--- Log opened Thu Oct 05 00:00:22 2017
20171005 00:02:24< shadowm> vultraz: Updated.
20171005 00:05:19< sevu> I thin it's reasonable to hide the idle AI. In MP create, you can give a side to a human player, an AI, nobody, or a fourth option, which seems to have the same effect as to give it to the host... For the AI you can set idle Ai, the only difference compared to setting nobody is that this side has a leader.. So, in short, there are duplicate options in the MP create screen 
20171005 00:06:42< sevu> (for campaigns the idle AI makes of cause sense. But there you define it in the source code. Thought I see that the commit enables it because otherwise this is buggy)
20171005 00:08:47< sevu> Sorry, my spelling is horrible. It's in the middle of the night, time to sleep.
20171005 00:10:39< mattsc> sevu: hmm, okay …
20171005 00:11:57< mattsc> something is needed on the SP side though, and in that case it is definitely not just a dev AI but something rather commonly used
20171005 00:12:34< mattsc> this looks like the idle_ai was originally meant for a different purpose than it is being used for now
20171005 00:17:20< celticminstrel> The idle AI is already hidden, isn't it?
20171005 00:17:46< mattsc> I just unhid it in order to fix issie 1955
20171005 00:17:53< celticminstrel> Eh?
20171005 00:18:12< mattsc> it’s explained in the commit message
20171005 00:18:26< celticminstrel> It'll take me awhile to get there. >_>
20171005 00:18:32< mattsc> ok
20171005 00:18:40< mattsc> well, I’ll be afk shortly anyway
20171005 00:19:11< mattsc> looks like while this fixes that issue, it might have another potentially undesirable effect
20171005 00:20:19< mattsc> I don’t see another easy solution though; it would probably require working with the AI config code
20171005 00:20:29< mattsc> as in, the code reading the configs
20171005 00:29:11-!- sevu [~Shiki@p54855482.dip0.t-ipconnect.de] has quit [Quit: Verlassend]
20171005 00:32:35< irker163> wesnoth: Ignacio R. Morelle wesnoth:master 1df298511881 / src/ (7 files in 3 dirs): campaignd: Send uploading clients a list of illegal names when any are found https://github.com/wesnoth/wesnoth/commit/1df298511881e30bdab2ff9a473631dd1dc2dfb0
20171005 00:50:10< irker163> wesnoth: Ignacio R. Morelle wesnoth:master ffa581d88bfd / changelog: Update changelog to account for add-ons server/client changes https://github.com/wesnoth/wesnoth/commit/ffa581d88bfdafe6d253d62c2994148c01abb85f
20171005 00:51:35-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20171005 00:52:08-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171005 00:59:45< celticminstrel> So regarding the newly-unhidden idle AI, we could easily just a) hardcode an exception and not list it, or b) add a "hidden" attribute in [ai].
20171005 01:04:38< mattsc> celticminstrel: sure, sounds good to me
20171005 01:05:05< mattsc> we could also do it the other way around: re-hide it and hard-code it’s inclusion in the configs
20171005 01:05:13< celticminstrel> Hm?
20171005 01:05:18< celticminstrel> What do you mean?
20171005 01:06:00< mattsc> It’s config was not read, and the AI’s id therefore not know, when the game was started without debug mode activated
20171005 01:06:47< mattsc> Anyways, I’m happy with the suggestion you made.
20171005 01:08:07< celticminstrel> I don't understand the suggestion you made.
20171005 01:08:57< mattsc> We could hard-code the inclusion of the idle_ai config (or just it’s empty stage) even if it is in a “hidden” directory.
20171005 01:09:22< mattsc> … if it is hidden in a directory that is not included by default.
20171005 01:09:34< mattsc> I’m not able to formulate anything that makes sense at the moment.
20171005 01:10:09< celticminstrel> So basically add {ais/dev/idle_ai.cfg} somewhere?
20171005 01:10:52< mattsc> Possibly.  I don’t really know at the moment how you would do it practically.
20171005 01:11:12< mattsc> *in practice?
20171005 01:14:00< vultraz> celticminstrel: what are "structured bindings"?
20171005 01:14:08< celticminstrel> Uhh. Context?
20171005 01:14:50< mattsc> celticminstrel: I’ll be afk for 45 min or so.  I like the idea of a hidden attribute.
20171005 01:15:28< vultraz> ah. nevermind
20171005 01:18:12< vultraz> are named locations supposed to be saved in the map data...?
20171005 01:18:43< celticminstrel> Yes.
20171005 01:19:00< vultraz> well this is causing ERRORS
20171005 01:19:20 * celticminstrel still wants to know what that was about structured bindings though.
20171005 01:19:23< celticminstrel> What sort of errors?
20171005 01:20:20< vultraz> celticminstrel: I was trying to assign `scenario["key"] = foo;` but in this case, `scenario` wasn't a config but the name of the base class. MSVC spit out an error telling me I needed latest c++ to use "structured bindings"
20171005 01:20:47< celticminstrel> Oh, I think structured bindings is probably this thing:
20171005 01:21:10< celticminstrel> auto [iter, inserted] = v.insert(at_iter, value);
20171005 01:21:16< vultraz> i see
20171005 01:21:29< celticminstrel> Not quite sure though.
20171005 01:21:43< vultraz> anyway, the error is an assertion fail in boost
20171005 01:22:05< vultraz> when i try to go to User Scenarios in MP Create if a user scenario has a named location in the map data
20171005 01:22:11< vultraz> " Gg, Gg, TEST LOC Gg, Gg, Gg,"
20171005 01:22:13< vultraz> for example
20171005 01:22:37< celticminstrel> As I understand it, that should be parsed as two named locations TEST and LOC, rather than one named "TEST LOC".
20171005 01:22:59< vultraz> oh dear
20171005 01:23:11< vultraz> then this is a more serious issue than I expected
20171005 01:23:15< celticminstrel> What's the assertion failure? Something about a value not existing?
20171005 01:23:57< vultraz> something in a "dynamic_set.hpp:640" for "m_check_invariants()"
20171005 01:24:04< celticminstrel> Ugh.
20171005 01:24:32< celticminstrel> What does the m_check_invariants() function check?
20171005 01:25:11< vultraz> I don't know
20171005 01:26:01< celticminstrel> ...open the file and look? You can get there entirely within MSVC (assuming that's what you're using).
20171005 01:26:33< vultraz> yes
20171005 01:26:47-!- travis-ci [~travis-ci@ec2-54-159-187-121.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 01:26:48< travis-ci> wesnoth/wesnoth#15169 (master - f097c3d : mattsc): The build failed.
20171005 01:26:48< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/283448797
20171005 01:26:48-!- travis-ci [~travis-ci@ec2-54-159-187-121.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 01:27:23-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20171005 01:29:38< vultraz> but i'd need to make a debug build
20171005 01:48:21< mattsc> Looking at the travis log, I don’t think I caused that error.
20171005 02:02:05-!- mattsc_ [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20171005 02:02:07-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Ping timeout: 260 seconds]
20171005 02:02:08-!- mattsc_ is now known as mattsc
20171005 02:14:28-!- travis-ci [~travis-ci@ec2-54-221-20-177.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 02:14:29< travis-ci> wesnoth/wesnoth#15171 (master - 1df2985 : Ignacio R. Morelle): The build passed.
20171005 02:14:29< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/283461302
20171005 02:14:29-!- travis-ci [~travis-ci@ec2-54-221-20-177.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 02:19:54< vultraz> 20171005 13:19:34 error ai/config: skipped AI config due to duplicate id [idle_ai]. Config contains:
20171005 02:19:54< vultraz> description = Dev AI: Idle AI
20171005 02:19:54< vultraz> id = idle_ai
20171005 02:19:54< vultraz> [stage]
20171005 02:19:54< vultraz>         name = empty
20171005 02:19:55< vultraz> [/stage]
20171005 02:24:03< vultraz> celticminstrel: ^
20171005 02:26:09< mattsc> Hmm, that appears to be because of my change, because that AI is now in both ai/ais/ and ai/dev.,  It should only happen if you start the game in debug mode.
20171005 02:27:20< mattsc> Fix one bug, introduce two problems.  Damned if you do, …  I should just retire for real.  :P
20171005 02:33:45< celticminstrel> ...why would you need a debug build? Just right-click the include line and select "Open document", or right-click a function call and select "Go to definition".
