PROJECT: Playest II

[J Bytheway]

Originally posted by Martin Gühmann
So you mean if I use a global variable that is set in the messagebox but is used in the original event handler to determine which of the options should be executed might work.

I doubt it. The global variable would be set on the local machine, but not remotely, so you'd still have the same problem. To fix this it is necessary the the way messagebox button code is handled in MP games to be significantly changed.

[J Bytheway]

Well, I'm back now, and I'd like to put together a new playtest release, but because I've been out of touch for a few weeks, I'm slightly wary. I also think my existing version is not right (because of BureauBert's problems), so I should probably start from scratch anyway...

[Martin Gühmann]

Originally posted by J Bytheway
(because of BureauBert's problems)

I must missed something here, what problems of BureauBert do you mean?

Originally posted by J Bytheway
Well, I'm back now, and I'd like to put together a new playtest release, but because I've been out of touch for a few weeks, I'm slightly wary. I also think my existing version is not right (because of BureauBert's problems), so I should probably start from scratch anyway...

Actual I would like to wait for Fromafar first so that we can include everything in the AlteredSource file thread maybe we can reduce the sprite limit at first to 255, so that we can include CityMod3. Another thing I would like do first is the restoration of backward compatibility of the civilisation.txt and I would also like to do a final solution of the trade issure first, and put it into an all file archieve.

-Martin

[J Bytheway]

Originally posted by Martin Gühmann
I must missed something here, what problems of BureauBert do you mean?

The problems he was having with AdvanceDB, etc., which you couldn't replicate.

Actual I would like to wait for Fromafar first so that we can include everything in the AlteredSource file thread maybe we can reduce the sprite limit at first to 255, so that we can include CityMod3. Another thing I would like do first is the restoration of backward compatibility of the civilisation.txt and I would also like to do a final solution of the trade issure first, and put it into an all file archieve.

I'd like to wait until all the altered source files are appropriately collated, so I'll probably wait until you put them together as and when you see fit.

Also, some bug reports of things I noticed when I was playing hotseat recently:

It doesn't work properly if you change the colour of the first player away from the first (You get the interface to one player while it is another player's turn and cannot do anything).

The messages about starting and stopping wonders are mostly sent to the wrong players.

[Fromafar]

Status update

Progress is slow, but steady. Martin's 2004.09.08 update and my 2004.09.09 update have been integrated now, together with MemberClass.cpp from the 2004.08.28 citymod3 update. Stability is still good, and mods are playable, even with the debug version. At least, if you are prepared to click away zillions of Assert pop-ups. It looks like most modded AI players are quite stubborn in trying to negotiate with barbarians and non-existing players each turn .

I have had a look at the change of the accepted library text size in GreatLibrary.cpp, but I fail to see its purpose. IMO, 4096 characters of text should suffice to provide some basic explanation or atmosphere. The additional memory is better spent on "real" game concepts. In particular when keeping the k_MAX_INTERP_LEN and k_MAX_GL_ENTRY constants at 4096, so the extra text will never be used.

[Martin Gühmann]

Originally posted by Fromafar
At least, if you are prepared to click away zillions of Assert pop-ups. It looks like most modded AI players are quite stubborn in trying to negotiate with barbarians and non-existing players each turn .

One problem with mods is that they use Diplomod and Peter has messed around with the ConsiderNewProposal function. At least that is one reason that mods are not 100% compartible. Possibly another problem are the other changes of Diplomod.

Originally posted by Fromafar
I have had a look at the change of the accepted library text size in GreatLibrary.cpp, but I fail to see its purpose. IMO, 4096 characters of text should suffice to provide some basic explanation or atmosphere. The additional memory is better spent on "real" game concepts. In particular when keeping the k_MAX_INTERP_LEN and k_MAX_GL_ENTRY constants at 4096, so the extra text will never be used.

For the German version 4096 characters aren't enough, I need some more otherwise one entry is truncated. And another problem is for example a japanese version even if I don't know the Great Libary.

