-Crawl- 4.1.1a released to Will Shattuck

Archived from groups: rec.games.roguelike.misc (More info?)

Greetings one and all!

I've just kicked, punched, and shaken the 4.1.1a code into something
that is at least pretending to be an executable and sent that off to
Will Shattuck (a fellow who gave an offer of hosting an executable in a
message in the "compiling 4.1.0a" thread).

Notes:
- This is my first ever successful or, perhaps better put, "successful"
compile using Bloodshed Dev-C++ -- Borland was not only not only
refusing to compile, it was doing so in a way that I couldn't figure out
how to do anything about, so I moved to plan B and, after much googling
and screaming, plan B seems to have worked. Thus don't necessarily blame
Brent for any problems -- post them in this thread and I'll see if
there's something I did wrong in the compile and can do something about.

- Darshan's sound-playing feature from the patch didn't make the compile
-- I promised my wife I'd be out of here by 8 and it was 7 when the
compiler reminded me that nope, I'm not gonna get that one for free
(thank God I'd re-read Darshan's patch notes just yesterday).

Thank-you's:
- Darshan for the de facto encouragement given by his advice that pushed
through the previous Borland compile
- Pekka Nurminen, without whose advice on mingw compiles during my
abortive "Bowcrawl" project in spring I would never have been able to
make this mingw-based compile today
- Jesse Welton, whose extremely timely note this European evening
regarding Brent's accidentally retaining his abl-show.o in the sources
directory resolved a problem that I would not have otherwise been able
to fix in time to release this evening
- Will once again for the hosting
- SC Skipsey once again for his fix to the windows-unsupported
"extra-random" seed problem (Brent's modified the sources to see around
it, but only if a certain flag is set, which isn't of course in the
linux makefile, on which Pekka's mingw makefile solution is based, so it
could have taken quite a while to have figured out what was going on
otherwise).
- anyone I forgot :-)

Erik

--
My sig file says something stupid at the moment and I don't have time to
change it, so you're seeing this instead.

Archived from groups: rec.games.roguelike.misc (More info?)

On 2 Aug 2005 13:23:40 -0700, "mrkrrtft" wrote:

>Windows Binaries:
>
>http://www.thewholeclan.com/will/crawl
>
>After tomorrow you can access it at:
>
>http://crawl.thewholeclan.com
>
>gotta love subdomains
>
>Enjoy.

Thanks for trying, but 771KB isn't the whole crawl, is it?
Winzip claims it's a broken file.

Archived from groups: rec.games.roguelike.misc (More info?)

On 2 Aug 2005 13:41:40 -0700, "Erik Piper" wrote:

>
>Lauri Vallo wrote:
>> On 2 Aug 2005 13:23:40 -0700, "mrkrrtft" wrote:
>>
>> >Windows Binaries:
>> >
>> >http://www.thewholeclan.com/will/crawl
>> >
>> >After tomorrow you can access it at:
>> >
>> >http://crawl.thewholeclan.com
>> >
>> >gotta love subdomains
>> >
>> >Enjoy.
>>
>> Thanks for trying, but 771KB isn't the whole crawl, is it?
>> Winzip claims it's a broken file.
>
>I also e-mailed it to myself at home and was able to unzip it
>successfully. However, it has a couple of bugs -- see the new thread
>I'm starting on the subject.

Yeah yeah it will always have some bugs. I won't wait forever to get
to play the game.

Every other crawl package I have is over 1 megabyte. Binaries are
close to two, so......

Archived from groups: rec.games.roguelike.misc (More info?)

On Tue, 02 Aug 2005 19:45:48 +0200, Erik Piper wrote:

>Greetings one and all!
>
>I've just kicked, punched, and shaken the 4.1.1a code into something
>that is at least pretending to be an executable and sent that off to
>Will Shattuck (a fellow who gave an offer of hosting an executable in a
>message in the "compiling 4.1.0a" thread).

