--- Log opened Tue Apr 04 00:00:07 2017
20170404 00:08:37-!- vultraz_iOS [uid24821@wesnoth/developer/vultraz] has joined #wesnoth-dev
20170404 00:14:06-!- RatArmy_ [~ratarmy@om126212084036.11.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 00:16:43-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20170404 00:17:19-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 00:21:37< vultraz_iOS> celticminstrel: progress?
20170404 00:21:52-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 260 seconds]
20170404 00:21:59< vultraz_iOS> building, that is
20170404 00:25:29-!- irker938 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20170404 00:25:29< irker938> wesnoth: Charles Dang wesnoth:master 30e0f14863c5 / src/ (5 files in 2 dirs): Updated ancient Copyright notices https://github.com/wesnoth/wesnoth/commit/30e0f14863c5a832bda967c0145280bd8efe46d9
20170404 00:29:34-!- RatArmy_ [~ratarmy@om126212084036.11.openmobile.ne.jp] has joined #wesnoth-dev
20170404 00:29:48-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20170404 00:29:50-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20170404 01:05:52< vultraz_iOS> :( the iterator branch isn't building either
20170404 01:06:58< celticminstrel> Yup.
20170404 01:07:07< vultraz_iOS> C:\Users\exodi\Documents\wesnoth\src\formula\variant.cpp|140|error: no match for 'operator==' (operand types are 'const boost::any' and 'const boost::any')|
20170404 01:07:12< celticminstrel> Probably needs a variant_value_base::equal_iterator.
20170404 01:07:49< celticminstrel> Which takes two boost::any and casts them to the value's iterator type.
20170404 01:08:01< vultraz_iOS> I see
20170404 01:08:33< vultraz_iOS> why did you add iterators for callable?
20170404 01:09:02< celticminstrel> Note that all the any_cast without catch is technically unsafe, though I've fairly sure the unsafe paths shouldn't be reachable with the way it's used currently.
20170404 01:09:11< celticminstrel> No strong reason, just figured it made sense to.
20170404 01:09:43< celticminstrel> Though I suspect the callable iterators actually won't work right now.
20170404 01:10:41< vultraz_iOS> alright, I'll work on this
20170404 01:10:47< vultraz_iOS> still no luck with the other branch?
20170404 01:11:10< celticminstrel> The problem with the callable iterators is that ATM begin() and end() return iterators into different copies of the inputs list.
20170404 01:11:21< celticminstrel> I haven't worked on anything since I pushed the iterator branch, sorry.
20170404 01:11:31< vultraz_iOS> ah
20170404 01:11:32< vultraz_iOS> k
20170404 01:14:49< mattsc> celticminstrel: is there anything specific you want me to test?  I doubt that I’ll be able to do any “comprehensive testsing of all things FAI” this week, but I could check whether some of the existing FAI implementations still work.
20170404 01:14:56< vultraz_iOS> why does inc/dec return any>
20170404 01:14:57< vultraz_iOS> ?
20170404 01:15:12< celticminstrel> It's probably impossible to make the current callable iterator design work while providing a separate begin/end, FTR.
20170404 01:15:18< mattsc> IIRC, there are still a couple FAIs in use in mainline, and I have one in one of my addons.
20170404 01:15:35< celticminstrel> vultraz_iOS: Basically, I'm using "any" to represent the internal iterator itself.
20170404 01:15:47< celticminstrel> So inc/dec takes an internal iterator as input, and outputs a new one.
20170404 01:16:28< vultraz_iOS> I assume equals just needs a bool
20170404 01:18:18< vultraz_iOS> ok, for the base it can just return true, i think
20170404 01:19:02< celticminstrel> vultraz_iOS: For the base it can probably just return true, yeah.
20170404 01:19:34< celticminstrel> And it should be fine for it to assume its inputs are both of the same, correct type.
20170404 01:20:16< celticminstrel> mattsc: Well, what I did for testing was -tformula and just watch for a turn or two to see if it appeared to be working. I can at least confirm that the guarding, priority, and patrol tests in that scenario worked.
20170404 01:20:41< celticminstrel> I didn't look too closely at what the leader was supposed to be doing (opening.fai and the FAI candidate actions).
20170404 01:21:21< celticminstrel> So basically...
20170404 01:21:51< celticminstrel> If you don't have time for comprehensive testing, that's the stuff I already tested.
20170404 01:22:06< celticminstrel> I don't have anything specific though that I want tested besides that.
20170404 01:22:39< celticminstrel> So you could probably just select a few things at random? >_>
20170404 01:24:00 * vultraz_iOS ponders why de-refing a map iterator needs to return a callable....
20170404 01:24:34< vultraz_iOS> instead of just the value
20170404 01:26:15< vultraz_iOS> ie, why does it return value_type
20170404 01:33:39< vultraz_iOS> ah, I guess I'm using equals()..
20170404 01:41:20-!- ToBeFree [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity]
20170404 01:44:01< vultraz_iOS> celticminstrel: I assume I should be using the equals() implementations in equals_iterator, yes>
20170404 01:44:03< vultraz_iOS> ?
20170404 01:47:02< mattsc> celticminstrel: okay, I’ll just poke around a little then
20170404 01:47:11< vultraz_iOS> or..
20170404 01:47:12< vultraz_iOS> hm...
20170404 01:47:15< vultraz_iOS> no...
20170404 01:47:20< vultraz_iOS> celticminstrel: how should I do this 
20170404 01:50:10< vultraz_iOS> oh i got it
20170404 01:50:21< vultraz_iOS> i'm NOT supposed to deref
20170404 01:53:50-!- RatArmy_ [~ratarmy@om126212084036.11.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 02:08:17-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 02:10:13< irker938> wesnoth: Charles Dang wesnoth:wfl_variant_iterator 55d7d83ae0fe / src/formula/ (variant.cpp variant.hpp variant_value.cpp variant_value.hpp): Completed implementation of new variant_iterator https://github.com/wesnoth/wesnoth/commit/55d7d83ae0fe47533efa09b967672823f00e15a3
20170404 02:10:15< vultraz_iOS> celticminstrel: ^ 
20170404 02:11:36< vultraz_iOS> celticminstrel: if no objections I'll rebase and merge
20170404 02:15:25-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20170404 02:17:08-!- JyrkiVesterinen [~JyrkiVest@87-100-174-255.bb.dnainternet.fi] has joined #wesnoth-dev
20170404 02:17:14-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 02:30:08-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20170404 02:30:52-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 02:37:28-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20170404 02:48:31< celticminstrel> [Apr 03@9:24:01pm] • vultraz_iOS ponders why de-refing a map iterator needs to return a callable....
20170404 02:48:33< celticminstrel> [Apr 03@9:24:35pm] vultraz_iOS: instead of just the value
20170404 02:48:34< celticminstrel> Uh, what?
20170404 02:49:16< vultraz_iOS> The key_value_pair thing?
20170404 02:49:42< celticminstrel> Oh, because the value type of a map is std::pair.
20170404 02:50:08< celticminstrel> More specifically, map<K, V>::value_type = std::pair<const K, V>
20170404 02:50:27< celticminstrel> ...
20170404 02:50:33< celticminstrel> Wow.
20170404 02:50:39-!- DeFender1031 [~DeFender1@217.132.17.168] has quit [Quit: I'm not back now.]
20170404 02:50:44-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 260 seconds]
20170404 02:50:56< celticminstrel> I accidentally made operator++ function the same as operator--
20170404 02:51:05< celticminstrel> Uh. Other way around, I mean.
20170404 02:51:09< celticminstrel> -- should call dec not inc
20170404 02:51:28< vultraz_iOS> Ah
20170404 02:51:29< vultraz_iOS> Can fix
20170404 02:51:42< celticminstrel> I have something else to fix too, probably.
20170404 02:52:03< vultraz_iOS> Alright, then you fix
20170404 02:52:49< vultraz_iOS> I assume you have no problems with my changes?
20170404 02:55:34< celticminstrel> I don't see any problems.
20170404 02:55:43< vultraz_iOS> :D
20170404 02:56:48< vultraz_iOS> should we remove  variant_iterator::operator++ with the int argument?
20170404 02:57:08< celticminstrel> No?
20170404 02:58:28< vultraz_iOS> ah, I see, it operates on a copy..
20170404 02:58:33< vultraz_iOS> but why the int
20170404 02:59:01< celticminstrel> It's just how C++ lets you overload postincrement.
20170404 02:59:10< vultraz_iOS> I see
20170404 02:59:52< vultraz_iOS> alright
20170404 03:02:35< vultraz_iOS> I have nothing more to add here
20170404 03:02:45< vultraz_iOS> squash and merge when you're done
20170404 03:04:09< vultraz_iOS> thanks for starting this, btw
20170404 03:04:18< vultraz_iOS> i couldn't figure out how to implement your suggestion
20170404 03:06:38-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 03:15:01-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20170404 03:15:22-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 03:20:58< celticminstrel> I thought I'd let you squash in case you have any last-minute formatting changes to apply, but I can squash if you prefer.
20170404 03:23:45-!- travis-ci [~travis-ci@ec2-107-20-32-152.compute-1.amazonaws.com] has joined #wesnoth-dev
20170404 03:23:46< travis-ci> wesnoth/wesnoth#13161 (wfl_variant_iterator - 55d7d83 : Charles Dang): The build failed.