20171005 02:34:50< vultraz> i dunno
20171005 02:34:54< vultraz> got some weird screen 
20171005 02:35:43< celticminstrel> ?!?!?
20171005 02:42:37< vultraz> "ucrtbase.pbd not loaded"
20171005 02:43:59< celticminstrel> That doesn't matter, it just means you don't have debugging symbols for something.
20171005 02:44:23< celticminstrel> It has nothing to do with Intellisense's ability to find the definition of a function.
20171005 02:46:12-!- sigurdfd [~SigurdFD@dynamic-acs-72-23-110-196.zoominternet.net] has joined #wesnoth-dev
20171005 02:46:53< vultraz> interestingly enough, the TEST LOC thing works fine in the context of just a user map
20171005 03:00:28-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20171005 03:00:34-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171005 03:17:42-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has joined #wesnoth-dev
20171005 03:21:14< irker163> wesnoth: Charles Dang wesnoth:master d957326828c7 / src/gui/widgets/window.hpp: GUI2/Window: fixed behavior regression from 2416f10ddd https://github.com/wesnoth/wesnoth/commit/d957326828c7b037a15c9b057e6862566367499f
20171005 03:21:17< irker163> wesnoth: Charles Dang wesnoth:master b8fe9fe7dc29 / src/ (3 files in 2 dirs): MP Create Game: don't allow proceeding if selected game has no sides https://github.com/wesnoth/wesnoth/commit/b8fe9fe7dc2985a5c38750335a014641673ed692
20171005 03:25:07< irker163> wesnoth: Tobias Frei wesnoth:master 5d18b5b01c8d / copyright: lol https://github.com/wesnoth/wesnoth/commit/5d18b5b01c8d3a9196d155f1a0df32942ddb8b6b
20171005 03:28:40< vultraz> there's still a weird issue, though
20171005 03:30:18< vultraz> if i remove the [side] tags from an mp scenario, it will still star
20171005 03:30:19< vultraz> t
20171005 03:30:28< vultraz> expected, since the game fills in [side] tags to match the map
20171005 03:30:48< vultraz> but there are no [side] tags once you get to Connect
20171005 03:30:50< vultraz> er
20171005 03:30:52< vultraz> yeah
20171005 03:31:05< vultraz> however, the same behavior works with a user map with no configuration 
20171005 03:31:06< vultraz> :|
20171005 03:36:13< shadowm> celticminstrel: Is there any chance c9191c98122d97310eddae041a9fd3496cd5b894 could have affected campaignd's passphrase hashing?
20171005 03:36:46< shadowm> Namely, by making the same input produce different hashes than before with the same salt?
20171005 03:37:25< shadowm> (See https://github.com/wesnoth/wesnoth/issues/2068 )
20171005 03:38:26< shadowm> I have a feeling that the passphrase problems (which a lot of people have reported) started when I deployed a new build with my unrelated changes.
20171005 03:40:00< shadowm> The previous build was from May 31 though, hm.
20171005 03:40:27< shadowm> And the commit was pushed *before* the build.
20171005 03:41:15< shadowm> Oh crap I'm looking at *trunk*.
20171005 03:41:36< shadowm> Okay, well, the previous 1.13 build was also from May 31st.
20171005 03:43:46< shadowm> So whatever went wrong with campaignd's passphrase hashing took place between May 31st (previous server build and run) and September 15th (when I rebuilt and restarted the server to deploy my botched file commit patch that again *can't possibly have corrupted the passphrases*).
20171005 04:07:48-!- travis-ci [~travis-ci@ec2-54-221-20-177.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 04:07:49< travis-ci> wesnoth/wesnoth#15173 (master - b8fe9fe : Charles Dang): The build has errored.
20171005 04:07:49< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/283503279
20171005 04:07:49-!- travis-ci [~travis-ci@ec2-54-221-20-177.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 04:28:26-!- sigurdfd [~SigurdFD@dynamic-acs-72-23-110-196.zoominternet.net] has quit []
20171005 05:17:27-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The computer falls into a deep sleep, to awake again some other day!]
20171005 05:19:23-!- travis-ci [~travis-ci@ec2-54-221-20-177.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 05:19:24< travis-ci> wesnoth/wesnoth#15174 (master - 5d18b5b : Tobias Frei): The build was broken.
20171005 05:19:24< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/283504020
20171005 05:19:24-!- travis-ci [~travis-ci@ec2-54-221-20-177.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 05:24:04-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has quit [Quit: .]
20171005 06:16:56-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has joined #wesnoth-dev
20171005 06:17:29-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20171005 06:20:33< zookeeper> mattsc, ohh, awesome.
20171005 06:22:24< irker163> wesnoth: Charles Dang wesnoth:master 2caf26641364 / src/game_initialization/ (configure_engine.cpp configure_engine.hpp): Configure Engine: formatting cleanup https://github.com/wesnoth/wesnoth/commit/2caf266413642a2732e4a2c6d738d709ea29fd91
20171005 06:22:27< irker163> wesnoth: Charles Dang wesnoth:master 2d3a79452b82 / src/game_initialization/configure_engine.hpp: Configure Engine: removed orphaned function declaration from 2df3a09c0c https://github.com/wesnoth/wesnoth/commit/2d3a79452b82d6ea0a833dd570df3a1305e4ef6a
20171005 06:22:30< irker163> wesnoth: Charles Dang wesnoth:master 2535fb0ff552 / src/game_initialization/ (connect_engine.cpp create_engine.cpp): Game Initialization: made use of emplace_back/std::move in a couple places https://github.com/wesnoth/wesnoth/commit/2535fb0ff5528283d90fcf173393aa988f534a7c
20171005 06:29:08< zookeeper> mattsc, so, unhiding an AI means making it available in the MP side controller assignment menu, right?
20171005 07:34:14-!- travis-ci [~travis-ci@ec2-54-159-187-121.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 07:34:15< travis-ci> wesnoth/wesnoth#15175 (master - 2535fb0 : Charles Dang): The build was fixed.
20171005 07:34:15< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/283542314
20171005 07:34:15-!- travis-ci [~travis-ci@ec2-54-159-187-121.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 08:13:22< irker163> wesnoth: Charles Dang wesnoth:master e3db7cc22561 / src/ (gui/dialogs/multiplayer/mp_options_helper.cpp saved_game.cpp): Save custom options data in a more concise way https://github.com/wesnoth/wesnoth/commit/e3db7cc225618ab013ad25d94909ff50f04fb67b
20171005 09:19:07-!- vn971 [~vasya@94.158.103.15] has joined #wesnoth-dev
20171005 09:25:01< AI0867> shadowm: I only noticed that yesterday. I'll be happy to fix the conflicts for you
20171005 09:37:45< vultraz> AI0867: i've merged his PR. now you need to fix conflicts for your PR :P
20171005 09:45:05-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171005 09:48:07-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Client Quit]
20171005 09:56:54< AI0867> that's fine with me
20171005 10:12:00-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has quit [Quit: Going offline, see ya! (www.adiirc.com)]
20171005 10:18:07-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has joined #wesnoth-dev
20171005 10:35:20< vn971> I wonder, can I set room topic for my room? I would like to.
20171005 10:35:57< vn971> I mean the game creation room.
20171005 10:36:55< vn971> For example, wesnoth-1.12 doesn't have sound notifications for people joining room. I'd like to set some explanation text in /topic that I'll reach back if I was AFK, telling not to wait too long.
20171005 10:42:35< Necrosporus> vn971, what do you think about https://github.com/wesnoth/wesnoth/pull/2070
20171005 10:48:43< vn971> Necrosporus: code wise, I won't be able to give any kind of estimate, unfortunately (I don't know cpp). In terms of the issue itself, I like the idea of a white list (as an alternative to black list).
20171005 10:59:57< Necrosporus> vn971, what is your opinion on !$& (') [,] which appear controversial?
20171005 11:10:14< Necrosporus> $[,] conflict with WML
20171005 11:14:19-!- irker163 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20171005 11:19:14-!- TadCarlucci [~lundberg@74.193.219.119] has joined #wesnoth-dev
20171005 11:19:21-!- TadCarlucci [~lundberg@74.193.219.119] has quit [Client Quit]
20171005 11:37:52< vn971> Necrosporus: I don't have any strict opinion on that specific issue. (As I understand it, [] are already used in addons, so the issue is non-trivial.)
