wno-unused-parameter -std=iso9899:199409 -pedantic
-O2 -I. -mno-cygwin -c -o dungeon.o dungeon.c
dungeon.c: In function `recharge_objects':
dungeon.c:524: warning: ISO C90 forbids mixed declarations and code
dungeon.c:540: error: initializer element is not constant
make: *** [dungeon.o] Error 1
C:\Projects\Exp\npp-0.5.1-src-WIP\src>make -f Makefile.win
gcc -DWINDOWS -Wall -Wextra -Wno-unused-parameter -std=iso9899:199409 -pedantic
-O2 -I. -mno-cygwin -c -o dungeon.o dungeon.c
dungeon.c: In function `recharge_objects':
dungeon.c:541: error: initializer element is not constant
make: *** [dungeon.o] Error 1
gcc -DWINDOWS -Wall -Wextra -Wno-unused-parameter -std=iso9899:199409 -pedantic
-O2 -I. -mno-cygwin -c -o feature.o feature.c
feature.c: In function `get_feat_num':
feature.c:1559: error: syntax error before '/' token
make: [feature.o] Error 1 (ignored)
gcc -DWINDOWS -Wall -Wextra -Wno-unused-parameter -std=iso9899:199409 -pedantic
-O2 -I. -mno-cygwin -c -o melee1.o melee1.c
melee1.c: In function `make_attack_ranged':
melee1.c:3986: error: `true' undeclared (first use in this function)
melee1.c:3986: error: (Each undeclared identifier is reported only once
melee1.c:3986: error: for each function it appears in.)
make: [melee1.o] Error 1 (ignored)

I recommend trying to build with:

CFLAGS += $(WARNINGS) -std=iso9899:199409 -pedantic -O2 -I.
or update all gcc-dependent makefiles to use the options of your choice.


There are also lots of warnings, some of them are scary.
Like:

cmd4.c: In function `do_cmd_change_name':
cmd4.c:215: warning: comparison is always false due to limited range of data typ
e
cmd4.c:226: warning: comparison is always false due to limited range of data typ
e
cmd4.c: In function `do_cmd_messages':
cmd4.c:387: warning: comparison is always false due to limited range of data typ
e
cmd4.c:397: warning: comparison is always false due to limited range of data typ
e
cmd4.c:407: warning: comparison is always false due to limited range of data typ
e
cmd4.c:414: warning: comparison is always false due to limited range of data typ
e
cmd4.c: In function `do_cmd_macro_aux':
cmd4.c:685: warning: comparison is always true due to limited range of data type

I find MinGW a lot more useful than lcc. It's a lot more standards-compliant (whichever standard you pick), and the warnings are much, much stronger.

Visual Studio can help too--the MSVC compiler catches different errors, although setting up a project file is a bit of a pain.

Edit:
PS: the version of cmd-know.c you are using is not copyright by me. (I had nothing to do with it. I want nothing to do with it...)

[nppangband]

I find MinGW a lot more useful than lcc. It's a lot more standards-compliant (whichever standard you pick), and the warnings are much, much stronger.

Thanks. I intended to download and workwith MinGW today, but I was killing bugs all day long today.

I think I know what most of those warnings are & the fixes are simple. I will get to work on them & get a clean compile with MinGW.

[ekolis]

Do you have a repository out there we can download the latest code from, or will we have to wait for the next WIP version?