20170404 03:23:46< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/218306034
20170404 03:23:46-!- travis-ci [~travis-ci@ec2-107-20-32-152.compute-1.amazonaws.com] has left #wesnoth-dev []
20170404 03:29:12 * celticminstrel pokes vultraz_iOS?
20170404 03:37:22< vultraz_iOS> hm?
20170404 03:38:01< vultraz_iOS> celticminstrel: I don't have any more formatting changes
20170404 03:38:34< irker938> wesnoth: Celtic Minstrel wesnoth:wfl_variant_iterator 22544781a258 / src/formula/ (callable.hpp callable_fwd.hpp variant.cpp variant_value.cpp variant_value.hpp): Some tweaks and fixes https://github.com/wesnoth/wesnoth/commit/22544781a258785015b594c6884656739189d601
20170404 03:38:49< celticminstrel> So should I squash that or do you have formatting changes now?
20170404 03:39:10 * vultraz_iOS reviews
20170404 03:40:05 * celticminstrel notes, in case you do squash, that the commit message should mention how callable has been also made iterable.
20170404 03:40:30< vultraz_iOS> eh, just one or two
20170404 03:40:40< vultraz_iOS> I'll do the squash then
20170404 03:40:51< vultraz_iOS> do you want me to push to branch and see what travis says or merge to master?
20170404 03:40:59-!- RatArmy_ [~ratarmy@om126212080043.11.openmobile.ne.jp] has joined #wesnoth-dev
20170404 03:41:04< celticminstrel> Seeing what Travis says does sound like a good idea.
20170404 03:45:35< vultraz_iOS> I still don't know what 'mutable' does
20170404 03:45:52< celticminstrel> It's like a counter to const.
20170404 03:46:02< celticminstrel> A mutable member variable is non-const even when the object is const.
20170404 03:46:10< vultraz_iOS> I see
20170404 03:46:11< celticminstrel> In other words, you can modify it on a const object.
20170404 03:51:09< vultraz_iOS> ok, squashed, doing a local build and then will push
20170404 03:52:21< vultraz_iOS> your latest changes were good
20170404 03:52:31< vultraz_iOS> very much reduced clutter.
20170404 03:54:02< celticminstrel> Also reduced copies by any-casting to a reference. I think.
20170404 03:55:29-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20170404 03:56:40-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 03:56:41-!- RatArmy_ [~ratarmy@om126212080043.11.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 03:57:20< vultraz_iOS> this code is looking so much better than it did a week ago
20170404 03:59:07< irker938> wesnoth: Celtic Minstrel wesnoth:wfl_variant_iterator 8483f8be1c04 / src/formula/ (6 files): Refactored variant_iterator implementation https://github.com/wesnoth/wesnoth/commit/8483f8be1c04bceef17d1a67b23f7207df0dc81a
20170404 03:59:45< vultraz_iOS> any way we can cancel the previous travis builds?
20170404 04:00:03< celticminstrel> Yes.
20170404 04:00:17< celticminstrel> They'll also error once it gets to them now. Since the commits no longer exist on github.
20170404 04:00:24< vultraz_iOS> ah
20170404 04:00:29< celticminstrel> At least, I'm pretty sure I've seen that happen before.
20170404 04:02:37< celticminstrel> Oh huh, if you rename the temp variable in operator++(int) and operator--(int) to iter, it'll reduce the diff slightly.
20170404 04:03:17< celticminstrel> BTW, can you try running wesnoth with -tformula on the iterator branch?
20170404 04:04:31< celticminstrel> Then just move your leader to the left and watch the AI turn.
20170404 04:05:45< celticminstrel> vultraz_iOS: One more change to variant.hpp:
20170404 04:05:55< celticminstrel> Replace line 204 with the following two lines:
20170404 04:06:22< celticminstrel> * param value A pointer to a variant value representing the container.
20170404 04:06:57< celticminstrel> * @param iter An underlying iterator for the underlying container.
20170404 04:07:06< celticminstrel> Both should be @param, whoops.
20170404 04:07:28< celticminstrel> Feel free to reword as well if you want.
20170404 04:13:39< vultraz_iOS> after merge
20170404 04:15:10< vultraz_iOS> celticminstrel: wesnoth crashes at end of turn
20170404 04:17:21-!- RatArmy_ [~ratarmy@om126212080043.11.openmobile.ne.jp] has joined #wesnoth-dev
20170404 04:17:33< vultraz_iOS> I have the stacktrace
20170404 04:17:55< vultraz_iOS> celticminstrel: https://pastebin.com/qjkiVW2C
20170404 04:18:19< vultraz_iOS> looks like the deref of an invlaid iterator
20170404 04:18:21< vultraz_iOS> invalid 
20170404 04:20:49< celticminstrel> Hmm...
20170404 04:22:07< vultraz_iOS> it's not a debug build
20170404 04:22:15< vultraz_iOS> so it's not that informative
20170404 04:24:11< celticminstrel> My best guess is that it happens on line 108 of opening.fai.
20170404 04:24:26< celticminstrel> But... there's no find function in the stack trace, so that guess might be off.
20170404 04:27:07< celticminstrel> It's not exactly easy to obtain the WFL stack trace from the C++ one. :P
20170404 04:27:20< celticminstrel> Though, it's semi-possible, which isn't the case with Lua.
20170404 04:27:52< celticminstrel> The most I can be sure of is that the crash occured in a where clause within an if() within another if().
20170404 04:28:13< vultraz_iOS> well that's actually pretty specific
20170404 04:28:21< celticminstrel> Yup.
20170404 04:28:58< celticminstrel> But no way to figure out which if() or which where clause.
20170404 04:29:12< celticminstrel> Even the fact that it's in opening.fai is pretty much only a guess that could be wrong.
20170404 04:29:22< celticminstrel> BTW, did the AI do anything at all before the crash?
20170404 04:29:39< vultraz_iOS> ran through its entire turn
20170404 04:29:58< vultraz_iOS> last thing was the uh... flashing lich i think?
20170404 04:30:47< vultraz_iOS> celticminstrel: yes, the lich moves, kill the archer, flashes black, then the crash happens
20170404 04:31:02< celticminstrel> Oh, if it got that far, then opening.fai was already run successfully.
20170404 04:31:27< celticminstrel> So it has to be one of the FAI candidate actions, then.
20170404 04:31:45< celticminstrel> In fact... I think it has to be the scouting one...
20170404 04:33:13-!- RatArmy_ [~ratarmy@om126212080043.11.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 04:33:15< celticminstrel> There's no filter function visible in the C++ stack trace, unless that's what entry #1 actually was...
20170404 04:39:16< celticminstrel> vultraz_iOS: My suggestion is to add if(!container_ || !iter_) return variant() to variant_iterator::operator*
20170404 04:39:39< celticminstrel> Actually... only the first is a pointer.
20170404 04:39:43< celticminstrel> So if(!container_)
20170404 04:40:02< celticminstrel> And add if(container_ to the operator++ and operator-- implementations before using ->
20170404 04:40:19< celticminstrel> See if that fixes the crash.
20170404 04:41:32-!- RatArmy_ [~ratarmy@om126212080043.11.openmobile.ne.jp] has joined #wesnoth-dev
20170404 04:42:14< vultraz_iOS> celticminstrel: now the crash is gone but the lich no longer moves to attack
20170404 04:42:26< celticminstrel> Huh.
20170404 04:43:14< celticminstrel> Well, I think it's fine to merge anyway, and we can fix the lich in another commit. (Don't forget to also make the doc changes to variant_iterator.)
20170404 04:43:27< celticminstrel> Oh right, we were waiting for Travis, too.
20170404 04:43:39< vultraz_iOS> what doc changes?
20170404 04:43:48< celticminstrel> [Apr 04@12:05:55am] celticminstrel: Replace line 204 with the following two lines:
20170404 04:43:49< celticminstrel> [Apr 04@12:06:22am] celticminstrel: * param value A pointer to a variant value representing the container.
20170404 04:43:50< celticminstrel> [Apr 04@12:06:56am] celticminstrel: * @param iter An underlying iterator for the underlying container.
20170404 04:43:55< celticminstrel> ^@param
20170404 04:45:11-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20170404 04:45:14-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 04:45:22< vultraz_iOS> oh
20170404 04:45:23< vultraz_iOS> wait
20170404 04:45:25< vultraz_iOS> i made a mistake
20170404 04:45:38< vultraz_iOS> I had if(container_) return variant()
20170404 04:45:53< celticminstrel> if(!container_) was what I said, right?
20170404 04:47:56< vultraz_iOS> yes
20170404 04:47:56< vultraz_iOS> I made a typo
20170404 04:47:56< celticminstrel> So the crash isn't fixed then?
20170404 04:47:56< vultraz_iOS> no
20170404 04:47:56< vultraz_iOS> it's worse
20170404 04:47:56< celticminstrel> Worse?
20170404 04:47:56< celticminstrel> How can it be worse?
20170404 04:47:56< vultraz_iOS> (also, for the record, when it was if(container_), it seems there was weird garbled output in console..)
20170404 04:47:56< vultraz_iOS> uh, it crashes earlier
20170404 04:47:56< vultraz_iOS> and just hangs
20170404 04:47:56< vultraz_iOS> after the wolf moves
20170404 04:47:56< celticminstrel> That's not necessarily worse, but okay.
