--- Log opened Sun Oct 08 00:00:26 2017
20171008 00:34:27-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20171008 00:35:01-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171008 00:43:22-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: KABOOM! It seems that I have exploded. Please wait while I reinstall the universe.]
20171008 00:43:44-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20171008 01:06:40-!- gfgtdf [~chatzilla@x4e368b26.dyn.telefonica.de] has quit [Quit: ChatZilla 0.9.93 [Firefox 56.0/20170926190823]]
20171008 01:12:17-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20171008 01:18:49-!- vultraz [uid24821@wesnoth/developer/vultraz] has quit [Quit: Connection closed for inactivity]
20171008 01:21:02-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20171008 01:50:38-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20171008 01:50:45-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171008 02:05:32-!- vultraz [uid24821@wesnoth/developer/vultraz] has joined #wesnoth-dev
20171008 02:23:22-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20171008 02:23:30-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171008 02:40:42-!- fabi_ [~fabi@2a02:810c:c840:2e65:9560:c9ab:51d8:5ab6] has joined #wesnoth-dev
20171008 02:44:56-!- fabi [~fabi@wesnoth/developer/fendrin] has quit [Ping timeout: 255 seconds]
20171008 02:56:39-!- sevu [~Shiki@p54854A40.dip0.t-ipconnect.de] has quit [Quit: Verlassend]
20171008 03:50:07-!- fabi [~fabi@2a02:810c:c840:2e65:6893:3caf:2b52:70ba] has joined #wesnoth-dev
20171008 03:50:08-!- fabi [~fabi@2a02:810c:c840:2e65:6893:3caf:2b52:70ba] has quit [Changing host]
20171008 03:50:08-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20171008 03:54:14-!- fabi_ [~fabi@2a02:810c:c840:2e65:9560:c9ab:51d8:5ab6] has quit [Ping timeout: 255 seconds]
20171008 04:35:16< EliDupree> LOL, so I extended my naive AI to learn how to move towards villages and the enemy leader, intending it to be a mediocre but fast-running AI that could be used for playouts for my Monte Carlo AI. To test how it behaved on a large map, I sent it up against the default AI on Cynsuan Battlefield… And it WON its first game
20171008 04:38:21-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has joined #wesnoth-dev
20171008 04:41:55< Necrosporus> EliDupree, and why did you expect it to lose?
20171008 04:42:34-!- sigurdfd [~SigurdFD@dynamic-acs-72-23-110-196.zoominternet.net] has quit []
20171008 04:42:36< EliDupree> Because it's a pretty stupid AI
20171008 04:43:09< EliDupree> Each unit goes towards the closest unowned village, which often means that a whole bunch of units going the same direction. Although that actually ended up being an advantage because it was natural grouping
20171008 04:57:33-!- fabi_ [~fabi@2a02:810c:c840:2e65:b0d9:ba4c:99ce:ea6b] has joined #wesnoth-dev
20171008 05:01:44-!- fabi [~fabi@wesnoth/developer/fendrin] has quit [Ping timeout: 255 seconds]
20171008 05:02:30< EliDupree> Of course, there's also the fact that Cynsuan Battlefield is super unbalanced
20171008 05:06:25< Necrosporus> EliDupree, well, default ai is even stupidier it seems
20171008 05:06:33< Necrosporus> It does often ignore villages for example
20171008 05:06:39< EliDupree> haha
20171008 05:06:44< Necrosporus> When it has chance to attack
20171008 05:06:59< EliDupree> Well, that's not necessarily stupid
20171008 05:07:07< Necrosporus> Even if such attack will leave only a scratch on attacked unit and get AI's units killed
20171008 05:07:14< JyrkiVesterinen> Or just the fact that it doesn't know that it's supposed to protect its leader above all else.
20171008 05:07:48< JyrkiVesterinen> When the AI brings its leader to battle, just attack it with everything you've got and you'll win. No need to care about casualties.