20171005 11:54:12< AI0867> I don't see why we can't simply stop authors from uploading add-ons with those characters, and leaving the existing ones
20171005 11:54:55< JyrkiVesterinen> The main problem is UMC authors who are waiting for 1.14 stable before porting their add-ons to it.
20171005 11:55:26< AI0867> that makes it even easier
20171005 11:55:39< JyrkiVesterinen> Sure, they will need to spend effort porting their add-ons, but that's not a valid argument to give them an even larger workload.
20171005 11:55:49< AI0867> the process of porting the add-on will now include removing illegal characters from filenames
20171005 11:56:00< Necrosporus> changing one or two lines isn't really that hard, IMO
20171005 11:56:15< JyrkiVesterinen> Now, we can deprecate/remove things in WML/Lua API/whatever if there is a big enough payoff.
20171005 11:56:52< JyrkiVesterinen> But, IMHO, the advantages of restricting allowed characters are nowhere near big enough to justify giving any extra work to UMC authors.
20171005 11:57:20< AI0867> disallowing [] would break 1 add-on on 1.12, according to the analysis
20171005 11:57:26< Necrosporus> JyrkiVesterinen, I think advantage of removing $[,] is sufficiently big
20171005 11:57:58< JyrkiVesterinen> It was threeadd-ons according to this comment: https://github.com/wesnoth/wesnoth/pull/2070#issuecomment-334347237
20171005 11:58:15< Necrosporus> It depends on what you call breaking
20171005 11:58:44< Necrosporus> if addon has one filename not referenced anywhere, and it will have to be removed, does it count as breaking?
20171005 11:59:11< JyrkiVesterinen> Yes, because it blocks the author from updating the add-on until he/she has removed the file.
20171005 11:59:24< JyrkiVesterinen> Even if removing it is the only thing he/she needs to do.
20171005 12:01:55< Necrosporus> What do you think about add-on which has copy of Ruffian.cfg which is probably copied from older version of wesnoth (has insignificant differences), and id Rufian (with one f)
20171005 12:02:44< JyrkiVesterinen> Is that related to the discussion about allowed characters?
20171005 12:03:10< Necrosporus> fixing this would involve removing cfg file and replacing all occurences of Rufian to Ruffian. And yes. It's one which has breaking case conflict
20171005 12:03:56< Necrosporus> Will be disallowed with another PR
20171005 12:04:48< JyrkiVesterinen> So, I assume that the add-on is also broken for players (i.e. simply crashes on load or something)?
20171005 12:04:59< Necrosporus> On linux it seems ok
20171005 12:05:21< Necrosporus> though the addon says "WIP" in description
20171005 12:05:30< Necrosporus> This one has other weird chars in filenames
20171005 12:05:37-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20171005 12:05:41< JyrkiVesterinen> Well, in that case I'd prefer not to block it.
20171005 12:06:17< JyrkiVesterinen> What exactly does the case conflict check do, anyway? Does it check that the add-on doesn't have, in the same directory, two filenames that differ only in case?
20171005 12:06:26< Necrosporus> Yes
20171005 12:06:53< JyrkiVesterinen> Such a check sounds good to me. It results in silently overwriting files on Windows.
20171005 12:07:03< JyrkiVesterinen> Even if the addon is currently allowed.
20171005 12:07:22-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Client Quit]
20171005 12:08:19< Necrosporus> OK, PC_Campaign addon, named The Strange Men (has ! problem too), there are three copies of same file [Vv]ector_abilities.cfg
20171005 12:08:27< Necrosporus> two of them are in same directory
20171005 12:08:33< JyrkiVesterinen> Ouch. :(
20171005 12:09:05< Necrosporus> Installing it on windows would be harmless, since files are same anyway
20171005 12:09:28< JyrkiVesterinen> All right, I'm willing not to count it as an addon that has the ! problem (since the case sensitivity check would block it anyway).
20171005 12:10:00< JyrkiVesterinen> That leaves "Defence of Gaer Derole" as the only addon that uses ! and we need to care about.
20171005 12:10:15< Necrosporus> It also has $ problem
20171005 12:10:16< JyrkiVesterinen> And just one addon is so low that I'm fine with blackilisting ! in filenames.
20171005 12:10:20< Necrosporus> it's only one with $
20171005 12:10:29< JyrkiVesterinen> (Same for $.)
20171005 12:13:02< Necrosporus> Altered and Altered New Sprites same to come from same author and are duplicates probably
20171005 12:13:35< Necrosporus> I didn't do detailed analysis, if one could be fully deprecated by another
20171005 12:13:45< Necrosporus> They are only ones with comma problem
20171005 12:14:00< Necrosporus> They have problems with other special chars as well
20171005 12:16:27< Necrosporus> So () [] , originate mostly from them
20171005 12:22:32< Necrosporus> JyrkiVesterinen, another point: are we going to also disallow uploading addon's which have broken picture?
20171005 12:23:02< JyrkiVesterinen> Don't ask me about it. AI is the one who has been implementing restrictions.
20171005 12:23:15< JyrkiVesterinen> And greally, it's a general question and you shouldn't ping anyone about it.
20171005 12:23:15< Necrosporus> Do you agree with it?
20171005 12:23:35< JyrkiVesterinen> IMO, image path validation is a good idea.
20171005 12:24:01< Necrosporus> Most of add-ons with weird filenames also have invalid image
20171005 12:24:24< Necrosporus> I can try to find ones which has weird filenames but good picture, perhaps it would be only one or two
20171005 12:26:38< JyrkiVesterinen> I think invalid images and abnormal filenames are distinct enough problems that I think an addon should be counted for filenames even if it has broken image references.
20171005 12:34:04< AI0867> JyrkiVesterinen: I added one more commit to #2074
20171005 12:34:45< JyrkiVesterinen> Yeah, I have seen it.
20171005 12:34:56< AI0867> Elvish_Hunter: any futher opinions on #2053?
20171005 12:34:57< JyrkiVesterinen> Looks fine to me. You're still carrying my approval.
20171005 12:35:16-!- irker945 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20171005 12:35:16< irker945> wesnoth: Alexander van Gessel wesnoth:master 7b3254964dbf / utils/travis/mp_test_executor.sh: Ensure that mp_test_executor exits https://github.com/wesnoth/wesnoth/commit/7b3254964dbf450fb0eb342a69849e7221201687
20171005 12:35:16< irker945> wesnoth: Alexander van Gessel wesnoth:master 2e5c5da0036b / utils/travis/mp_test_executor.sh: MP test: treat failed init of SDL_video as a 'successful' test https://github.com/wesnoth/wesnoth/commit/2e5c5da0036be2e68b42e70f3153b799236b5b69
20171005 12:35:17< irker945> wesnoth: Alexander van Gessel wesnoth:master fea609d30920 / utils/travis/mp_test_executor.sh: Add missing break https://github.com/wesnoth/wesnoth/commit/fea609d30920e9de4c029cf43b2fd7c377f6263a
20171005 12:35:18< irker945> wesnoth: Alexander van Gessel wesnoth:master fc0674bcb653 / utils/travis/mp_test_executor.sh: Change *_RUNNING vars to string comparisons https://github.com/wesnoth/wesnoth/commit/fc0674bcb653f9fbfbaa64e22b20639f1c2c9057
20171005 12:35:19< irker945> wesnoth: Alexander van Gessel wesnoth:master 87bda82c2859 / utils/travis/mp_test_executor.sh: Merge pull request #2074 from AI0867/mp-test-fixes https://github.com/wesnoth/wesnoth/commit/87bda82c285914a7a7744b26bd2bbdd724c6f56a
20171005 12:51:13-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Ping timeout: 248 seconds]
20171005 12:52:02-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171005 12:54:02-!- sevu [~Shiki@p5485609A.dip0.t-ipconnect.de] has joined #wesnoth-dev
20171005 12:54:50-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171005 13:09:15-!- PaulLapierre[m] [paullapier@gateway/shell/matrix.org/x-rtmeycehnuyzjplp] has joined #wesnoth-dev
20171005 13:11:58-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20171005 13:13:27 * PaulLapierre[m] sent a long message: PaulLapierre[m]_2017-10-05_13:13:26.txt <https://matrix.org/_matrix/media/v1/download/matrix.org/nGomqpdzGcferyKrVfQfkrWD>
20171005 13:15:19< Necrosporus> Probably $side.gold
20171005 13:15:27< Necrosporus> What is the context of your code?