20170404 04:47:56< celticminstrel> Oh wait, after the wolf...
20170404 04:47:56< vultraz_iOS> let me examine thta gabled output..
20170404 04:47:56< celticminstrel> Before recruiting?
20170404 04:48:02< celticminstrel> That garbled output needs to be fixed somehow BTW.
20170404 04:48:41< vultraz_iOS> ERROR: TYPE ERROR: expected object but found null (null())
20170404 04:48:41< vultraz_iOS>   `╝ⁿ
É═g

20170404 04:48:41< vultraz_iOS>   
20170404 04:48:41< vultraz_iOS>   
20170404 04:48:41< vultraz_iOS>   
20170404 04:48:42< vultraz_iOS>   X╧L≡çX└MêX°OΦåXtè∩
20170404 04:48:51< vultraz_iOS> the fuck is this
20170404 04:48:59< celticminstrel> It doesn't help to show the garbled part. :P
20170404 04:49:12< celticminstrel> Hmm.
20170404 04:49:27< celticminstrel> I wonder if that's related to the crash or something separate.
20170404 04:49:42-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 258 seconds]
20170404 04:49:46< vultraz_iOS> when I change to if(!container_) the output is gone
20170404 04:49:53< celticminstrel> Ah, okay.
20170404 04:50:18< celticminstrel> Kinda odd though.
20170404 04:50:20< vultraz_iOS> ok...
20170404 04:50:34< vultraz_iOS> now I went back to if(!container_) and we're in the same place as before
20170404 04:50:45< vultraz_iOS> lich attacks, crash
20170404 04:50:47< celticminstrel> So the stack trace still traces it to the operator* and not to anything in variant_callable?
20170404 04:51:12< vultraz_iOS> let me get a fresh stack
20170404 04:52:08-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20170404 04:52:14< celticminstrel> Also, is there any sort of error message related to the crash? Something shown by the debugger or something, like SIGXXX or whatever.
20170404 04:52:23< vultraz_iOS> it
20170404 04:52:27< vultraz_iOS> s a segfault, yes
20170404 04:52:52< vultraz_iOS> new stack https://pastebin.com/UeCUyaxB
20170404 04:53:37< celticminstrel> Oh huh, it crashed in unit formulas this time.
20170404 04:53:49< celticminstrel> What was the last thing that happened before it crashed?
20170404 04:53:56< celticminstrel> ie the last thing the AI did.
20170404 04:54:02< vultraz_iOS> i didn't see this time
20170404 04:54:08< celticminstrel> Ah, oh well.
20170404 04:54:13< vultraz_iOS> tried again without debugger and it's back to the lich
20170404 04:54:21< celticminstrel> Probably won't help that much.
20170404 04:54:31< celticminstrel> So it's a segfault...
20170404 04:54:49< celticminstrel> And it's not in variant_callable::deref_iterator, seemingly.
20170404 04:55:00< celticminstrel> Which suggests that the if(!container_) should fix it...
20170404 04:55:15< celticminstrel> :|
20170404 04:55:24< celticminstrel> ...oh wait.
20170404 04:55:56< celticminstrel> There's also the possibility that "this" is null.
20170404 04:55:56< vultraz_iOS> what?
20170404 04:56:07< vultraz_iOS> I see
20170404 04:56:09< celticminstrel> Though... given the context, that seems unlikely...
20170404 04:56:11< vultraz_iOS> how?
20170404 04:56:21< celticminstrel> Since this is a variant_iterator, and I don't think we use those through pointers.
20170404 04:56:39< celticminstrel> So if "this" is not null, and container_ is not null, then what the... OH
20170404 04:56:51< celticminstrel> Add = nullptr to container_ in the header.
20170404 04:57:00< vultraz_iOS> what?
20170404 04:57:06< celticminstrel> Try it, see if it helps.
20170404 04:57:12< celticminstrel> (Or in the default constructor if you prefer.)
20170404 04:57:44< celticminstrel> Pointers are "primitive" types, so they are not initialized by default, which means they start with an arbitrary value.
20170404 04:57:46-!- RatArmy_ [~ratarmy@om126212080043.11.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 04:58:02< celticminstrel> Which means that the !container_ test might be failing when it should pass.
20170404 04:58:13< celticminstrel> So try making sure the default constructor initializes it to nullptr.
20170404 04:58:37< vultraz_iOS> celticminstrel: fixed
20170404 04:58:50< celticminstrel> \o/
20170404 05:00:19< vultraz_iOS> might fix that output too
20170404 05:00:34< celticminstrel> Well, fixing the output can be a separate commit. :P
20170404 05:00:51< celticminstrel> If you can figure out where the output is coming from, great, please do fix it.
20170404 05:01:21< celticminstrel> (The output isn't directly related to the variant_iterator stuff - I was also getting it when working on the action_callable stuff, for example.)
20170404 05:01:57< irker938> wesnoth: Charles Dang wesnoth:wfl_variant_iterator 2d1c8692e5a5 / src/formula/variant.cpp: Added some safety checks and correct initialization to variant_iterator https://github.com/wesnoth/wesnoth/commit/2d1c8692e5a50955b63f60d9875229b6101edb33
20170404 05:02:00< irker938> wesnoth: Charles Dang wesnoth:wfl_variant_iterator bbd52026d24f / src/formula/variant.hpp: Fixup a doc comment https://github.com/wesnoth/wesnoth/commit/bbd52026d24fc3f04cc737aa19a3095a869aaec3
20170404 05:02:10< vultraz_iOS> now we're good
20170404 05:02:30< celticminstrel> You're going to squash those three together?
20170404 05:02:47< vultraz_iOS> should I?
20170404 05:03:25< celticminstrel> Well, you don't have to, but it seems wierd to me to squash everything else but not squash that. The comment, at least, should definitely be squashed into the main commit.
20170404 05:04:03< vultraz_iOS> alright, ill squash themm
20170404 05:04:19< celticminstrel> Waiting for Travis first?
20170404 05:04:24< vultraz_iOS> nah
20170404 05:04:28< irker938> wesnoth: Celtic Minstrel wesnoth:wfl_variant_iterator aca6cae04b0d / src/formula/ (6 files): Refactored variant_iterator implementation https://github.com/wesnoth/wesnoth/commit/aca6cae04b0d61ca6d9033db184e596eebca5f8e
20170404 05:08:27-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 240 seconds]
20170404 05:08:35-!- TC04 [~quassel@venus.arosser.com] has joined #wesnoth-dev
20170404 05:09:03< vultraz_iOS> yeah, it seems if you get rid of commits travis cancels the build
20170404 05:09:30< celticminstrel> No, I cancelled them.
20170404 05:10:08< vultraz_iOS> oh
20170404 05:10:18< vultraz_iOS> well then :P
20170404 05:11:51< vultraz_iOS> it's interesting that the C++14 build has been working for a long time
20170404 05:12:57-!- crimson_pingvin [~crimson_p@ec2.happyspork.com] has joined #wesnoth-dev
20170404 05:14:11-!- travis-ci [~travis-ci@ec2-107-20-32-152.compute-1.amazonaws.com] has joined #wesnoth-dev
20170404 05:14:12< travis-ci> wesnoth/wesnoth#13163 (wfl_variant_iterator - 8483f8b : Celtic Minstrel): The build is still failing.
20170404 05:14:12< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/218320948
20170404 05:14:12-!- travis-ci [~travis-ci@ec2-107-20-32-152.compute-1.amazonaws.com] has left #wesnoth-dev []
20170404 05:14:27-!- Netsplit *.net <-> *.split quits: TC01, crimson_penguin
20170404 05:14:33-!- crimson_pingvin is now known as crimson_penguin
20170404 05:14:33-!- crimson_penguin [~crimson_p@ec2.happyspork.com] has quit [Changing host]
20170404 05:14:33-!- crimson_penguin [~crimson_p@wesnoth/developer/crimsonpenguin] has joined #wesnoth-dev
20170404 05:15:21< celticminstrel> Segfault in the formula tests, huh.
20170404 05:15:42< vultraz_iOS> yes
20170404 05:15:45< vultraz_iOS> that's what we just fixed :/
20170404 05:15:47< vultraz_iOS> wasn't it?
20170404 05:15:48< celticminstrel> That's probably the same issue you just fixed though.
20170404 05:15:54< celticminstrel> Yes, that.
20170404 05:17:14< vultraz_iOS> if this passes, then we need to work on the memory leak stuff
20170404 05:17:48< celticminstrel> Curiously, the error I'm getting ther only occurs in that one place.
20170404 05:17:52< celticminstrel> ^there
20170404 05:17:59< celticminstrel> In variant_value_translator.hpp.
20170404 05:18:06-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 05:18:11< celticminstrel> None of the other make_shared instances raise the error.
20170404 05:20:04< vultraz_iOS> where is this file?
20170404 05:20:17< celticminstrel> src/ai/default
20170404 05:20:31< celticminstrel> Sorry, composite not default
20170404 05:21:31< vultraz_iOS> possibly because action_callable has no ctor
20170404 05:22:06< celticminstrel> It implicitly has a default constructor (since formula_callable has a constructor that accepts 0 args).