20171008 05:08:35< Necrosporus> Actually you even want to attack with units which it can kill on retaliation first
20171008 05:08:35< EliDupree> Yeah, that's the main reason my Monte Carlo AI has such a high win rate on the tiny map
20171008 05:10:03< Necrosporus> Though I have noticed that AI became a bit smarter
20171008 05:10:16< Necrosporus> For example it did move wounded units to villages sometimes
20171008 05:13:01< shadowm> loonycyborg: Reminder that i've yet to hear of you regarding #2068.
20171008 05:50:50< JyrkiVesterinen> Necrosporus: I tried out disallowing backslashes in file paths.
20171008 05:51:28< Necrosporus> And?
20171008 05:51:35< JyrkiVesterinen> The simplest approach (disallowing them at the filesystem layer) doesn't work. It gets passed paths which are created in ways like "query the userdata directory, then append the rest of the path"
20171008 05:51:50< JyrkiVesterinen> (where the userdata directory contains backslashes)
20171008 05:52:11< Necrosporus> So it disallows all userdata?
20171008 05:52:36< Necrosporus> The function which queries userdata location could convert slashes, though
20171008 05:52:39< JyrkiVesterinen> Yes, including what the core of the game needs to start up.
20171008 05:52:57< JyrkiVesterinen> I believe we'd need that kind of slash conversion in too many places.
20171008 05:53:02< JyrkiVesterinen> It's not feasible.
20171008 05:53:17< Necrosporus> then it should be on higher level, like WML parser?
20171008 05:53:30< shadowm> The WML parser doesn't know anything about paths. You mean the WML preprocessor.
20171008 05:53:42< JyrkiVesterinen> Yes, I think it's a reasonable approach.
20171008 05:53:51< shadowm> AAnd also wesnoth.dofile and whatever the other one is called
20171008 05:53:59< JyrkiVesterinen> And I'd do it at the layer of the WML tag implementation.
20171008 05:53:59< Necrosporus> Another approach is do all check like that only on uploading
20171008 05:54:28< Necrosporus> I think it should improve performance
20171008 05:54:28< shadowm> WML tag? This isn't about the preprocessor substitution path?
20171008 05:54:39< Necrosporus> Both
20171008 05:54:47< Necrosporus> image=wrong\path.png
20171008 05:54:56< shadowm> That's not a WML tag.
20171008 05:55:16< shadowm> That's one possible name for an attribute found in at least a million WML tags.
20171008 05:55:19< Necrosporus> {~add-ons\Bad_Addon\macros\stuff.cfg}
20171008 05:55:36< celticminstrel> Pretty sure dofile uses the same C++ function.
20171008 05:55:56< shadowm> And that's not hyperbole, because WML tags defined by Lua can refer to images however they want. Might as well tell image::get_image()/the image locator code to scan for backslashes.
20171008 05:56:03< Necrosporus> Both places load resources on Windows when you use backslash, but wont load it on other OSes which is the problem
20171008 05:56:07< celticminstrel> To resolve the relative path to an absolute path.
20171008 05:56:09< shadowm> Aaaaaaand waste some time checking long strings every time.
20171008 05:56:19< shadowm> In case Wesnoth's performance wasn't already questionable.
20171008 05:57:00< Necrosporus> This check could only be done if platform is Windows, then performance won't degrade for other OSes
20171008 05:57:01< JyrkiVesterinen> I think checking the path in image loading shouldn't cause any noticeable performance loss.
20171008 05:57:13< JyrkiVesterinen> The cost of scanning a string is nothing compared with disk read.
20171008 05:57:16< celticminstrel> Anyway, IMO we should just merge 2077 and close 2070.
20171008 05:57:32< Necrosporus> I am for 2070
20171008 05:57:35< JyrkiVesterinen> celticminstrel: Agreed.
20171008 05:57:40< shadowm> JyrkiVesterinen: A single string? Sure.
20171008 05:58:08< celticminstrel> And ideally we'd ensure all backslashes are changed to slashes when displaying paths, but still accept either interchangeably.
20171008 05:58:14< Necrosporus> My analysis shows that whitelist will only trouble porting 7 add-ons and only two of them will need to do more than renaming files
20171008 05:58:17< shadowm> I'll let you figure out how many image paths are actually accessed during terrain builder initialization.
20171008 05:58:34< shadowm> And also during the first display of more complex unit animations.