20171005 13:16:06< Necrosporus> If you want to give some gold there could be better ways to do so
20171005 13:16:31< PaulLapierre[m]> i want to show the amount of gold the player have
20171005 13:17:18< Necrosporus> You need to store side into variable, then use $foo.gold where foo is your variable name
20171005 13:17:39< Necrosporus> Perhaps it could be done in simpler way, but mainline code stores side first
20171005 13:18:13< shadowm> AI0867: The 1.14 add-ons server will be a new instance FYI, so anything that would affect existing content is an issue.
20171005 13:18:31< shadowm> AI0867: The same goes for 1.15, 1.16, 1.17...
20171005 13:19:54 * PaulLapierre[m] sent a long message: PaulLapierre[m]_2017-10-05_13:19:53.txt <https://matrix.org/_matrix/media/v1/download/matrix.org/eATlLVpRalBkMcXeIqcGXnJU>
20171005 13:20:09< shadowm> I hope you realize that most people won't click random links.
20171005 13:20:38< AI0867> actually, I think that's a message longer than a single IRC line, transformed this way by his gateway
20171005 13:20:41< shadowm> AI0867: I honestly find that you're going overboard with this.
20171005 13:20:43< AI0867> he's not using an IRC client
20171005 13:20:55< shadowm> Most people do not know that either.
20171005 13:21:01< Necrosporus> PaulLapierre[m], you need [store_side] tag
20171005 13:21:30< PaulLapierre[m]> i can do that into a Option tag ?
20171005 13:21:52-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20171005 13:22:15< shadowm> The whitespace restriction was already questionable and resulted in a few rather poorly-worded complaints, so I'm not looking forward to dealing with whoever next decides that they need to have a non-ASCII filename in their add-on.
20171005 13:22:27< shadowm> *questionable enough
20171005 13:22:30< AI0867> shadowm: you're talking about the filenames thing? We could make this thing optional for existing instances, but it only affects new uploads as it is
20171005 13:22:46< shadowm> (Because you are not the one who's running the server or handling support requests.)
20171005 13:23:21< shadowm> Did you even read what I said about 1.14?
20171005 13:23:56< Necrosporus> And what that means, stuff from 1.13 will need to be reaploaded?
20171005 13:24:01< JyrkiVesterinen> "Only affecting new uploads" isn't a very important point. It means more effort for a UMC author who wants to update theis addon.
20171005 13:24:08< JyrkiVesterinen> *their
20171005 13:24:28< JyrkiVesterinen> Necrosporus: Exactly.
20171005 13:26:43< AI0867> 15:23 < shadowm> Did you even read what I said about 1.14? ← I interpreted it differently, as in, that there would not be content violating these rules on it, rather than that everything would have to be reuploaded
20171005 13:27:01< shadowm> That's a very convenient interpretation.
20171005 13:27:58-!- TC02 [~quassel@london.acm.jhu.edu] has quit [Ping timeout: 264 seconds]
20171005 13:28:26< Necrosporus> If non-ascii is allowed, then probably some more or less complex checks will be needed, to ban control chars, unmatched surrogate pairs and...
20171005 13:28:58< JyrkiVesterinen> It isn't important to disallow Unicode control characters IMO.
20171005 13:29:02< AI0867> filenames are stored as UTF-8 in wesnoth, so it would just need to check whether the UTF-8 is valid in that regard
20171005 13:29:11< JyrkiVesterinen> Unmatched surrogate pairs aren't possible in UTF-8.
20171005 13:29:33< AI0867> and that the filename is at most 255 characters long
20171005 13:29:35< Necrosporus> Surrogate chars could be encoded as UTF-8 btw
20171005 13:29:46< AI0867> yes, but they're not valid
20171005 13:30:05< Necrosporus> Then more complex check will be needed
20171005 13:30:15< JyrkiVesterinen> And use of supplementary Unicode planes is easy to detect, as only in that case there will be bytes that match the bit pattern 0b11110xxx.
20171005 13:30:17< JyrkiVesterinen> See https://en.wikipedia.org/wiki/UTF-8
20171005 13:31:58< AI0867> implementation details are implementation details. The main question right now is if we want to keep the blacklist system with some extra checks, or if we want to switch to whitelist
20171005 13:32:21< JyrkiVesterinen> Necrosporus: no, a more complex check isn't needed. All characters which end up as surrogate pairs in UTF-16 are in supplementary planes, and the check I told above is everything it takes to detect them.
20171005 13:32:47< JyrkiVesterinen> I have already told my preference: keep the blacklist.
20171005 13:33:36< Necrosporus> Wait
20171005 13:33:59< Necrosporus> Halves of surrogate pairs are in BMP by themself
20171005 13:34:07< Necrosporus> They can be encoded into UTF-8
20171005 13:35:48< JyrkiVesterinen> Uh, that's not the case as far as I know.
20171005 13:35:50< JyrkiVesterinen> "Since the ranges for the high surrogates, low surrogates, and valid BMP characters are disjoint, it is not possible for a surrogate to match a BMP character"
20171005 13:38:46< Necrosporus> What about security problem with filenames https://img.raymond.cc/blog/wp-content/uploads/2015/03/rtlo-spoof-extension.png
20171005 13:38:57< JyrkiVesterinen> Okay. A bit more research, and I found this: "Since RFC 3629 (November 2003), the high and low surrogate halves used by UTF-16 (U+D800 through U+DFFF) and code points not encodable by UTF-16 (those after U+10FFFF) are not legal Unicode values, and their UTF-8 encoding must be treated as an invalid byte sequence."
20171005 13:39:12< JyrkiVesterinen> So, the "is valid UTF-8" check takes care of surrogates.
20171005 13:39:36< Necrosporus> depends on how is it done
20171005 13:39:44< Necrosporus> Maybe restrict to 2-byte UTF-8?
20171005 13:40:18< Necrosporus> Which encodes anything up to U+8000
20171005 13:40:23< AI0867> no, that excludes half the BMP, but again, these are implementation details
20171005 13:40:33< JyrkiVesterinen> I'd like to allow 3-byte. It contains CJK characters, and it's plausible that UMC authors would want to use them in filenames.
20171005 13:40:53< Necrosporus> What about players?
20171005 13:41:10< Necrosporus> I as a player would not like to see files with unreadable name (depending on font)
20171005 13:42:02< JyrkiVesterinen> You're not supposed to see them, you know. Players aren't supposed to browse add-on directories.
20171005 13:42:10< Necrosporus> They are also impossible to type normally
20171005 13:42:23< Necrosporus> So shell experience is degraded
20171005 13:42:24< AI0867> add-on filenames are not part of the UI
20171005 13:42:29< Necrosporus> I mostly browser files from shell
20171005 13:42:49< Necrosporus> JyrkiVesterinen, I might want to copy some files into my UMC
20171005 13:43:10< JyrkiVesterinen> That's a developer use case rather than a player use case.
20171005 13:43:13< Necrosporus> And I always browse filenames to read code of scenarios I have played after I finish them
20171005 13:43:22< JyrkiVesterinen> Not important enough to disallow characters, IMHO.
20171005 13:44:26< Necrosporus> By the way, allowing unicode creates sort of problem similar to one which you noticed in Windows
20171005 13:44:30-!- fabi [~fabi@wesnoth/developer/fendrin] has quit [Remote host closed the connection]
20171005 13:44:56< Necrosporus>  chars such as é could be represented in two ways
20171005 13:45:01< Necrosporus> One way is single codepoint
20171005 13:45:23< Necrosporus> Other way is two codepoints, one for e and other for combining accent
20171005 13:45:46< Necrosporus> Some filesystems allow both to exist in same place, for example linux filesystem
20171005 13:46:01< Necrosporus> Mac and Windows tries to normalize filenames so they would conflict
20171005 13:46:52< JyrkiVesterinen> That's an unlikely problem to occur in practice.
20171005 13:47:05< AI0867> it's pretty much equivalent to case conflicts
20171005 13:47:13< JyrkiVesterinen> Developers being *allowed* to use Unicode in file names doesn't mean that they will *necessarily* do it.