20170404 05:22:33< celticminstrel> Anyway, many of the other make_shared calls are also action_callables.
20170404 05:23:04< vultraz_iOS> i see
20170404 05:25:51< celticminstrel> My suspicion is that it's something to do with incomlete types.
20170404 05:26:28< vultraz_iOS> try including a bunch of stuff?
20170404 05:27:19< vultraz_iOS> also, do review my use of make_shared
20170404 05:27:27< vultraz_iOS> might be a few places to use fake_ptr
20170404 05:27:38< vultraz_iOS> i didn't really pay attention, just focused on getting it built
20170404 05:27:51< celticminstrel> There are a few errors in your commit, yes. I don't remember off the top of my head, but I intend to go over it again once I fix my error.
20170404 05:27:57< vultraz_iOS> alright
20170404 05:32:02< celticminstrel> I managed to transform the error to "can't convert from shared_ptr<formula_callable> to variant"
20170404 05:32:08< celticminstrel> Which is obviously false though.
20170404 05:32:17< celticminstrel> So something's still missing.
20170404 05:36:19-!- SigurdFD [~SigurdFD@dynamic-acs-72-23-110-196.zoominternet.net] has quit []
20170404 05:38:07< celticminstrel> Now I'm getting more issues...
20170404 05:43:58-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20170404 05:50:49-!- JyrkiVesterinen [~JyrkiVest@87-100-174-255.bb.dnainternet.fi] has quit [Quit: .]
20170404 05:51:59-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Ping timeout: 240 seconds]
20170404 06:00:10-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20170404 06:02:24< celticminstrel> Looks like the errors go away if the variant constructor takes the shared_ptr by reference. Weird.
20170404 06:02:39< celticminstrel> Well, certainly doesn't hurt to take it by reference though.
20170404 06:03:40< celticminstrel> ...wait, now I'm getting them again.
20170404 06:04:27< vultraz_iOS> travis passed
20170404 06:05:36< celticminstrel> Okay, so apparently I have to not only take it by reference but also have a templated constructor that matches base classes. ???
20170404 06:06:18< vultraz_iOS> msvc is weird
20170404 06:06:20< celticminstrel> I wonder if we could get Jyrki or gfgtdf to check if the wfl_memleak_fix_attempt builds on MSVC later than 2013, because if it does I'm inclined to wrap the templated constructor in an ifdef.
20170404 06:06:22< vultraz_iOS> none of this happens for me
20170404 06:06:34< vultraz_iOS> on gcc
20170404 06:08:19< celticminstrel> ...didn't I just get an error-free build a few minutes ago? I thought I did, but now all the errors are back...
20170404 06:13:17-!- travis-ci [~travis-ci@ec2-54-82-95-13.compute-1.amazonaws.com] has joined #wesnoth-dev
20170404 06:13:18< travis-ci> wesnoth/wesnoth#13165 (wfl_variant_iterator - aca6cae : Celtic Minstrel): The build passed.
20170404 06:13:18< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/218329848
20170404 06:13:18-!- travis-ci [~travis-ci@ec2-54-82-95-13.compute-1.amazonaws.com] has left #wesnoth-dev []
20170404 06:13:45< irker938> wesnoth: Celtic Minstrel wesnoth:master aca6cae04b0d / src/formula/ (6 files): Refactored variant_iterator implementation https://github.com/wesnoth/wesnoth/commit/aca6cae04b0d61ca6d9033db184e596eebca5f8e
20170404 06:14:41< celticminstrel> Feel free to delete the branch, too.
20170404 06:15:46< vultraz_iOS> can i also delete gui_renaming
20170404 06:16:00< celticminstrel> Was there something else on it?
20170404 06:16:20< vultraz_iOS> no
20170404 06:16:22< vultraz_iOS> 0 ahead
20170404 06:16:28< celticminstrel> Then yeah.
20170404 06:17:01< vultraz_iOS> and CelticMinstrel-patch-1?
20170404 06:17:08< vultraz_iOS> it has "Fix [animate_unit]text= rejecting translatable strings (GNA25625)"
20170404 06:17:11< vultraz_iOS> i think this was merged
20170404 06:17:20< celticminstrel> So apparently it works if the varaint constructor takes a non-const ref but not if it takes a const ref...
20170404 06:17:21< vultraz_iOS> though it does say 1 ahead..
20170404 06:18:08< celticminstrel> It's merged.
20170404 06:18:22< vultraz_iOS> delete?
20170404 06:18:52< celticminstrel> If I didn't bother to give a branch a name, it's probably a case where I committed from the website, but for whatever reason didn't merge immediately (possibly to allow for a Travis check), so yeah, delete.
20170404 06:19:15< celticminstrel> patch-1 or whatever should be deleted unless there's an open PR.
20170404 06:23:29< celticminstrel> vultraz_iOS: What's with the FIXME! Not safe? comment in attack.cpp?
20170404 06:23:44< vultraz_iOS> i thought it was creating a shared_ptr from this
20170404 06:24:34< celticminstrel> Ah, it's more roundabout, but you're right.
20170404 06:24:50-!- noy_ [~Noy@S01067cb21b205894.vs.shawcable.net] has joined #wesnoth-dev
20170404 06:24:50-!- noy_ [~Noy@S01067cb21b205894.vs.shawcable.net] has quit [Changing host]
20170404 06:24:50-!- noy_ [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 06:25:22-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20170404 06:25:22-!- noy_ is now known as noy
20170404 06:26:36< irker938> wesnoth: Charles Dang wesnoth:master 19e0d86e6dc3 / src/ (formula/function.cpp wesnoth.cpp): Removed some Visual Leak Detector-related code https://github.com/wesnoth/wesnoth/commit/19e0d86e6dc3387196d8c2a8b5236b0828d97178
20170404 06:26:39< irker938> wesnoth: Charles Dang wesnoth:master ee454a8f542d / src/formula/variant.cpp: Marked a few things as static https://github.com/wesnoth/wesnoth/commit/ee454a8f542dd6e88ab42b2634363b366fb7d07a
20170404 06:26:40< vultraz_iOS> and other similar places in the code that reference 'this'
20170404 06:27:04< celticminstrel> IOW anywhere a safe_call_result is constructed.
20170404 06:27:26< celticminstrel> I just went through the two callable_objects.cpp files to catch them all.
20170404 06:27:59< celticminstrel> Marking those as static makes them file-local, which makes sense.
20170404 06:33:40-!- JyrkiVesterinen [~JyrkiVest@85-76-66-240-nat.elisa-mobile.fi] has joined #wesnoth-dev
20170404 06:36:35< JyrkiVesterinen> 20170404 06:06:20< celticminstrel> I wonder if we could get Jyrki or gfgtdf to check if the wfl_memleak_fix_attempt builds on MSVC later than 2013, because if it does I'm inclined to wrap the templated constructor in an ifdef.
20170404 06:36:40< celticminstrel> Found another case where something this was passed to the callable constructor.
20170404 06:36:47< JyrkiVesterinen> I can only test when I get home, about ten hours from now.
20170404 06:36:52< celticminstrel> Ah.
20170404 06:37:05< celticminstrel> ^-something
20170404 06:37:45< vultraz_iOS> i can't find this "filter" code you referred to yesterday
20170404 06:38:24< celticminstrel> Huh?
20170404 06:38:35< celticminstrel> You mean like in eg units/filter.cpp?
20170404 06:38:46< vultraz_iOS> oh
20170404 06:38:48< vultraz_iOS> THAT filtering
20170404 06:39:00< celticminstrel> What other filtering would I be talking about?
20170404 06:40:28< vultraz_iOS> hmmm
20170404 06:40:29< vultraz_iOS> ok, I see
20170404 06:41:10< vultraz_iOS> ok, so I already have map_formula_callable..
20170404 06:41:29< vultraz_iOS> so I need a wfl::formula
20170404 06:41:32< vultraz_iOS> and call evaluate
20170404 06:41:45< vultraz_iOS> but you say it needs to take an action table?
20170404 06:42:11< celticminstrel> The canvas code already has a function_symbol_table IIRC. Just change it to an action_function_symbol_table and pass that in.
20170404 06:42:28< vultraz_iOS> it has no such thing
20170404 06:42:34< vultraz_iOS> it has a map_formula_callable
20170404 06:42:38< celticminstrel> Pretty sure it does somewhere...
20170404 06:42:47< celticminstrel> I might be thinking of the window code...
20170404 06:43:23< vultraz_iOS> ok the window code does
20170404 06:43:57< celticminstrel> You don't necessarily need to include the function table in the canvas class.
20170404 06:44:02< vultraz_iOS> so do you mean me to add an action_function_symbol_table  member to the canvas?
20170404 06:44:11< celticminstrel> It could also be a locally declared table.
20170404 06:44:17< vultraz_iOS> static?
20170404 06:44:33< JyrkiVesterinen> vultraz_iOS: Do you think we should also remove the Debug_with_VLD configuration from the Visual Studio project?
20170404 06:44:48< vultraz_iOS> I assume after I call evaluate I then have all the actions in the action_function_symbol_table?
20170404 06:44:58< celticminstrel> No.
20170404 06:45:15< vultraz_iOS> or do I call execute variant
20170404 06:45:19< celticminstrel> The action_function_symbol_table simply makes the set_var() function available to the formula.