20171008 05:58:40< JyrkiVesterinen> Breaking seven add-ons is way too many IMO.
20171008 05:58:54< Necrosporus> What about 2 add-ons?
20171008 05:58:57< celticminstrel> Breaking one addon is way too many.
20171008 05:59:07< celticminstrel> Unless there's really no way around it.
20171008 05:59:10< Necrosporus> Then don't disallow ? in names
20171008 05:59:22< celticminstrel> Well, that's a problem though.
20171008 05:59:45< celticminstrel> Can this addon be installed on Windows at all?
20171008 05:59:48< Necrosporus> What if we ask add-on developers?
20171008 05:59:54< Necrosporus> celticminstrel, it will error
20171008 06:00:05< celticminstrel> Have you tried it?
20171008 06:00:05< shadowm> <add-on developer> whats a filename
20171008 06:00:12< celticminstrel> XD
20171008 06:00:45< Necrosporus> Anyway, let's ask if they are willing to rename some files or prefer Wesnoth to keep them allowed
20171008 06:01:08< celticminstrel> That sounds like a terrible idea somehow.
20171008 06:01:43< EliDupree> As an add-on developer, I say break ALL the add-ons :-P
20171008 06:01:50< JyrkiVesterinen> Can you be the one to ask them to rename the files, Necrosporus?
20171008 06:01:58< shadowm> As an add-on developer, I say don't break add-ons.
20171008 06:02:04< celticminstrel> ^
20171008 06:02:10< shadowm> And I'm pretty sure I have more spare time in my hands than the average add-on developer.
20171008 06:02:29 * celticminstrel is an add-on dev too FTR, though obviously not of as popular an add-on.
20171008 06:04:58-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20171008 06:06:49-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The computer falls into a deep sleep, to awake again some other day!]
20171008 06:09:14-!- fabi_ [~fabi@2a02:810c:c840:2e65:b0d9:ba4c:99ce:ea6b] has quit [Ping timeout: 255 seconds]
20171008 06:11:01-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20171008 06:16:06-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20171008 06:36:39< JyrkiVesterinen> Okay, looks like I found a better place for the backslash check: https://github.com/wesnoth/wesnoth/blob/1ee9a2e13d42189935c5e3bda299930bfb5400e2/src/filesystem_boost.cpp#L1100-L1120
20171008 06:37:04< JyrkiVesterinen> That function is already intended for developer-supplied paths, which sould only use forward slashes.
20171008 06:37:07< JyrkiVesterinen> Let's try it...
20171008 07:03:06< shadowm> loonycyborg: Never mind, I think I won't need your help after all. I'll need celticminstrel's instead.
20171008 07:04:49-!- irker911 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20171008 07:04:49< irker911> wesnoth: Jyrki Vesterinen wesnoth:no_backslashes 5f09cf3f8d1b / RELEASE_NOTES changelog src/filesystem_boost.cpp: Disallow specifying file paths with backslashes https://github.com/wesnoth/wesnoth/commit/5f09cf3f8d1b0ffd1ae6a2c2c0af36e01c4b5ef4
20171008 07:08:46< Necrosporus> Is  unrenameable=yes supposed to work inside [side] ?
20171008 07:08:54< Necrosporus> warning unit: Unknown attribute 'unrenameable' discarded.
20171008 07:08:59< Necrosporus> In 1.12
20171008 07:09:06< vultraz> not that i know of
20171008 07:09:36< Necrosporus> But there is unit attribute which sets units in way it can't be renamed
20171008 07:09:56< Necrosporus> [side] is supposed to take unit attributes (and apply to leader)
20171008 07:10:22-!- fabi_ [~fabi@2a02:810c:c840:2e65:1092:37ac:7145:c7bd] has joined #wesnoth-dev
20171008 07:14:29-!- fabi [~fabi@wesnoth/developer/fendrin] has quit [Ping timeout: 255 seconds]
20171008 07:22:18< loonycyborg> shadowm: I don't even know much about addon passphrases. I didn't touch that part myself.