And as MAX_ENTRY is only relevant for the parsing buffer:

 		const int MAX_ENTRY = 8192;
		char the_name[MAX_NAME];

And is later copied into an array with the correct size:

char * name_copy = new char[strlen(the_name)+1];
char * entry_copy = new char[strlen(the_entry)+1];
strcpy(name_copy, the_name);
strcpy(entry_copy, the_entry);

And finally the char arrays with the length that is only needed are put into the hash table.

m_great_library_info->Add_To_Hash_Table(name_copy,entry_copy);

In the end only 4096 are needed in additon to the original version and all of the entries of the German Great Libary can be parsed without truncating them.

-Martin

[Fromafar]

Can you find out what sizes are needed (worst case) for the German texts? If we don't update k_MAX_INTERP_LEN and k_MAX_GL_ENTRY, you may store the full text, but it will get truncated later in stringUtils_interpret. OTOH, I don't like increasing those values too much. When you are calling stringUtils_interpret from GreatLibraryWindow::LoadText, you end up with 4 such monster blocks on the stack at the same time.

Maybe I am a bit too paranoid here, but I don't know what stack size is considered normal for Windows programs.

[Martin Gühmann]

Originally posted by Fromafar
Can you find out what sizes are needed (worst case) for the German texts? If we don't update k_MAX_INTERP_LEN and k_MAX_GL_ENTRY, you may store the full text, but it will get truncated later in stringUtils_interpret. OTOH, I don't like increasing those values too much. When you are calling stringUtils_interpret from GreatLibraryWindow::LoadText, you end up with 4 such monster blocks on the stack at the same time.

Maybe I am a bit too paranoid here, but I don't know what stack size is considered normal for Windows programs.

Well according to that this piece of information your 4 big fat terrible mosters turn into 4 tiny little critters. And I don't see that the "/F (Set Stack Size)" option is set so you should have 1MB for your stack.

-Martin

[Fromafar]

OK, it probably is mostly harmless then. When your daily work sometimes involves having to fit everything into about 300 bytes of stack, you learn to be suspicious of statements like "just increase the size of [...] shouldn't be the problem". We just have to keep in mind that it takes as little as 128 tiny little critters to mess up with the program giganticly.

With the 2004.09.04 sstateevent update, the 2004.08.15 personality.cdb is working fine, so the only thing left to do is the inclusion of the army movement changes. Oh, and maybe having a poll about which civilisations to include in the game.

[Martin Gühmann]

Originally posted by Fromafar
Oh, and maybe having a poll about which civilisations to include in the game.

Maybe we should update then the interface so that I don't have to put in and out the civs. And actual I would be careful about the Celts, it looks like someone merged the Irish and the Welsh to form the Celts. Doing this is a little bit inacurate historicly. From Caesar we know that the Gauls called themselves Celts. Therefore the possible city names have to come from a wider area than from the British Isles.

-Martin

[Martin Gühmann]

Here is another crash log.

  0x00629513  [?CheckForEasterEggs@ChatWindow@@QAEHPAD@Z + 0x1c3]
  0x00629990  [?ChatCallback@ChatWindow@@SAXPAVaui_Control@@KKPAX@Z + 0x90]
  0x005c1a65  [?HitEnter@aui_TextField@@SAXPAUHWND__@@@Z + 0x35]
  0x005c1ecf  [?TextFieldWindowProc@@YGJPAUHWND__@@IIJ@Z + 0x5f]
  0xbff7363b  [(kernel) + 0x0]

I think it is now time that we can test all together again. Therefore I provided a new All source pack from that a new playtest version can be compiled. A lot of new stuff is coming.

-Martin

[J Bytheway]

Thanks Martin, I'll try and put it together tomorrow.

[Martin Gühmann]

Originally posted by J Bytheway
Thanks Martin, I'll try and put it together tomorrow.

I hope you didn't downloaded it, yet. Obviously due to a copy'n'paste array the link in the altered source thread pointed to the all pack from August. If you have already downloaded it I must apologize for this mistake.

-Martin

[J Bytheway]

I was a little puzzled by that date - I noticed it just now and assumed it was just a wrong filename. I'll go get the new one.

[J Bytheway]

FYI, the links still don't quite work - you should capitalize the "All" .