20170404 06:45:19< vultraz_iOS> on the output
20170404 06:45:36< celticminstrel> If you declare it locally in the function, any functions the formula defines will be usable only in that formula.
20170404 06:46:18< celticminstrel> If you declare it in the canvas, functions would persist through all evaluations, though if you were doing it that way you might want to take a leaf from the window's book and offer a separate key solely for declaring the functions.
20170404 06:46:22< vultraz_iOS> JyrkiVesterinen: possibly
20170404 06:46:47< vultraz_iOS> celticminstrel: actually wonder if i can utilize that key
20170404 06:46:56< celticminstrel> ???
20170404 06:47:19< celticminstrel> Are you saying you want to make any functions declared in the window accessible from the canvas?
20170404 06:47:26< vultraz_iOS> JyrkiVesterinen: i don't think anyone's using it
20170404 06:47:39< celticminstrel> If so, you'd pass the window's function table when constructing the action_function_symbol_table.
20170404 06:47:50< vultraz_iOS> celticminstrel: I'm considering
20170404 06:48:02< vultraz_iOS> I don't want to fuck up this design
20170404 06:48:28< celticminstrel> Note that technically there's a minor issue with what I just described.
20170404 06:48:46< celticminstrel> It's that if the window happened to declare a function named set_var or safe_call, it would be inaccessible.
20170404 06:49:00< celticminstrel> Because the action_function_symbol_table's definition of them takes priority.
20170404 06:49:22< celticminstrel> But it's minor because I doubt anyone would declare a function with those names.
20170404 06:50:51< vultraz_iOS> how do you declare a function again?
20170404 06:51:08< vultraz_iOS> I know the FFL syntax (name->command stuff) but I don't think we have that..
20170404 06:51:38< irker938> wesnoth: Charles Dang wesnoth:master 8bf345cb4f6d / src/gui/core/canvas.cpp: GUI2/Canvas: use emplace_back when constructing shapes https://github.com/wesnoth/wesnoth/commit/8bf345cb4f6dcee7690f9c865b3a5ef02eef373e
20170404 06:51:40< celticminstrel> def name(arguments) formula;
20170404 06:52:23< celticminstrel> BTW, I don't think that's quite what the FFL syntax is.
20170404 06:52:56< celticminstrel> 1) I doubt the name is part of the syntax, unless it's optional. 2) "command stuff" is a formula.
20170404 06:53:19< vultraz_iOS> -> is essentially trailing return type
20170404 06:53:36< vultraz_iOS> it declares that foo returns something to be evaluated as a command
20170404 06:53:39< celticminstrel> Uh, no, it's the maps-to operator. :P
20170404 06:53:48< vultraz_iOS> not in FFL
20170404 06:54:05< celticminstrel> That's a generic statement.
20170404 06:54:18< celticminstrel> And probably the reason why FFL and C++11 chose it for these purposes.
20170404 06:54:36< vultraz_iOS> oh wait 
20170404 06:54:40< vultraz_iOS> def() is a thing there too
20170404 06:54:41< vultraz_iOS> i forgot
20170404 06:54:47< celticminstrel> I don't recall exactly how FFL works, but I doubt -> would mean it returns a command.
20170404 06:55:11< celticminstrel> (BTW, my new action_callable is essentially like FFL command, I think.)
20170404 06:55:22< vultraz_iOS> name: "def()->command stuff",
20170404 06:55:39< vultraz_iOS> actually commands
20170404 06:55:40< vultraz_iOS> but whatever
20170404 06:55:59< celticminstrel> Are you sure about that?
20170404 06:56:07< vultraz_iOS> https://github.com/anura-engine/wesnoth2/blob/master/data/objects/titlescreen_controller.cfg#L16
20170404 06:56:26< celticminstrel> Okay yeah, that's what I thought.
20170404 06:56:47< celticminstrel> The "-> xyz" part indicates the return type. So I guess your statement earlier is accurate.
20170404 06:56:53< celticminstrel> [Apr 04@02:53:19am] vultraz_iOS: -> is essentially trailing return type
20170404 06:57:22< celticminstrel> And "stuff" as "a formula".
20170404 06:57:27< vultraz_iOS> yes
20170404 06:57:36< celticminstrel> And it's presumable def(args) rather than just def() :P
20170404 06:57:45< vultraz_iOS> yeah
20170404 06:57:49< vultraz_iOS> just that in this case it takes no args
20170404 06:58:08< vultraz_iOS> looks like in FFL the command is just 'set'
20170404 06:58:21< vultraz_iOS> though that doesn't apply to us yet since we haven't implemented object properties 
20170404 06:58:25< vultraz_iOS> so set_var is accurate
20170404 06:59:11< vultraz_iOS> still think we should work to implement more of the functionality of FFL and integrate it in with WML like FFL did with FSON.
20170404 07:00:15< vultraz_iOS> One thing that would be nice is a load_file command
20170404 07:00:31< celticminstrel> There's a run_file() function in the AI, actually.
20170404 07:00:54< celticminstrel> Feel free to move it into src/formula and add it to the action_function_symbol_table.
20170404 07:00:55< vultraz_iOS> I see
20170404 07:01:10< vultraz_iOS> Should definitely be generic and used instead of macro inclusions
20170404 07:01:41< celticminstrel> Should also make sure it's secure in terms of how it finds the file.
20170404 07:01:51< celticminstrel> It probably is, but I haven't really looked at it.
20170404 07:02:01< vultraz_iOS> Actually, I think we should decide on a dedicated formula key for use in wml. Like lua has <<< >>> and variable formulas have $()
20170404 07:02:19< celticminstrel> Not sure what you mean.
20170404 07:02:26< celticminstrel> <<>> is not dedicated for Lua.
20170404 07:02:50< vultraz_iOS> Some way to denote that the value of a key is in WFL
20170404 07:03:22< celticminstrel> There is technically no way to denote that it's Lua either ATM.
20170404 07:03:32< celticminstrel> You can write Lua with regular quotes.
20170404 07:03:42< celticminstrel> (As long as it contains no curly braces.)
20170404 07:03:47< celticminstrel> (Or something like that.)
20170404 07:04:27< vultraz_iOS> I see
20170404 07:05:28< celticminstrel> https://github.com/CelticMinstrel/wml-pygments/blob/master/wesnoth.py#L327
20170404 07:05:57< celticminstrel> My code for detecting whether an attribute is WFL.
20170404 07:06:26< celticminstrel> ...apparently it's wrong, too. Some of those FAIs should be WFL.
20170404 07:06:46< celticminstrel> Well, only one of them, I think.
20170404 07:07:48< celticminstrel> Ah, wait, no.
20170404 07:07:58< celticminstrel> It's correct. I missed the WFL a little further up.
20170404 07:08:35< celticminstrel> So basically, if you want a specific key for formula stuff, "formula" is what's already used in filters and thus is probably a good choice.
20170404 07:08:52< celticminstrel> I think that code might also not detect the new movetype patching syntax.
20170404 07:09:00< celticminstrel> Or GUI2 WFL.
20170404 07:09:07< celticminstrel> It almost certainly doesn't detect that.
20170404 07:13:42< celticminstrel> I'm looking at the FAI candidate action execution code, and it looks really weird.
20170404 07:14:39< celticminstrel> It iterates over my_units_flt with index-based iteration.
20170404 07:15:04< vultraz_iOS>  a lot of the code's been using that
20170404 07:15:06< vultraz_iOS> for some reasn
20170404 07:15:08< celticminstrel> And then it uses the index to index my_units.
20170404 07:15:09< vultraz_iOS> the old code
20170404 07:15:32< celticminstrel> Which makes me wonder if it's even actually setting the right unit.
20170404 07:16:23< vultraz_iOS> as for formula==
20170404 07:16:27< vultraz_iOS> formula=*
20170404 07:16:28-!- atarocch [~atarocch@93.56.160.28] has joined #wesnoth-dev
20170404 07:16:47< vultraz_iOS> i kinda meant a delim to allow for WFL for any key
20170404 07:16:55< vultraz_iOS> we're not there yet, though
20170404 07:17:03< celticminstrel> I think that might be impossible.
20170404 07:17:28< celticminstrel> WFL requires an object to act upon. What would the object be if you used such a general syntax?
20170404 07:17:50< celticminstrel> Also, when would the WFL be evaluated?
20170404 07:18:12< celticminstrel> It might not be impossible, but those questions would at least need to be answered.
20170404 07:19:00< vultraz_iOS> hm
20170404 07:19:01< vultraz_iOS> true
20170404 07:19:12< vultraz_iOS> considering our design it does cause problem
20170404 07:19:49< celticminstrel> I think this code doesn't actually need my_units_flt.
20170404 07:20:00< celticminstrel> I should probably keep these changes to a separate commit though...
20170404 07:20:15-!- travis-ci [~travis-ci@ec2-23-22-154-243.compute-1.amazonaws.com] has joined #wesnoth-dev
20170404 07:20:16< travis-ci> wesnoth/wesnoth#13166 (master - aca6cae : Celtic Minstrel): The build has errored.