20171008 07:29:07< irker911> wesnoth: doofus-01 wesnoth:master 94f3a4c2de23 / data/core/ (76 files in 2 dirs): new terrain variation - Elf castle ruins (#2064) https://github.com/wesnoth/wesnoth/commit/94f3a4c2de23aec2e7f6deadd0b08a0215f44d90
20171008 07:45:21< loonycyborg> I'll review that pm anyway just in case
20171008 07:57:49-!- Oebele [~quassel@143.177.58.202] has joined #wesnoth-dev
20171008 08:17:18-!- travis-ci [~travis-ci@ec2-54-90-86-45.compute-1.amazonaws.com] has joined #wesnoth-dev
20171008 08:17:19< travis-ci> wesnoth/wesnoth#15206 (no_backslashes - 5f09cf3 : Jyrki Vesterinen): The build passed.
20171008 08:17:19< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/284966897
20171008 08:17:19-!- travis-ci [~travis-ci@ec2-54-90-86-45.compute-1.amazonaws.com] has left #wesnoth-dev []
20171008 08:19:48-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20171008 08:23:47-!- fabi_ [~fabi@2a02:810c:c840:2e65:1092:37ac:7145:c7bd] has quit [Ping timeout: 255 seconds]
20171008 09:00:00-!- mjs-de [~mjs-de@x4db52ac3.dyn.telefonica.de] has joined #wesnoth-dev
20171008 09:26:13-!- fabi_ [~fabi@2a02:810c:c840:2e65:f1c5:5a57:5850:2af1] has joined #wesnoth-dev
20171008 09:30:23-!- fabi [~fabi@wesnoth/developer/fendrin] has quit [Ping timeout: 255 seconds]
20171008 09:45:00-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has quit [Quit: .]
20171008 09:47:08-!- louis94 [~~louis94@91.138-200-80.adsl-dyn.isp.belgacom.be] has joined #wesnoth-dev
20171008 10:08:11-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20171008 10:08:18-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171008 10:28:01-!- Oebele [~quassel@143.177.58.202] has quit [Remote host closed the connection]
20171008 10:29:18-!- irker911 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20171008 10:34:38-!- fabi [~fabi@2a02:810c:c840:2e65:4c8:baca:4725:5b5b] has joined #wesnoth-dev
20171008 10:34:41-!- fabi [~fabi@2a02:810c:c840:2e65:4c8:baca:4725:5b5b] has quit [Changing host]
20171008 10:34:41-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20171008 10:38:47-!- fabi_ [~fabi@2a02:810c:c840:2e65:f1c5:5a57:5850:2af1] has quit [Ping timeout: 255 seconds]
20171008 11:26:22-!- louis94 [~~louis94@91.138-200-80.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 260 seconds]
20171008 11:27:07-!- sevu [~Shiki@p54854813.dip0.t-ipconnect.de] has joined #wesnoth-dev
20171008 11:40:03-!- fabi_ [~fabi@2a02:810c:c840:2e65:456f:9dda:a359:fbe0] has joined #wesnoth-dev
20171008 11:44:02-!- fabi [~fabi@wesnoth/developer/fendrin] has quit [Ping timeout: 255 seconds]
20171008 11:56:31-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has joined #wesnoth-dev
20171008 12:26:35-!- fabi_ [~fabi@2a02:810c:c840:2e65:456f:9dda:a359:fbe0] has quit [Quit: Konversation terminated!]
20171008 12:27:51-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20171008 12:37:01-!- fabi [~fabi@wesnoth/developer/fendrin] has quit [Quit: Konversation terminated!]
20171008 12:55:12-!- vultraz [uid24821@wesnoth/developer/vultraz] has quit [Quit: Connection closed for inactivity]
20171008 13:17:46-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171008 13:58:21< Necrosporus> How hard it would be to implement an option which makes image not found errors more prominent?