20171005 13:47:19< AI0867> they already do
20171005 13:47:35< JyrkiVesterinen> Not many, according to Necrosporus's own analysis.
20171005 13:47:39< AI0867> see the cyrillic /[original]/
20171005 13:48:12< JyrkiVesterinen> And that case, for example, doesn't have any problems. As Necro noted, the file isn't referenced anywhere.
20171005 13:50:58< JyrkiVesterinen> And indeed, a conflict would require that an addon would have the same filename encoded in both ways. Who in their right mind would do that?!
20171005 13:51:32< Necrosporus> Same can be said about case conflicts
20171005 13:51:46< Necrosporus> Who in right mind will include three copies of same macro file in their addon?
20171005 13:52:19< JyrkiVesterinen> Why would they necessarily be the same?
20171005 13:52:38< Necrosporus> In this case they all were same. In other addon they were different
20171005 13:52:38< AI0867> they were, in this case
20171005 13:53:19< JyrkiVesterinen> For example, you could have some sort of fantasy world, where there are two characters named "ufns" and "UFNS", differing only in case.
20171005 13:53:38< JyrkiVesterinen> (Hey, this is a fantasy world, where people can have strange ways to name their children.)
20171005 13:53:54< JyrkiVesterinen> And then just imagine that an artist has drawn a sprite for both.
20171005 13:53:56< JyrkiVesterinen> Boom.
20171005 13:54:15< Necrosporus> So maybe allow case-different addons?
20171005 13:54:30< Necrosporus> And use some hooks to workaround windows and mac problems
20171005 13:54:48< JyrkiVesterinen> Whereas "name contains accented characters and is encoded in both ways" is something I can only imagine happening if someone is trying to exploit Wesnoth somehow.
20171005 13:55:29< Necrosporus> Windows supports case-sensitive filesystems
20171005 13:55:34< JyrkiVesterinen> Allowing case-different addons would be difficult. It's easier to just ban them.
20171005 13:55:37< Necrosporus> It just has to be access differently
20171005 13:55:57< JyrkiVesterinen> To be exact, there is a registry key to set Windows to case-sensitive mode.
20171005 13:56:01< JyrkiVesterinen> That key is global.
20171005 13:56:08< Necrosporus> Wesnoth can set it %)
20171005 13:56:18< JyrkiVesterinen> No one uses it because too many applications break when it's enabled.
20171005 13:56:32< Necrosporus> I think there are other ways
20171005 13:56:38< JyrkiVesterinen> Because obviously barely anyone tests their software in case-sensitive mode.
20171005 13:57:57< Necrosporus> I can imagine files ending up in one place by going by different route
20171005 13:58:14< Necrosporus> Like one copied from Mac and other from Windows onto linux machine
20171005 13:59:11< JyrkiVesterinen> It isn't enough if that happens.
20171005 13:59:29< JyrkiVesterinen> It would also be required that the author somehow doesn't notice that the file was duplicated.
20171005 13:59:47< Necrosporus> How did it happen with other addons?
20171005 13:59:49-!- fabi_ [~fabi@2a02:810c:c840:2e65:cc6c:be08:611d:ad7d] has joined #wesnoth-dev
20171005 13:59:51-!- travis-ci [~travis-ci@ec2-54-145-67-88.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 13:59:52< travis-ci> wesnoth/wesnoth#15178 (master - 87bda82 : Alexander van Gessel): The build has errored.
20171005 13:59:53< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/283671175
20171005 13:59:53-!- travis-ci [~travis-ci@ec2-54-145-67-88.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 14:00:22< Necrosporus> author somehow didn't notice that he had three files with same name and content
20171005 14:00:23< JyrkiVesterinen> Also, it would mean that one file originally had a non-normalized file name. AFAIK, that's not possible on any operating system unless you explicitly try it.
20171005 14:00:45< Necrosporus> Mac does normalize into e + ', no?
20171005 14:01:08< JyrkiVesterinen> Because that kind of case duplication has likely happened for other UMC authors, too.
20171005 14:01:18< JyrkiVesterinen> In 99% of cases, they notice it and fix it.
20171005 14:01:19< Necrosporus> There are also characters like Å and µ
20171005 14:01:59< JyrkiVesterinen> But then someone ran into that duplication AND failed to miss it. And thus, add-ons with identical-except-case filenames ended up in the server.
20171005 14:02:16< JyrkiVesterinen> For file name representation, it's extremely unlikely to happen.
20171005 14:03:01< JyrkiVesterinen> It requires a file with a name with multiple representations to begin with. Your search found three such files across all Wesnoth addons.
20171005 14:03:20< JyrkiVesterinen> Then it requires that the author has somehow input the file name in non-normalized form.
20171005 14:03:35< JyrkiVesterinen> Then the author needs to copy the file to Windows/Mac and back.
20171005 14:03:46< JyrkiVesterinen> And finally he needs to miss that the name is now duplicated.
20171005 14:03:50< JyrkiVesterinen> *file
20171005 14:04:41< JyrkiVesterinen> Such a possibility, that's only barely possible, is nowhere near strong enough reason to block UMC authors from using Unicode in filenames at all.
20171005 14:04:57< Necrosporus> So what we have is numberous potential problems, each of which seem particularly small separately
20171005 14:05:31< JyrkiVesterinen> Every problem you have pointed out is so small that I simply don't find them worth caring about.
20171005 14:05:39< JyrkiVesterinen> Same is true for all of them combined.
20171005 14:06:07< Necrosporus> Well, regarding FAT problems I run into them now and then
20171005 14:06:40< Necrosporus> I use ext4 on my USB drives mostly, but then they are unreadable by TV...
20171005 14:06:59< JyrkiVesterinen> Have you mounted the file system with right character encoding?
20171005 14:07:29< JyrkiVesterinen> Linux assumes that FAT32 is UTF-8 encoded, but Windows defaults to local encoding when formatting (ISO 8859-15 here).
20171005 14:07:41< Necrosporus> I know about that, but sometimes I forget to write -o iocharset=utf8,codepage=... stuff when mounting them
20171005 14:07:50< JyrkiVesterinen> I needed to override the encoding in /etc/fstab to get it to work.
20171005 14:08:07< Necrosporus> Really?
20171005 14:08:14< JyrkiVesterinen> Hm. I see.
20171005 14:08:16< Necrosporus> I thought linux defaults to cp437
20171005 14:08:43< Necrosporus> Or anything which is hardcoded when kernel is compiled
20171005 14:08:50< JyrkiVesterinen> At least here I recall it assuming UTF-8.
20171005 14:09:10< JyrkiVesterinen> Might be that it's a kernel build option, and our kernels use a different default.
20171005 14:09:12< Necrosporus> You can download vanilla kernel and check
20171005 14:09:19< Necrosporus> Yes, it's build option
20171005 14:09:29< JyrkiVesterinen> I think I won't bother. Too much effort.
20171005 14:09:47< Necrosporus> Settable in make menuconfig
20171005 14:09:55-!- fabi_ [~fabi@2a02:810c:c840:2e65:cc6c:be08:611d:ad7d] has quit [Quit: Konversation terminated!]
20171005 14:10:41< Necrosporus> Anyway it means that wesnoth add-ons with filenames from extended latin cannot be transfered on FAT USB-sticks between windows which uses Cyrillic script
20171005 14:10:48< Necrosporus> and vice versa
20171005 14:11:12< JyrkiVesterinen> Yeah, that's true.
20171005 14:11:22< JyrkiVesterinen> Still, I consider it a minor problem.
20171005 14:11:31-!- fabi_ [~fabi@2a02:810c:c840:2e65:cc6c:be08:611d:ad7d] has joined #wesnoth-dev
20171005 14:11:46< JyrkiVesterinen> One doesn't exactly need to transfer add-ons in thumb drives very often.
20171005 14:11:51< Necrosporus> Necessety to rename some files before uploading is minor too
20171005 14:12:28< Necrosporus> What if you play on Android and want to keep add-ons on internal flash?
20171005 14:12:35< Necrosporus> It is fat-formatted 
20171005 14:12:41< Necrosporus> or sd-card
20171005 14:12:59< JyrkiVesterinen> Well, backwards compatibility is something I don't want to sacrifice because of minor (almost theoretical) problems.
