elexis

From my personal player pov, 200 would be ok for me and 300 have always appeared like pure spam games. (Also often players just the 8 players without reducing the population limit, which totally doesn't work performance wise if everyone reaches the limit.)

Hm, the WonderVictory message hadn't been changed, that also uses "will have won".

The problem only appears to people who administrate a multiplayer lobby (almost noone) and who sends a "message of the day" (almost noone of that subset), so got an ocean of bigger fishs to fry. Stopping to send messages you don't want to receive is probably the thing you want But I agree that implementing "message of the day support" is to address any claims that our code is broken, so committed support of these messages in r20005. You will now get a message box instead of an error. If you don't want those, don't send the messages.

Ah, so now it's "hi", not "hello". Was that an "announcement" or "message of the day"? gameboy, when this issue is fixed, you will still get "hi" and "hello", just a message box instead of a warning.

refs D740

Well, the code works as the comments intend it to work: Double-clicking: // Select similar units regardless of rank Triple-clicking: // Select units matching exact template name (same rank) It's just not expected by the user I guess.

The templates are indeed identical, so it's a code bug, not a template bug. The triple-click code must check for the template name instead of the SelectionGroup name. Reading the code (input.js), that's exactly what it does. // Check for double click or triple click if (ev.clicks == 2) { // Select similar units regardless of rank templateToMatch = GetEntityState(clickedEntity).identity.selectionGroupName; if (templateToMatch) matchRank = false; else // No selection group name defined, so fall back to exact match templateToMatch = GetEntityState(clickedEntity).template; } else // Triple click // Select units matching exact template name (same rank) templateToMatch = GetEntityState(clickedEntity).template; Since that comment exists since r11178, I'm lucky it's not another bug of the alpha 22 doubleclick patch The rank parsing was odd in different places already. @Imarok did some work with that already, but he's busy atm. Can you report this bug at http://trac.wildfiregames.com/?

The chat window label still didn't pass the long-strings test. So just going to add the auto-resizing code later. The option categories will be changed, then the labels will be changed in that way as well. I'll keep that as a review for then The string fixes and the Ken Wood release name were committed in revision 20000.

Checked every string in the GUI now that we try to use title case consistently. The new option categories proposed in D797 don't intersect with these stringchanges. The prelobby has buttons that contain sentences that don't end with periods. Exception ot hte title case rule works for me for now. The Developer overlay consistently doesn'tuse title case and don't see a reason to change that. That ingame "Game Speed" button was wrong. There is a "Player name" label in the summary screen, but that doesn't look like it wants to use title case. Strings like "Stone exchanged" follow and it seems more natural this way to me. If you disagree, I can change this string later too. The multiple selection details capture bar had shown "Capture points: foo". "Hitpoints" and "Capture points" were not used, so deleted.

That's not a patch, that's the file changed by the patch linked in that ticket. That patch does the opposite of what you want, so again: do not apply the patch.

(And if we're already posting links, the videos "flat earth clues" and "two hundred proofs the earth is not a spinning globe" were more entertaining presentations of oddities, casual bullshit and science fiction than that website IMO)

Just for completeness, there are also some random map scripts introduced by _kali in Alpha 20 and Alpha 21 that use heightmaps. _kali used custom tools to create the maps. It should be better documented. (We still need to reproduce the process, so that we can convert these maps to a round shape.) Never used it, nor read the code. But I assume the highest quality wins. The atlas menu entry stating that JPEG files can be loaded is a leftover from the time it was supported, see 17870 / #2828. The texture decoding was introduced here: https://trac.wildfiregames.com/changeset/12308/ps/trunk/source/tools/atlas/GameInterface/Handlers/MapHandlers.cpp There weren't too many changes https://code.wildfiregames.com/source/0ad/browse/ps/trunk/source/tools/atlas/GameInterface/Handlers/MapHandlers.cpp?view=blame I don't expect the results to vary (other than loss of quality when compressing the image again). Maybe Vladislav knows more.

Does anyone see a way of adding these techs without new icons? There are images of statues of most of these deities that could be taken as a template. Remembering some commits and looking at the incomplete 2D Art part of the credits section, we can see @Pureon , @wowgetoffyourcellphone and @Lion.Kanzen made some icons. Maybe ask Lion if he has the qualification, time and interest to do so.

I wonder too when I see all the old random maps. Many have no terrain height differences to speak of and look extremely ugly to me. For example the all-time-lobby-favorite map Mainland (here with trees removed): This has become much better in the last years. For example wraitii's Pyrennean Sierra map from 2012 has a nice mountain range and slowly increasing height. _kalis maps from alpha 20 / 2015 introduced the gameplay-wise really interesting bluffs (most prominent map Ambush). With Alpha 21 / 2016 _kali had added real-world maps (most probably excluding curvature) Red Sea, Mediterranean and so forth that look much better. In early 2017, when Hannibal_Barca wanted to give us more flat earth maps, we had identified the recipe for beautiful maps: The createBumps call, which should be more often used for these ugly maps. Like mario galaxy? 0 A.D. maps are not large enough to reveal any curvature of the theoretical dimensions, because largest map size * TERRAIN_TILE_SIZE = 512 * 4 = 2048 meters if I'm not mistaken. (Other than the 0 A.D. part, it's an interesting topic but as offtopic as it might get.)

Do _not_ use this patch (https://www.ejabberd.im/node/4857). Just tell me if you have "hello" in your ejabberd config at /etc/ejabberd/ejabberd.cfg

@gameboy So it's not the pidgin message but your ejabberd server config. search for "hello" in /etc/ejabberd/ejabberd.cfg

@gameboy which message did you send in pidgin? "announcement" or "message of the day"? Please send "not hello" in pidgin. Check whether you still get "hello" in 0ad.

Posted something in the ticket . @gameboy do you mean "message of the day" or "announcement" with "send the message to all users online"? Can you try this? 1. apply the warn(uneval(msg)); patch above 2. Send a pidgin message "Not hello" 3. Join lobby with 0AD and post the warnings to the forums. 4. Repeat step 3 two times

Okay thanks. @gameboy How did you send the "hello" message in pidgin? Which button did you press exactly in pidgin? You sent it only once in pidgin but get the error on each login?

SpiritOfTheLaw is tha man! He didn't make a video about this by any chance?

Ok, that's good so far! @gameboy can you test this patch and give the output?

That's weird, I never get it when logging in with an existing account in alpha 22 or svn. Does anyone else get these warnings in that case? gameboy, you get it on every login? If so, try please try this patch and post the warning log or screenshot: lobby_debug.patch

I'd definitely fail the elevator test (in the best possible scenario)

When you tested what exactly? Wildfire Games svn lobby or your own lobby? When creating a new account or when logging in with an existing account?

As of alpha 22, the engine can deal with 8 players * 150 pop, so 12 players * 100 pop and 16 players * 75 pop are quite realistic. All the random maps will need consideration, but nothing impossible. We also had few instances with 8 players and 16 observers already, so there's no real overhead (only the rejoin process is a bit laggy). If we already do some effort to break the barrier, we might just as well go all the way (i.e. 256, unless we need to allocate memory for the maximum in advance somewhere).