20170404 07:20:16< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/218340985
20170404 07:20:16-!- travis-ci [~travis-ci@ec2-23-22-154-243.compute-1.amazonaws.com] has left #wesnoth-dev []
20170404 07:21:21< vultraz_iOS> in anura the "context" is the object that owns the properties
20170404 07:21:28< celticminstrel> Okay, so I was wrong. This code is correct as long as all the units are actually valid units.
20170404 07:21:48< vultraz_iOS> though I'm not sure what the FSON context is...
20170404 07:22:12< celticminstrel> If one of the units is invalid (not a unit), then it might pass the wrong unit to the evaluation formula, though...
20170404 07:27:40-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 07:32:08-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 260 seconds]
20170404 07:32:23< celticminstrel> Kinda wish variant_iterator had an operator->
20170404 07:32:37< celticminstrel> You'd need a variant_ref helper class to implement it.
20170404 07:33:11< vultraz_iOS> hm?
20170404 07:33:13< vultraz_iOS> operator->?
20170404 07:33:25< celticminstrel> With most iterators you can write iter->member
20170404 07:33:28< vultraz_iOS> oh, right
20170404 07:33:32< vultraz_iOS> deref access
20170404 07:33:36< celticminstrel> With variant_iterator you have to write (*iter).member
20170404 07:34:01< celticminstrel> This is because operator* returns a temporary value, rather than a reference.
20170404 07:34:16< celticminstrel> Which makes the obvious implementation of operator-> impossible.
20170404 07:36:27< vultraz_iOS> right
20170404 07:36:33< vultraz_iOS> that would be a reference to a temporary
20170404 07:36:59< celticminstrel> Pointer, but yes.
20170404 07:37:14< celticminstrel> Some of the config iterators also have this problem though, and they do support operator->
20170404 07:37:21< celticminstrel> So it's not insurmountable.
20170404 07:37:40< celticminstrel> IIRC it's the config child iterators (not attribute iterators) that have the problem.
20170404 07:42:17< irker938> wesnoth: Charles Dang wesnoth:master f8d13b3f5687 / src/gui/ (38 files in 7 dirs): GUI2: deployed emplace_back en-masse where appropriate https://github.com/wesnoth/wesnoth/commit/f8d13b3f5687b93c6da0308ea80b0ed96bbbe000
20170404 07:42:21< celticminstrel> ...function_table.cpp:588+
20170404 07:42:29< celticminstrel> WHY IS IT TAKING THE ADDRESS OF A TEMPORARY LOCATION
20170404 07:52:36< vultraz_iOS> celticminstrel: if that is so how would it even build
20170404 07:52:53< celticminstrel> That's an interesting story.
20170404 07:53:15< celticminstrel> The key is that it's indirect.
20170404 07:53:41< celticminstrel> And loc() returns a reference, fooling the compiler into thinking it's safe to take its address.
20170404 07:55:20< vultraz_iOS> I see
20170404 07:55:23< vultraz_iOS> that's not good
20170404 07:55:25< vultraz_iOS> do fix
20170404 07:55:46< celticminstrel> Already made the commit. I'll cherry-pick it and push to master as soon as I'm finished going over your commit.
20170404 07:59:24< vultraz_iOS> so much emplace still to deploy..
20170404 08:26:02< irker938> wesnoth: Celtic Minstrel wesnoth:wfl_memleak_fix_attempt d0caf2ed144d / src/formula/ (callable_fwd.hpp variant.cpp variant.hpp): MSVC 2013 fixups https://github.com/wesnoth/wesnoth/commit/d0caf2ed144df3e7c454801bd4cbfd30ed2f5661
20170404 08:26:04< irker938> wesnoth: Celtic Minstrel wesnoth:wfl_memleak_fix_attempt d0883ab0fa93 / src/ai/formula/function_table.cpp: Typo fixes https://github.com/wesnoth/wesnoth/commit/d0883ab0fa93df1cb409421f2211545ce9f7bd73
20170404 08:26:06< irker938> wesnoth: Celtic Minstrel wesnoth:wfl_memleak_fix_attempt e3e69b67cc96 / src/ai/formula/function_table.cpp: Fix potential access to invalid memory in FormulaAI https://github.com/wesnoth/wesnoth/commit/e3e69b67cc96a492326a85b9cb714a8aa2640806
20170404 08:26:08< irker938> wesnoth: Celtic Minstrel wesnoth:wfl_memleak_fix_attempt f18c6dd36977 / src/ (17 files in 5 dirs): Cleanup, more use of fake_ptr() where needed, less use of get(), more auto, some https://github.com/wesnoth/wesnoth/commit/f18c6dd369770f944419819ea3b52c8ea447ed3a
20170404 08:35:37< irker938> wesnoth: Celtic Minstrel wesnoth:master 56b26d26163d / src/ai/formula/function_table.cpp: Typo fixes https://github.com/wesnoth/wesnoth/commit/56b26d26163d0979a457d121b5062daf55c892ed
20170404 08:35:39< irker938> wesnoth: Celtic Minstrel wesnoth:master 1216adb837b6 / src/ai/formula/function_table.cpp: Fix potential access to invalid memory in FormulaAI https://github.com/wesnoth/wesnoth/commit/1216adb837b6efa58f9fc644b8e850ccdd5591c4
20170404 08:38:46-!- celticminstrel is now known as celmin|sleep
20170404 08:48:41-!- JyrkiVesterinen [~JyrkiVest@85-76-66-240-nat.elisa-mobile.fi] has quit [Quit: .]
20170404 08:54:50-!- travis-ci [~travis-ci@ec2-107-20-32-152.compute-1.amazonaws.com] has joined #wesnoth-dev
20170404 08:54:51< travis-ci> wesnoth/wesnoth#13168 (master - f8d13b3 : Charles Dang): The build passed.
20170404 08:54:51< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/218360725
20170404 08:54:51-!- travis-ci [~travis-ci@ec2-107-20-32-152.compute-1.amazonaws.com] has left #wesnoth-dev []
20170404 09:15:56-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 09:20:24-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 260 seconds]
20170404 09:21:57< matthiaskrgr> guess what kind of compiled warnings are back  :D :P  
20170404 09:22:00-!- JyrkiVesterinen [~JyrkiVest@85-76-66-240-nat.elisa-mobile.fi] has joined #wesnoth-dev
20170404 09:22:01< matthiaskrgr> *iler
20170404 09:28:43< matthiaskrgr> https://pastebin.com/iiJ5N4TX 
20170404 09:31:24< JyrkiVesterinen> Hmm. For now we might even want to keep the first one to remind us that it's not implemented properly.
20170404 09:31:25< JyrkiVesterinen> https://github.com/wesnoth/wesnoth/commit/aca6cae04b0d61ca6d9033db184e596eebca5f8e#diff-2b18cd67d43260cb0bcf172edab83ddeR355
20170404 09:55:26< vultraz_iOS> oh that
20170404 09:55:28< vultraz_iOS> blah
20170404 10:07:22< vultraz_iOS> celmin|sleep: how would that function be implemented properly
20170404 10:11:03< irker938> wesnoth: Charles Dang wesnoth:master c4064f68f6d2 / src/gui/ (9 files in 3 dirs): GUI2: deployed map (and minor vector) emplace where appropriate https://github.com/wesnoth/wesnoth/commit/c4064f68f6d2616add0df2f9dc59994ed67db87d
20170404 10:11:06< irker938> wesnoth: Charles Dang wesnoth:master 2304148302cf / src/ (game_events/action_wml.cpp playturn_network_adapter.cpp): Use emplace_back in cases of push_back + config_of https://github.com/wesnoth/wesnoth/commit/2304148302cf98143c422c5f96d11def9eeb38d3
20170404 10:22:11< irker938> wesnoth: Charles Dang wesnoth:master ff38db42ab8c / src/gui/core/timer.cpp: Small fixup to c4064f6 https://github.com/wesnoth/wesnoth/commit/ff38db42ab8cb312338c9b0f30cd9c6fcd085af7
20170404 10:22:14< irker938> wesnoth: Charles Dang wesnoth:master 8fe671481a2a / src/formula/variant_value.hpp: Added missing override specifiers https://github.com/wesnoth/wesnoth/commit/8fe671481a2ae3171735b860a209326721927efe
20170404 10:22:23< vultraz_iOS> matthiaskrgr: ^
20170404 10:26:58-!- RatArmy_ [~ratarmy@om126200125153.15.openmobile.ne.jp] has joined #wesnoth-dev
20170404 10:28:26-!- Duthlet [~Duthlet@dslb-188-106-146-119.188.106.pools.vodafone-ip.de] has joined #wesnoth-dev
20170404 10:30:45< matthiaskrgr> looking good so far 
20170404 10:30:56< matthiaskrgr> 53% done
20170404 10:43:56-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170404 10:50:31-!- RatArmy_ [~ratarmy@om126200125153.15.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 10:57:19< matthiaskrgr> fixed
20170404 11:04:17-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 11:08:40-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 260 seconds]
20170404 11:20:46-!- RatArmy_ [~ratarmy@om126200125153.15.openmobile.ne.jp] has joined #wesnoth-dev
20170404 11:52:52-!- Appleman1234 [~Appleman1@pl19787.ag1212.nttpc.ne.jp] has quit [Ping timeout: 268 seconds]
20170404 11:52:53-!- RatArmy_ [~ratarmy@om126200125153.15.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 12:17:54-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 12:21:49-!- RatArmy_ [~ratarmy@om126200125153.15.openmobile.ne.jp] has joined #wesnoth-dev
20170404 12:22:24-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 260 seconds]
20170404 12:24:44-!- Kwandulin [~Kwandulin@p200300760F3E7D83E44FE3AB4458A2F0.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170404 13:04:47-!- Appleman1234 [~Appleman1@pl42526.ag1212.nttpc.ne.jp] has joined #wesnoth-dev
20170404 13:21:25-!- RatArmy_ [~ratarmy@om126200125153.15.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 13:22:25-!- irker938 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20170404 13:26:48-!- Appleman1234 [~Appleman1@pl42526.ag1212.nttpc.ne.jp] has quit [Ping timeout: 260 seconds]
20170404 14:16:20< celmin|sleep> vultraz_iOS: The implementation would be rougly the same as for containers, except casting to a different type.