20171005 14:13:40< Necrosporus> Оригиналы also has problems with [], which I strongly suggest to diallow because WML
20171005 14:14:49< shadowm> Not. All. Files. Are. WML.
20171005 14:16:08< Necrosporus> By the way, will wesnoth load image files with [] in name at all?
20171005 14:16:46< shadowm> Not all images are used in animations.
20171005 14:17:09< shadowm> Not all files are WML.
20171005 14:17:14< shadowm> Not all WML is event code.
20171005 14:17:16-!- fabi_ [~fabi@2a02:810c:c840:2e65:cc6c:be08:611d:ad7d] has quit [Quit: Konversation terminated!]
20171005 14:17:28< shadowm> Not all code is WML.
20171005 14:17:32< JyrkiVesterinen> To be exact, Android uses a custom filesystem. It was FAT32 until Android 4.4, but after that it has been something else.
20171005 14:17:37< shadowm> Not all files have to be used by Wesnoth.
20171005 14:17:50-!- fabi_ [~fabi@2a02:810c:c840:2e65:cc6c:be08:611d:ad7d] has joined #wesnoth-dev
20171005 14:18:10< JyrkiVesterinen> Android 8.0 introduced a driver called SDCardFS to speed up that FS.
20171005 14:18:17< JyrkiVesterinen> https://arstechnica.com/gadgets/2017/09/android-8-0-oreo-thoroughly-reviewed/8/#h7
20171005 14:19:14< Necrosporus> Android also screwed up USB storage mode :(
20171005 14:20:27< Necrosporus> Android 4.x is still widely used
20171005 14:20:40< Necrosporus> even 2.x is still found on devices which you can find in store
20171005 14:20:47-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has quit [Quit: Going home]
20171005 14:22:02< Necrosporus> I can easily imagine people who want to bring Wesnoth with them on USB drive
20171005 14:22:06< Necrosporus> including add-on
20171005 14:31:38< AI0867> 16:19 < Necrosporus> Android also screwed up USB storage mode :( ← to be fair, offering a drive as "mass storage", means offering exclusive access
20171005 14:32:01< AI0867> which is why android moved away from doing that, to MTP and such
20171005 14:32:31< AI0867> the problem was that you had to unmount the SD card so the attached computer could use it
20171005 14:33:03< AI0867> which is impossible if you don't have an SD card, just emulated storage
20171005 14:33:17< Necrosporus> It did work somehow
20171005 14:33:21< AI0867> and all apps that want to use the "external storage" will fail as long as it's unmounted
20171005 14:33:24< AI0867> not really
20171005 14:33:49< Necrosporus> It does work for my phone
20171005 14:34:04< Necrosporus> OK, kill those apps
20171005 14:34:16< AI0867> it was partially functional, which is why it was even on there, but it was broken from the start
20171005 14:37:56-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20171005 14:42:25< sevu> I had for a few years all my important data on a FAT partition, and used it for windows ans linux. I screwed up all umlaute, thought the reason was that I didn't knew that I have to specify special options in the fstab tho have linux treat it normally. I think I installed the windows wesnoth on the FAT partition, because this allowed to store the userdata for windows in the fat too, and on linux I used a symlink to point to the fat. But 
20171005 14:42:25< sevu> the place to do sth. about is was on the linux side, not on the wesnoth side.
20171005 14:56:30-!- fabi_ [~fabi@2a02:810c:c840:2e65:cc6c:be08:611d:ad7d] has quit [Quit: Konversation terminated!]
20171005 14:57:02-!- fabi_ [~fabi@2a02:810c:c840:2e65:318f:e759:31db:2ceb] has joined #wesnoth-dev
20171005 15:00:16-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20171005 15:04:46< mattsc> zookeeper: thanks; it took a while of poking around to find what was actually happening
20171005 15:04:59< mattsc> zookeeper: Yes, unhiding them makes the AI config available to the engine, which fixes the bug, but also means it is now shown in the MP menu.
20171005 15:05:31< mattsc> The options to deal with this would be (AFAICT): 1. hard-coding special treatment for the idle AI, 2. adding a hidden= attribute to the [ai] tag, 3. or not caring that it now shows up in the MP menu.
20171005 15:05:40< mattsc> celmin suggested #1 and #2.  So far I think I like #2 best.
20171005 15:07:05< zookeeper> mattsc, well, #2 seems like the most obvious choice
20171005 15:07:38< Necrosporus> I am OK with 3. What's wrong with being able to select Idle AI?
20171005 15:08:02< Necrosporus> maybe you want an enemy which doesn't fight back
20171005 15:08:14< Necrosporus> Or it's same as just idle?
20171005 15:09:19< mattsc> zookeeper: yeah
20171005 15:09:31< sevu> you can already set to have an empty side. It's the same like the idle AI, just that the Idle AI has a leader. So, it's a duplicate option.
20171005 15:09:45< Necrosporus> What if you want a leader?
20171005 15:10:03< Necrosporus> I mean perhaps you want to fight a bunch of units which just stay still
20171005 15:10:23< Necrosporus> Will "Idle" allow it?
20171005 15:11:06< sevu> It's about the MP create screen. You can specify the idle AI in scenario code.
20171005 15:11:10< mattsc> Necrosporus: this falls under “can already be done” and is such an “unusual” case that it does not make sense to specifically provide a menu option for it
20171005 15:11:22< mattsc> ^ as sevu just said
20171005 15:12:48< Necrosporus> In this case a question is about specifically remove menu option for it
20171005 15:13:05< Necrosporus> By the way there are duplicate menu options already
20171005 15:13:16< mattsc> zookeeper: there’s another minor issue with the change I made.  I kept the idle AI also in data/ai/dev for backward compatibility.  But that means that it is loaded twice in debug mode and produces the error/warning message that vultraz showed.
20171005 15:13:35< Necrosporus> For example in local MP, what is difference between assigning side to "Local Player" and to my_mp_name ?
20171005 15:13:57< mattsc> The obvious eay solution is to remove it from dev/ and add a message into the “compatibility breaking” list.
20171005 15:14:06< mattsc> *easy
20171005 15:14:50< mattsc> not sure if that’s the best solution though
20171005 15:15:25< zookeeper> so, uh, what usecases would that break?
20171005 15:15:28< Necrosporus> can it be fixed while leaving it in dev?
20171005 15:15:49< zookeeper> a direct file inclusion that probably no one uses?
20171005 15:16:18< mattsc> If somebody has ‘{ai/dev/idle_ai_cfg}’ in a scenario, rather than [ai]ai_algorithm=idle_ai’.
20171005 15:16:27< zookeeper> right, i'll grep for idle_ai_cfg
20171005 15:16:54< mattsc> okay; I already did so in mainline and replaced the one occurence (in a test scenario), but I don’t know about add-ons
20171005 15:17:37< zookeeper> no matches for idle_ai_cfg in 1.12 nor 1.13 add-on *.cfg's
20171005 15:17:44< Necrosporus> ./data/ai/scenarios/scenario-AI_Arena_small.cfg
20171005 15:17:55< zookeeper> (or did you mean idle_ai.cfg?)
20171005 15:18:05< zookeeper> oh, you did
20171005 15:18:23< mattsc> Oops. yes.  Should have copy/pasted instead of typed it in.
20171005 15:18:24< Necrosporus> data/ai/aliases/idle.cf
20171005 15:18:48< mattsc> Necrosporus: no.  Not any more.
20171005 15:19:20< zookeeper> mattsc, nope, no matches
20171005 15:19:53< mattsc> zookeeper: cool, thanks.  Let’s just get rid of it then.
20171005 15:20:07< zookeeper> seems safe to me
20171005 15:25:13< mattsc> Just doing a few tests, to make sure it actually takes care of the error and does not accidentally introduce the Liberty bug again.
20171005 15:25:37< mattsc> It shouldn’t, but I’ve screwed up simpler things before …  :P
20171005 15:32:49< zookeeper> sadly, so have we all
20171005 15:35:06< irker945> wesnoth: mattsc wesnoth:master 70bd919631e8 / data/ai/dev/idle_ai.cfg: Idle AI: remove from data/ai/dev after all https://github.com/wesnoth/wesnoth/commit/70bd919631e86bafd08716ccc6f42a9e84c9ccf5
20171005 15:35:27< mattsc> vultraz: ^ this takes care of the duplicate AI start-up error
20171005 15:36:47< mattsc> So that leaves adding a hidden= attribute to the [ai] tag.