Eagerly waiting for 4.1.2! (in a not magic locked .zip)
*smile*

Archived from groups: rec.games.roguelike.misc (More info?)

"Erik Piper" <erikNOSPAM@sky.cz> wrote in message
newscobgd$q17$1@domitilla.aioe.org...
> Greetings one and all!
>
> I've just kicked, punched, and shaken the 4.1.1a code into something that
> is at least pretending to be an executable and sent that off to Will
> Shattuck (a fellow who gave an offer of hosting an executable in a message
> in the "compiling 4.1.0a" thread).
>
> Notes:
> - This is my first ever successful or, perhaps better put, "successful"
> compile using Bloodshed Dev-C++ -- Borland was not only not only refusing
> to compile, it was doing so in a way that I couldn't figure out

Another option would be to use Microsoft's C++ Compiler Toolkit
(& Microsoft SDK).

I compiled 4.1.2 alpha - it took a couple of hours to fix all errors
& check all warnings.

Here is just list of places where errors / warnings were issued.
Most of those <foo> may not be initialized warnings are not valid, but
there are some cases that should be fixed.
(In addition to these, current makefiles do not work, so I compiled using
batch files. Then you need snprinf implementation and modify AppHdr.h &
libw32c.cc to support Microsoft's compiler)

>> abl-show.cc:

&& (do_invoke == Curr_abil[loopy].is_invocation))

warning - comparision between integer & boolean.

>> abyss.cc:

you.x_pos = x;
you.y_pos = y;

warning - x & y might not be initialized

>> acr.cc:

if (you.attribute[ATTR_WAS_SILENCED] != its_quiet)

warning - comparision between integer & boolean or two bools,
bools a, b do not have same state if ( a && !b ) || ( !a && b ).

>> beam.cc:

if (!in_bounds( tx, ty ))

warning: tx, ty might not be initialized

#ifdef WIN32CONSOLE
if (!beam.is_tracer)
setBuffering(oldValue);
#endif

warning: oldValue might not be initialized

>> fight.cc:

if (ur_armed && !weapon_skill( you.inv[wpn] ) == SK_STAVES)

warning: second part never holds, probably should be
weapon_skill( you.inv[wpn] ) != SK_STAVES?

>> Kills.cc:

kill_def::kill_def &k = kills[descriptor];

error: it should be just kill_def. (kill_def::kill_def refers to
constructor)

k = kill_def::kill_def(mon);

error: just kill_def instead of kill_def::kill_def.

std::map<___monster_desc, kill_def::kill_def,
___monster_desc::less_than>::

error: just kill_def instead of kill_def::kill_def.

const kill_def::kill_def &k = iter->second;

error: just kill_def instead of kill_def::kill_def.

std::map<___monster_desc, kill_def::kill_def,
___monster_desc::less_than>::

error: just kill_def instead of kill_def::kill_def.

>> kills.h:

struct less_than
{
bool operator () ( const ___monster_desc &m1,
const ___monster_desc &m2)

error: should be const function.

std::map<___monster_desc, kill_def::kill_def,
___monster_desc::less_than> kills;

error: kill_def instead of kill_def::kill_def.

>> libw32c.cc:

if (curstype == current_cursor)

warning: comparision between bool & integer.

>> macro.cc:

warning: there was a warning that num variable may be used before
initialized,
but I do not remember exact location.

>> monstuff.cc:

mon->target_x = tx;
mon->target_y = ty;

warning: tx, ty may be used before initialized

>> spells4.cc:

if (create_monster( mon, beh, 3 ) != -1)

warning: mon may be used before initialized

>> travel.cc:

for (int dir = 0; dir < 8; (dir += 2) == 8 && (dir = 1))

warning about that dir = 1 assignment.. well, that assignment could be
easily avoided.

>> travel.h:

struct less_than {
bool operator () ( const level_id &first, const level_id &second ) {

error: operator() should be const function

Archived from groups: rec.games.roguelike.misc (More info?)