20170404 14:24:56< vultraz_iOS>  I see
20170404 14:25:43< celmin|sleep> formula_input_vector::const_iterator IIRC
20170404 14:31:34-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 264 seconds]
20170404 14:31:50-!- JyrkiVesterinen [~JyrkiVest@85-76-66-240-nat.elisa-mobile.fi] has quit [Quit: .]
20170404 14:39:59-!- Appleman1234 [~Appleman1@pl41368.ag1212.nttpc.ne.jp] has joined #wesnoth-dev
20170404 14:45:39-!- mkdroid [~null@unaffiliated/matthiaskrgr] has joined #wesnoth-dev
20170404 14:52:03-!- celmin|sleep is now known as celticminstrel
20170404 15:05:00< vultraz_iOS> ...I've gotten the flashing images in the storyscreen to work :D
20170404 15:06:12-!- Appleman1234 [~Appleman1@pl41368.ag1212.nttpc.ne.jp] has quit [Ping timeout: 260 seconds]
20170404 15:06:47< vultraz_iOS> trick was they were never really flashing at all
20170404 15:07:51< celticminstrel> ???
20170404 15:08:17< vultraz_iOS> for some reason I was thinking they vanished..
20170404 15:08:19< vultraz_iOS> they never vanished
20170404 15:08:32< vultraz_iOS> just flashed red/white by drawing a red image, then a white image, then a red image..
20170404 15:08:41< celticminstrel> Ah.
20170404 15:09:13< vultraz_iOS> so what I've done is set up a time to run with the dialog
20170404 15:09:22< vultraz_iOS> if there are no floating images for a part, it does nothing
20170404 15:09:55< vultraz_iOS> if there are, it will draw one, then schedule the next one
20170404 15:10:09< vultraz_iOS> right now i have it running every 100 ms.
20170404 15:10:31< vultraz_iOS> which...actually means it ceil rounds every delay value to multiples of 100
20170404 15:10:35< vultraz_iOS> hmmm
20170404 15:10:39-!- mkdroid [~null@unaffiliated/matthiaskrgr] has quit [Quit: I'll be back!]
20170404 15:11:33< vultraz_iOS> well, another solution is to use non-repeating timers
20170404 15:11:53< vultraz_iOS> ie, when we switch parts, we see we have images to draw, draw the first
20170404 15:12:00< vultraz_iOS> which sets a timer for the next
20170404 15:12:05< vultraz_iOS> which will set a timer for the next
20170404 15:12:06< vultraz_iOS> etc etc
20170404 15:12:36< vultraz_iOS> actually this is probably better since then delay values are precisely honored 
20170404 15:14:53< vultraz_iOS> yes, let's do that
20170404 15:15:45-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 15:16:10-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 15:17:29< zookeeper> vultraz_iOS, but i told you they're not flashing days ago already
20170404 15:17:45< vultraz_iOS> apparently it didn't register
20170404 15:17:51< zookeeper> ah well, that may be
20170404 15:17:56< vultraz_iOS> apparently also me *seeing* it myself didnt register :|
20170404 15:18:20-!- gfgtdf [~chatzilla@x4e36999c.dyn.telefonica.de] has joined #wesnoth-dev
20170404 15:20:08< vultraz_iOS> perrrfect
20170404 15:20:13< vultraz_iOS> now the timing is precise 
20170404 15:20:19< vultraz_iOS> and the code is cleaner
20170404 15:20:24< vultraz_iOS> oh I'm a genius 
20170404 15:20:37< vultraz_iOS> in that I used the same paradigm I used before :P
20170404 15:20:42< vultraz_iOS> *sosmart*
20170404 15:30:02< vultraz_iOS> ok, now I just need to figure out how to cut the chain if the part changes mid-draw..
20170404 15:31:29< celticminstrel> ?
20170404 15:32:59< vultraz_iOS> if you move on to the next/previous part while images are drawing, you don't want them to keep drawing :P
20170404 15:33:10< vultraz_iOS> ESPECIALLY if the next part has its own images
20170404 15:34:40< celticminstrel> Urgh, I thought I fixed these errors. :|
20170404 15:37:34-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Ping timeout: 246 seconds]
20170404 15:37:50-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170404 15:50:42-!- JyrkiVesterinen [~JyrkiVest@89-166-118-212.bb.dnainternet.fi] has joined #wesnoth-dev
20170404 15:55:16-!- irker367 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20170404 15:55:16< irker367> wesnoth: Celtic Minstrel wesnoth:wfl_memleak_fix_attempt 35262450dbe7 / src/formula/ (callable_fwd.hpp variant_value.hpp): More fixups https://github.com/wesnoth/wesnoth/commit/35262450dbe787298bf33db52e590c14158a94a3
20170404 15:56:17-!- Appleman1234 [~Appleman1@pl21277.ag1212.nttpc.ne.jp] has joined #wesnoth-dev
20170404 15:59:05< celticminstrel> That branch also appears to break everything :(
20170404 15:59:33< celticminstrel> In that none of the FormulaAI scripts in the formula test scenario work. Most complain "expected object but found int".
20170404 16:00:52< irker367> wesnoth: Celtic Minstrel wesnoth:wfl_memleak_fix_attempt 306782eb782f / src/formula/callable.hpp: Make fake_ptr() non-const, maybe that'll solve some of the clang errors https://github.com/wesnoth/wesnoth/commit/306782eb782fd85d33c9ebaacf7fc4fc27665b5e
20170404 16:06:36< irker367> wesnoth: Jyrki Vesterinen wesnoth:master 3e0a22ec9f29 / projectfiles/VC12/ (liblua.vcxproj wesnoth.sln wesnoth.vcxproj wesnothd.vcxproj wesnothlib.vcxproj): Remove the Debug_with_VLD configuration from the Visual Studio project https://github.com/wesnoth/wesnoth/commit/3e0a22ec9f292e46f8809d4483b7f4c53e6b016a
20170404 16:16:16< irker367> wesnoth: Charles Dang wesnoth:master cd0e015a8d04 / src/gui/dialogs/ (story_viewer.cpp story_viewer.hpp): Story Viewer: implemented floating image handling https://github.com/wesnoth/wesnoth/commit/cd0e015a8d0476886ecf8f0efff4c4dd0c84f03b
20170404 16:17:58< vultraz_iOS> now just need to tackle the base rect thingy
20170404 16:18:11< vultraz_iOS> but not right now
20170404 16:19:58< JyrkiVesterinen> celticminstrel: wfl_memleak_fix_attempt does not build with MSVC2015.
20170404 16:20:00< JyrkiVesterinen> 4>i:\battle for wesnoth\wesnoth\src\formula\callable.hpp(63): error C2662: 'wfl::formula_callable_ptr wfl::formula_callable::fake_ptr(void)': cannot convert 'this' pointer from 'const wfl::formula_callable' to 'wfl::formula_callable &' (compiling source file ..\..\src\chat_command_handler.cpp)
20170404 16:20:00< JyrkiVesterinen> 4>  i:\battle for wesnoth\wesnoth\src\formula\callable.hpp(63): note: Conversion loses qualifiers (compiling source file ..\..\src\chat_command_handler.cpp)
20170404 16:20:39< celticminstrel> Is that including the commit I just pushed?
20170404 16:20:50< celticminstrel> Oh, wait.
20170404 16:20:57< celticminstrel> That commit could've caused it... :|
20170404 16:21:13< JyrkiVesterinen> Yes, I built with the latest commit.
20170404 16:22:51< celticminstrel> I might ask you to test again once I can get both Travis and MSVC 2013 happy with it.
20170404 16:29:21-!- JyrkiVesterinen [~JyrkiVest@89-166-118-212.bb.dnainternet.fi] has quit [Quit: Going offline for an hour or two]
20170404 16:42:55-!- Kwandulin [~Kwandulin@p200300760F3E7D83E44FE3AB4458A2F0.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20170404 17:01:11< irker367> wesnoth: Celtic Minstrel wesnoth:wfl_memleak_fix_attempt 6a99d23841dd / src/formula/callable.hpp: More fixups https://github.com/wesnoth/wesnoth/commit/6a99d23841dde7d225782994b6dcb6faf62e4066
20170404 17:09:25< celticminstrel> What's with Travis not announcing recently...