20171005 15:37:29< mattsc> I could probably figure out how to do that, but I assume somebody like celmin can do so about a million times faster than I.
20171005 15:37:44-!- Oebele [~quassel@143.177.58.202] has joined #wesnoth-dev
20171005 15:38:21< mattsc> … and updating the changelog.  I’ll do that after we’ve taken care of the above.
20171005 16:00:11-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has joined #wesnoth-dev
20171005 16:01:07-!- mattsc [~mattsc@wesnoth/developer/mattsc] has left #wesnoth-dev []
20171005 16:01:38-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20171005 16:02:31< JyrkiVesterinen> Necrosporus, regarding Unicode and FAT32: I recalled that FAT32 allows Unicode filenames, it stores long file names as UCS-2.
20171005 16:03:10< JyrkiVesterinen> I don't actually understand why mounting with wrong character encoding on Linux causes trouble.
20171005 16:04:25< JyrkiVesterinen> It may have something to do with short file names, which use one byte per character and are stored with locale-dependent encoding. But Linux should prefer long file names (which are always UCS-2).
20171005 16:04:33-!- fabi__ [~fabi@2a02:810c:c840:2e65:c516:b584:73d3:243b] has joined #wesnoth-dev
20171005 16:05:10< JyrkiVesterinen> Also, UCS-2 can only represent characters in the BMP, which is in fact the biggest reason why supplementary planes shouldn't be allowed (as I have always advocated).
20171005 16:08:24-!- fabi_ [~fabi@2a02:810c:c840:2e65:318f:e759:31db:2ceb] has quit [Ping timeout: 246 seconds]
20171005 16:35:41< irker945> wesnoth: Paŭlo Ebermann wesnoth:master 02907ebe7543 / data/tools/wesnoth/wmltools3.py: Fix the output by --progress. https://github.com/wesnoth/wesnoth/commit/02907ebe75433177f245c12e0eedcef5e3dc5ba8
20171005 16:35:43-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has joined #wesnoth-dev
20171005 16:44:01-!- DeFender1031 [~DeFender1@46-116-196-40.bb.netvision.net.il] has joined #wesnoth-dev
20171005 17:02:06< AI0867> modern windows systems are more likely to use exFAT, which does have full UTF-16 support, rather than the UCS-2 support of VFAT
20171005 17:02:51< JyrkiVesterinen> I run Windows 10, and I have kept my FAT partition as FAT32 for compatibility.
20171005 17:03:13< JyrkiVesterinen> I rpobably don't need it for anything, but I don't have any compelling reason to upgrade it to ExFAT either.
20171005 17:08:03-!- nore [~ncourant@sas.eleves.ens.fr] has quit [Ping timeout: 240 seconds]
20171005 17:09:26-!- nore [~ncourant@sas.eleves.ens.fr] has joined #wesnoth-dev
20171005 17:10:36< AI0867> original FAT32 is even configurable to use either 8-bit or 16-bit codepages (probably including UCS-2)
20171005 17:11:55-!- travis-ci [~travis-ci@ec2-50-17-115-78.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 17:11:56< travis-ci> AI0867/wesnoth#34 (windows-case-check - 1d3e2d2 : Alexander van Gessel): The build has errored.
20171005 17:11:56< travis-ci> Build details : https://travis-ci.org/AI0867/wesnoth/builds/283745313
20171005 17:11:56-!- travis-ci [~travis-ci@ec2-50-17-115-78.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 17:12:01-!- fabi_ [~fabi@2a02:810c:c840:2e65:69c8:3ea7:3d00:cf0d] has joined #wesnoth-dev
20171005 17:12:46< JyrkiVesterinen> FAT32 was developed after LFN, so it may be related to that.
20171005 17:13:02< JyrkiVesterinen> (LFN support landed in Windoed 95, but it was Win98 that introduced FAT32.)
20171005 17:15:48-!- fabi__ [~fabi@2a02:810c:c840:2e65:c516:b584:73d3:243b] has quit [Ping timeout: 240 seconds]
20171005 17:25:03-!- travis-ci [~travis-ci@ec2-54-198-22-147.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 17:25:04< travis-ci> wesnoth/wesnoth#15180 (master - 70bd919 : mattsc): The build passed.
20171005 17:25:04< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/283757128
20171005 17:25:04-!- travis-ci [~travis-ci@ec2-54-198-22-147.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 17:31:29-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has quit [Quit: [endlevel]]
20171005 17:41:08-!- sevu [~Shiki@p5485609A.dip0.t-ipconnect.de] has quit [Quit: Verlassend]
20171005 17:43:04-!- Appveyor [~Appveyor@74.205.54.20] has joined #wesnoth-dev
20171005 17:43:04< Appveyor> wesnoth - Release Alexander van Gessel 1d3e2d2: Windows case check Failed
20171005 17:43:04< Appveyor> Details: https://ci.appveyor.com/project/AI0867/wesnoth/build/Wesnoth-VS2017-windows-case-check-3
20171005 17:43:08-!- Appveyor [~Appveyor@74.205.54.20] has left #wesnoth-dev []
20171005 17:47:49-!- Appveyor [~Appveyor@74.205.54.20] has joined #wesnoth-dev
20171005 17:47:49< Appveyor> wesnoth - Debug Alexander van Gessel 1d3e2d2: Windows case check Failed
20171005 17:47:49< Appveyor> Details: https://ci.appveyor.com/project/AI0867/wesnoth/build/Wesnoth-VS2017-windows-case-check-3
20171005 17:47:53-!- Appveyor [~Appveyor@74.205.54.20] has left #wesnoth-dev []
20171005 18:19:27-!- fabi__ [~fabi@2a02:810c:c840:2e65:bc3f:d9ab:526f:d6e9] has joined #wesnoth-dev
20171005 18:23:46-!- fabi_ [~fabi@2a02:810c:c840:2e65:69c8:3ea7:3d00:cf0d] has quit [Ping timeout: 264 seconds]
20171005 18:34:56-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20171005 18:36:56-!- travis-ci [~travis-ci@ec2-54-145-67-88.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 18:36:57< travis-ci> wesnoth/wesnoth#15181 (master - 02907eb : Paŭlo Ebermann): The build passed.
20171005 18:36:57< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/283783725
20171005 18:36:57-!- travis-ci [~travis-ci@ec2-54-145-67-88.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 18:39:00< shadowm> JyrkiVesterinen: FAT32 was introduced in Windows 95 OEM Service Release 2, actually.
20171005 18:39:42< JyrkiVesterinen> Well, doesn't change my point.
20171005 18:39:51< shadowm> "OEM Service Release" versions were only licensed to OEMs.
20171005 18:41:38< shadowm> OSR 1 was essentially Windows 95 SP1 with a few additional fixes I believe, but OSR 2 introduced FAT32 support (in both real and protected mode). OSR 2.1 introduced USB support and compatibility with an early version of the Windows NT driver model, and OSR 2.5 added IE 4 as a built-in component.
20171005 18:51:33< AI0867> ugh
20171005 18:51:44< AI0867> I was once asked to help fix someone's computer
20171005 18:51:58< AI0867> so I brought all the drivers they were missing... on a USB stick
20171005 18:52:18-!- Appveyor [~Appveyor@74.205.54.20] has joined #wesnoth-dev
20171005 18:52:18< Appveyor> wesnoth - Release Alexander van Gessel 61b4034: Move function into anonymous namespace Failed
20171005 18:52:18< Appveyor> Details: https://ci.appveyor.com/project/AI0867/wesnoth/build/Wesnoth-VS2017-windows-case-check-5
20171005 18:52:22-!- Appveyor [~Appveyor@74.205.54.20] has left #wesnoth-dev []
20171005 18:52:23< AI0867> it was a pre-OSR 2.1 win95 install
20171005 19:10:12-!- sigurdfd [~SigurdFD@dynamic-acs-72-23-110-196.zoominternet.net] has joined #wesnoth-dev
20171005 19:26:56-!- fabi_ [~fabi@2a02:810c:c840:2e65:b051:dba4:97e7:10b] has joined #wesnoth-dev
20171005 19:30:58-!- fabi__ [~fabi@2a02:810c:c840:2e65:bc3f:d9ab:526f:d6e9] has quit [Ping timeout: 264 seconds]
20171005 19:37:33-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20171005 19:38:55-!- Appveyor [~Appveyor@74.205.54.20] has joined #wesnoth-dev
20171005 19:38:55< Appveyor> wesnoth - Release Alexander van Gessel 9e049b5: Forgotten semicolon Failed
20171005 19:38:55< Appveyor> Details: https://ci.appveyor.com/project/AI0867/wesnoth/build/Wesnoth-VS2017-windows-case-check-6
20171005 19:38:59-!- Appveyor [~Appveyor@74.205.54.20] has left #wesnoth-dev []
20171005 19:44:12< sigurdfd> zookeeper: Did you see my DiD suggestion on Discord recently? Any thoughts?