20171008 13:59:09< Necrosporus> For example make them spawn a dialog box similar to one which you see if file included through {} is not found
20171008 13:59:43< Necrosporus> UMC authors might want to enable it while developing their add-ons to spot image problems
20171008 14:18:22-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20171008 14:25:46< Necrosporus> This is new, a map which has only ice and keeps
20171008 15:08:05-!- sevu [~Shiki@p54854813.dip0.t-ipconnect.de] has quit [Remote host closed the connection]
20171008 15:57:10-!- louis94 [~~louis94@91.138-200-80.adsl-dyn.isp.belgacom.be] has joined #wesnoth-dev
20171008 16:25:15-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20171008 16:25:22-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171008 17:27:24-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20171008 17:27:43-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171008 18:02:01-!- louis94 [~~louis94@91.138-200-80.adsl-dyn.isp.belgacom.be] has quit [Ping timeout: 240 seconds]
20171008 18:04:17-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Ping timeout: 255 seconds]
20171008 18:04:18-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171008 18:09:28-!- sevu [~Shiki@p54854813.dip0.t-ipconnect.de] has joined #wesnoth-dev
20171008 18:25:34-!- stikonas_ is now known as stikonas
20171008 19:43:40-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20171008 19:43:47-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171008 19:55:40-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20171008 19:55:46-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171008 20:23:22-!- sevu [~Shiki@p54854813.dip0.t-ipconnect.de] has quit [Quit: Verlassend]
20171008 20:39:51-!- mjs-de [~mjs-de@x4db52ac3.dyn.telefonica.de] has quit [Remote host closed the connection]
20171008 20:40:42-!- vultraz [uid24821@wesnoth/developer/vultraz] has joined #wesnoth-dev
20171008 21:10:13-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has quit [Quit: .]
20171008 21:11:21-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 240 seconds]
20171008 22:01:02< shadowm> celticminstrel: Can I rely on the hex hashes being 30 characters long then?
20171008 22:01:48< shadowm> Apparently the Base64 ones are 22 chars.
20171008 22:27:42< shadowm> Okay, turns out I can't rely on that.
20171008 22:27:57< shadowm> How am I supposed to tell a hex hash apart from a Base64 one then?
20171008 22:28:56< shadowm> It wouldn't be a problem if the hex hash serialization didn't use a subset of the characters Base64 uses.
20171008 22:30:00< shadowm> Seems like it's going to be easier to just go over the logs for every setpass command I've used and hopefully I've still got all the PMs I received.
20171008 22:36:26< shadowm> ... You know, this is very frustrating.
20171008 22:36:42< shadowm> I just realized I don't have any add-ons to test with.
20171008 22:37:39< shadowm> I guess I can fabricate my own test case using my own hashes and salts from 1.12.
20171008 22:41:04< shadowm> celticminstrel: You said base64_digest() should work, right? Because it doesn't.
20171008 22:42:07-!- fabi [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20171008 22:42:25< shadowm> I took the salt and hash for one of my campaigns from 1.12 into the server configuration for my local test instance, and the server still rejects my passphrase even though it's the same one I used on the 1.12 server.
20171008 22:42:33< shadowm> So that must mean base64_digest() changed too.
20171008 22:47:36< shadowm> Ohhhhh.
20171008 22:47:46< shadowm> Scratch all that.
20171008 22:47:59< shadowm> I forgot about authenticate() and I was patching only set_passphrase().
20171008 22:48:36< shadowm> Not the dumbest thing I've ever done, sadly.
20171008 22:55:53-!- irker220 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20171008 22:55:53< irker220> wesnoth: Ignacio R. Morelle wesnoth:master 2d836ddcd9bb / changelog src/campaign_server/campaign_server.cpp: campaignd: Use base64-encoded hashes for add-on passphrases again https://github.com/wesnoth/wesnoth/commit/2d836ddcd9bb6a657f4b9cc22a4a49d90fd0f276
20171008 23:01:23-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20171008 23:39:00-!- travis-ci [~travis-ci@ec2-107-22-24-24.compute-1.amazonaws.com] has joined #wesnoth-dev
20171008 23:39:01< travis-ci> wesnoth/wesnoth#15214 (master - 2d836dd : Ignacio R. Morelle): The build has errored.
20171008 23:39:01< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/285259208
20171008 23:39:01-!- travis-ci [~travis-ci@ec2-107-22-24-24.compute-1.amazonaws.com] has left #wesnoth-dev []
--- Log closed Mon Oct 09 00:00:27 2017