Erik Piper wrote:
> Lauri Vallo wrote:
>
>>On Tue, 02 Aug 2005 19:45:48 +0200, Erik Piper wrote:
>>
>>
>>>Greetings one and all!
>>>
>>>I've just kicked, punched, and shaken the 4.1.1a code into something
>>>that is at least pretending to be an executable and sent that off to
>>>Will Shattuck (a fellow who gave an offer of hosting an executable in a
>>>message in the "compiling 4.1.0a" thread).
>>
>>Eagerly waiting for 4.1.2! (in a not magic locked .zip)
>>*smile*
>
>
> I'm not sure what to do about it -- both the 4.1.1 zip and the new one
> I sent Will today for 4.1.2 have unzipped fine on my win2k machine,
> which would seem like a good sign since they were produced on a winxp
> machine. I've probably been using the zipper/unzipper supplied with my
> favorite 2-pane manager, Servant Salamander, for every step, but I
> can't imagine that making a difference. Moreover, I'd think that Will
> would have tested the zip before releasing it... hmm, actually I never
> tested what Will posted, just my own zip. But I can't imagine him
> having done anything with it before posting (other than changing the
> name from *.aaa back to *.zip due to gmail not liking zips)... this is
> all pretty weird. :-(
>
> e.
>

could it be that they have been uploaded with a ftp client in ascii
rather then binary mode?

Archived from groups: rec.games.roguelike.misc (More info?)

Erik Pipe" wrote:
>kenneth marken wrote:

>> could it be that they have been uploaded with a ftp client in ascii
>> rather then binary mode?
>
>That could well have been the problem. Been. It's fixed. I posted a
>message on it via aieo.org, but they're having chronic troubles lately,
>so I might get the word out faster from here.

Thanks for solving it.

Now if I could get it to run..
So far 4.1.1 and 4.1.2 give me this on win98:

Duplicated class 'x' in list_order.
Error setting input code page to 437. (errno: 120)

Archived from groups: rec.games.roguelike.misc (More info?)

Lauri Vallo wrote:

> Erik Pipe" wrote:
>
>> kenneth marken wrote:
>
>
>
>>> could it be that they have been uploaded with a ftp client in ascii
>>> rather then binary mode?
>>
>>
>> That could well have been the problem. Been. It's fixed. I posted a
>> message on it via aieo.org, but they're having chronic troubles lately,
>> so I might get the word out faster from here.
>
>
>
> Thanks for solving it.
> Now if I could get it to run..
> So far 4.1.1 and 4.1.2 give me this on win98:
>
> Duplicated class 'x' in list_order.
> Error setting input code page to 437. (errno: 120)


Glad to see a confirmation of that first message. :-) On my two testing
computers (WinXP EN and Win2K CZ), I get it too (never saw it long
enough to confirm the precise text), but the program then goes on to
run. So the no-go seems to be in the code page problem. Try setting
char_set = ASCII in init.txt and see if that helps.

e.

Archived from groups: rec.games.roguelike.misc (More info?)

Erik Piper wrote:

>Glad to see a confirmation of that first message. :-) On my two testing
>computers (WinXP EN and Win2K CZ), I get it too (never saw it long
>enough to confirm the precise text), but the program then goes on to
>run. So the no-go seems to be in the code page problem. Try setting
>char_set = ASCII in init.txt and see if that helps.

*pets his dos prompt*

Thanks, I now tried that, and IBM and DEC, and starting with crawl
-plain, but nothing changes.

Archived from groups: rec.games.roguelike.misc (More info?)

"Erik Piper" <erik@sky.cz> writes:
> Lauri Vallo wrote:
>> >Error setting input code page to 437. (errno: 120)
[...]
> http://svn.haxx.se/dev/archive-2003-02/1507.shtml

> Hmmmm...

Edit libw32c.cc and remove the lines starting with

InputCP = GetConsoleCP();

down to the end of init_libw32c(). I haven't verified that it compiles
and works, but it should.

--
Darshan Shaligram <scintilla@gmail.com> Deus vult