20171005 19:44:30< zookeeper> umm, no, i only sporadically check there
20171005 19:44:31 * zookeeper looks
20171005 19:46:13< zookeeper> nope, i don't see anything
20171005 19:46:42< zookeeper> oh there it is
20171005 19:48:58< zookeeper> well, i'm kind of a supporter of the current format. not the exact implementation obviously, but i don't share some peoples' categorical disdain for the concept.
20171005 19:51:30< zookeeper> if someone wanted to improve the implementation instead, then that'd make me slightly more thrilled
20171005 19:52:04< sigurdfd> ok. what about this adjustment to what I posted...
20171005 19:53:06< sigurdfd> ... in third time through, scenario objectives of Victory - "Defeat another foolish hero or die trying"; No defeat objectives listed...
20171005 19:53:48< sigurdfd> ...and if you beat the 3rd foolish hero, it goes to epilog, if you die, it does what DiD curretly does, either way, the campaign ends.
20171005 19:54:42< sigurdfd> does that cover what you mean by improve the implementation?
20171005 19:55:58< zookeeper> well, the objective change is a good idea, but i don't really see why you'd want to force it to end after the Xth foolish hero
20171005 19:57:07< zookeeper> i mean, the main problem is that you can cheese your way through way too many foolish heroes, or possibly even forever
20171005 19:57:19< zookeeper> and that's unequivocally bad
20171005 19:57:27< sigurdfd> to deal with the people who have to look at the code/forums to figure out how to finish the campaign
20171005 19:58:22< zookeeper> well, see, if the difficulty ramps up quickly enough, and the dialogue gives enough hints that they're not experiencing an outright bug, then that should take care of any confusion
20171005 19:58:28< sigurdfd> ...and to make sure players don't keep playing s12 excessively
20171005 20:00:26-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20171005 20:00:37< zookeeper> i mean, there's currently 5 foolish hero types. i think the third one should already be excessively difficult so that almost everyone dies there, and to get to the fifth should be practically impossible without massive cheesing and luck.
20171005 20:02:58< zookeeper> in fact, with that in place i suppose you could end the campaign after all the heroes have been encountered. because in practise no regular players would get to that point anyway.
20171005 20:08:34< sigurdfd> ok then. Does something along the lines of the above, with 3rd scenario excessevlyi difficult and starting the "..or die trying" objective, and if player manages to beat 5 they get the epilog ending of going on and on; sound reasonable?
20171005 20:10:09< zookeeper> sure
20171005 20:11:12< zookeeper> not the only changes i think the scenario needs, though. the ending dialogue should have race-specific variants because obviously that's not how orcs talk, and it'd be good if there was some changes to the map like increasing amounts of ruin and decay or whatever, and possibly heroes coming from different directions.
20171005 20:12:51< zookeeper> also the big thing about difficulty is somehow making sure that you can't just hide in a corner and kill everyone with drain while assassinating the foolish hero with a bunch of nightgaunts.
20171005 20:14:25< sigurdfd> ok. I'll see what I can do with that after I finish the DM pr and some stuff with the beginner campaigns.
20171005 20:14:50< zookeeper> great
20171005 20:32:23-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20171005 20:33:37< DeFender1031> some dialogue showing his progressively hopeless state of mind and need for the only remaining way out might help there too
20171005 20:36:23-!- fabi_ [~fabi@2a02:810c:c840:2e65:b051:dba4:97e7:10b] has quit [Ping timeout: 255 seconds]
20171005 20:37:32-!- travis-ci [~travis-ci@ec2-50-17-115-78.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 20:37:33< travis-ci> AI0867/wesnoth#35 (windows-case-check - 61b4034 : Alexander van Gessel): The build passed.
20171005 20:37:33< travis-ci> Build details : https://travis-ci.org/AI0867/wesnoth/builds/283837506
20171005 20:37:33-!- travis-ci [~travis-ci@ec2-50-17-115-78.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 20:38:53-!- louis94 [~~louis94@91.138-200-80.adsl-dyn.isp.belgacom.be] has joined #wesnoth-dev
20171005 20:49:02< zookeeper> DeFender1031, certainly. it's pretty light on dialogue currently.
20171005 20:51:31< DeFender1031> For example, perhaps having him muttering to himself "no way out now, no way out now." after a while would be a good idea
20171005 20:51:52< DeFender1031> "alone in the darkness"
20171005 20:51:56< DeFender1031> creepy stuff like tht
20171005 20:53:31< zookeeper> "nasty heroeses!"
20171005 20:54:47< sigurdfd> :D
20171005 20:55:31< DeFender1031> not quite the direction I was going, but sure. :P
20171005 21:00:38-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has quit [Quit: .]
20171005 21:19:29-!- louis94 [~~louis94@91.138-200-80.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 248 seconds]
20171005 21:19:29-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 248 seconds]
20171005 21:22:03-!- Oebele [~quassel@143.177.58.202] has quit [Remote host closed the connection]
20171005 21:24:30-!- travis-ci [~travis-ci@ec2-50-17-115-78.compute-1.amazonaws.com] has joined #wesnoth-dev
20171005 21:24:31< travis-ci> AI0867/wesnoth#36 (windows-case-check - 9e049b5 : Alexander van Gessel): The build has errored.
20171005 21:24:31< travis-ci> Build details : https://travis-ci.org/AI0867/wesnoth/builds/283840379
20171005 21:24:31-!- travis-ci [~travis-ci@ec2-50-17-115-78.compute-1.amazonaws.com] has left #wesnoth-dev []
20171005 21:40:48-!- fabi_ [~fabi@2a02:810c:c840:2e65:a5a5:868f:be39:916] has joined #wesnoth-dev
20171005 21:45:22-!- fabi [~fabi@wesnoth/developer/fendrin] has quit [Ping timeout: 264 seconds]
20171005 21:57:21< Necrosporus> AI0867, exFAT has patent problem so it's not supported out of the box anywhere but in Windows, right? Plus it is default only for 64G+ volumes
20171005 21:59:09< Necrosporus> I know that long filenames on flash drives formatted in Windows (modern versions, XP+) tend to break when mounted without supplying proper codepage... Now you say that FAT32 supports unicode...
20171005 22:09:11-!- irker945 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20171005 22:25:35-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20171005 22:37:53-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20171005 22:39:48-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171005 22:44:25< shadowm> Is [textdomain] processed every time an add-on is loaded, or only during the core game initialization process?
20171005 22:45:58< shadowm> Okay, I just tried it out myself. The correct answer is the latter.
20171005 22:46:17< shadowm> Which is... um. A bit disappointing. But I can live with it.
20171005 22:49:16-!- fabi [~fabi@2a02:810c:c840:2e65:9462:686:b4a7:f68] has joined #wesnoth-dev
20171005 22:49:16-!- fabi [~fabi@2a02:810c:c840:2e65:9462:686:b4a7:f68] has quit [Changing host]
20171005 22:49:16-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20171005 22:53:11-!- fabi_ [~fabi@2a02:810c:c840:2e65:a5a5:868f:be39:916] has quit [Ping timeout: 255 seconds]
20171005 23:12:05< Necrosporus> I have figured why filenames were garbled
20171005 23:42:56-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20171005 23:55:42-!- fabi_ [~fabi@2a02:810c:c840:2e65:e9db:6011:3296:372e] has joined #wesnoth-dev
20171005 23:59:46-!- fabi [~fabi@wesnoth/developer/fendrin] has quit [Ping timeout: 264 seconds]
--- Log closed Fri Oct 06 00:00:23 2017