20170404 17:33:17-!- JyrkiVesterinen [~JyrkiVest@89-166-118-212.bb.dnainternet.fi] has joined #wesnoth-dev
20170404 17:33:53-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 268 seconds]
20170404 17:34:06-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 17:35:29-!- noy [~Noy@wesnoth/developer/noy] has quit [Client Quit]
20170404 17:36:09-!- Kwandulin [~Kwandulin@p200300760F3E7D83ECBE2B9C17BC0DBB.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170404 17:43:13< celticminstrel> Should I close PR973 or just comment on everything wrong about it?
20170404 17:43:19< celticminstrel> vultraz_iOS: ^
20170404 17:47:02< gfgtdf> celticminstrel: i think commit 68ba6f40e4cc53c3c9 of that pr does make sense though
20170404 17:47:07< celticminstrel> Also tying to decide what's a good wiki page to link.
20170404 17:47:37< celticminstrel> gfgtdf: Removing the fresh_staats call?
20170404 17:48:14< gfgtdf> celticminstrel: yes, at lest http://gna.org/bugs/?24383 is still an open bug and it could fix it.
20170404 17:49:28< celticminstrel> Huh, I'm surprised someone would consider stats to be so important.
20170404 17:52:34< celticminstrel> gfgtdf: If you want to cherry-pick that commit, please also take 30266ee, 8a5b4eb, 586a86e, aca596c, and squash all five together.
20170404 17:55:53-!- noy [~Noy@24.114.41.198] has joined #wesnoth-dev
20170404 17:55:57-!- noy [~Noy@24.114.41.198] has quit [Changing host]
20170404 17:55:57-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 17:59:46< gfgtdf> hmm ye, for now i think its easier to wait for them to cleanup their pr.
20170404 18:00:15< celticminstrel> Perhaps relevant is that the submitter and the author of that commit are different.
20170404 18:02:55< celticminstrel> I'm still inclined to just close the PR, but I want to wait for vultraz_iOS's opinion.
20170404 18:06:32-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20170404 18:06:46-!- Shiki [~Shiki@p548555DC.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170404 18:08:29< celticminstrel> Ooh! I think I found the cause of the junk that WFL outputs to console in the event of an error.
20170404 18:08:48< celticminstrel> Every formula has a name. It is stored as a const char*. It is uninitialized unless you call set_name.
20170404 18:10:26< celticminstrel> It seems like only function expressions set the name.
20170404 18:10:49< celticminstrel> Thus, every formula expression has a bogus entry in the call stack consisting of the contents of an arbitrary area of memory.
20170404 18:19:37-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20170404 18:22:29-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 18:29:17-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170404 18:30:27-!- travis-ci [~travis-ci@ec2-54-82-95-13.compute-1.amazonaws.com] has joined #wesnoth-dev
20170404 18:30:28< travis-ci> wesnoth/wesnoth#13176 (master - cd0e015 : Charles Dang): The build has errored.
20170404 18:30:28< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/218527541
20170404 18:30:28-!- travis-ci [~travis-ci@ec2-54-82-95-13.compute-1.amazonaws.com] has left #wesnoth-dev []
20170404 18:31:29< celticminstrel> Error seems to be just a connectivity failure.
20170404 18:34:54< irker367> wesnoth: Celtic Minstrel wesnoth:master 36dbd7cf0288 / src/formula/ (function.cpp function.hpp variant.cpp variant.hpp): Fix garbage showing up in WFL stack traces https://github.com/wesnoth/wesnoth/commit/36dbd7cf02883857125e80ef1283320b28a2bef3
20170404 18:35:37-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20170404 18:37:46-!- vultraz_iOS [uid24821@wesnoth/developer/vultraz] has quit [Quit: Connection closed for inactivity]
20170404 18:49:54-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20170404 19:14:18-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170404 19:14:50-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170404 19:40:10-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 19:40:10-!- Kwandulin [~Kwandulin@p200300760F3E7D83ECBE2B9C17BC0DBB.dip0.t-ipconnect.de] has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
20170404 19:44:46-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 264 seconds]
20170404 19:47:12-!- aidanhs [~aidanhs@81.4.110.234] has quit [Quit: WeeChat 0.4.2]
20170404 19:47:31-!- aidanhs [~aidanhs@81.4.110.234] has joined #wesnoth-dev
20170404 20:05:35-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 20:08:51-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170404 20:09:23-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170404 20:18:58-!- Appleman1234 [~Appleman1@pl21277.ag1212.nttpc.ne.jp] has quit [Ping timeout: 240 seconds]
20170404 20:20:12< irker367> wesnoth: Celtic Minstrel wesnoth:wfl_memleak_fix_attempt cb7ff94ce291 / src/ (21 files in 6 dirs): Implement a destructor subscription system for formula_callable https://github.com/wesnoth/wesnoth/commit/cb7ff94ce291a593291b14d2716ca46b4384cdff
20170404 20:20:47< celticminstrel> I can't find any way to satisfy Travis while also making it build in MSVC 2013.
20170404 20:21:40< celticminstrel> I think that commit should at least fix the latest error reported by Travis.
20170404 20:22:54< celticminstrel> (I also squashed everything together.)
20170404 20:35:18< gfgtdf> celticminstrel: what does msvc 2013 complain about?
20170404 20:36:11< celticminstrel> gfgtdf: Failure to convert std::shared_ptr<some_formula_callable_subclass> to wfl::variant
20170404 20:37:40< gfgtdf> is it expected tp call the explicit variant(const formula_callable* callable) ctor?
20170404 20:37:49< celticminstrel> Yes.
20170404 20:38:04< celticminstrel> But it now takes a shared_ptr<const formula_callable> instead.
20170404 20:39:47< irker367> wesnoth: Celtic Minstrel wesnoth:wml_tag_porting 8c39d2bd8d38 / / (3 files in 2 dirs): Properly port [kill] to Lua https://github.com/wesnoth/wesnoth/commit/8c39d2bd8d3880eff883d2421a2db82e23142ac4
20170404 20:40:33< gfgtdf> celticminstrel: and the cde that calls that ctor has the subclass defined (instead of just forward declared)?
20170404 20:40:36< gfgtdf> code*
20170404 20:41:11< celticminstrel> Given that it's using std::make_shared in 99% of cases... presumably it does?
20170404 20:41:42< celticminstrel> If it didn't, there'd instead be an error about attempting to instantiate an incomplete class, right?
20170404 20:43:39< gfgtdf> yes i'd think so.
20170404 20:47:10-!- Duthlet [~Duthlet@dslb-188-106-146-119.188.106.pools.vodafone-ip.de] has quit [Ping timeout: 264 seconds]
20170404 20:52:16< gfgtdf> celticminstrel: hmm trvis seems to fail at the boost tests after your commit
20170404 20:54:27< celticminstrel> Well, I suppose that's progress.
20170404 20:56:25-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20170404 21:01:24-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170404 21:01:56-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170404 21:03:53< irker367> wesnoth: Celtic Minstrel wesnoth:wfl_memleak_fix_attempt 9db764a653ab / src/tests/test_formula_core.cpp: Fix tests https://github.com/wesnoth/wesnoth/commit/9db764a653abfe6e925cf2bec965a46dbaab7655
20170404 21:05:03-!- RatArmy_ [~ratarmy@om126234112233.16.openmobile.ne.jp] has joined #wesnoth-dev
20170404 21:11:23-!- noy [~Noy@wesnoth/developer/noy] has quit [Quit: noy]
20170404 21:18:22-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 264 seconds]
20170404 21:24:43-!- Appleman1234 [~Appleman1@pl21277.ag1212.nttpc.ne.jp] has joined #wesnoth-dev
20170404 21:31:19-!- RatArmy_ [~ratarmy@om126234112233.16.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 21:41:48-!- JyrkiVesterinen [~JyrkiVest@89-166-118-212.bb.dnainternet.fi] has quit [Quit: .]
20170404 21:42:31-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 21:42:31-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20170404 21:42:58-!- RatArmy_ [~ratarmy@om126234112233.16.openmobile.ne.jp] has joined #wesnoth-dev
20170404 21:47:01-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 21:49:00-!- atarocch [~atarocch@93.56.160.28] has quit [Remote host closed the connection]
20170404 21:53:57-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170404 21:54:29-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170404 22:00:10-!- noy [~Noy@wesnoth/developer/noy] has quit [Quit: noy]
20170404 22:13:31-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20170404 22:14:09-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 22:16:33-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20170404 22:16:46-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170404 22:25:51-!- noy [~Noy@S01067cb21b205894.vs.shawcable.net] has joined #wesnoth-dev
20170404 22:25:57-!- noy [~Noy@S01067cb21b205894.vs.shawcable.net] has quit [Changing host]
20170404 22:25:57-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 22:38:47-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20170404 22:41:42-!- RatArmy_ [~ratarmy@om126234112233.16.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 22:47:30-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170404 22:48:02-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170404 22:49:47-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 22:53:51-!- RatArmy_ [~ratarmy@om126234112233.16.openmobile.ne.jp] has joined #wesnoth-dev
20170404 23:04:11-!- vultraz_iOS [uid24821@wesnoth/developer/vultraz] has joined #wesnoth-dev
20170404 23:06:45-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20170404 23:09:53-!- RatArmy_ [~ratarmy@om126234112233.16.openmobile.ne.jp] has quit [Read error: Connection reset by peer]
20170404 23:13:15-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 23:35:56-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 258 seconds]
20170404 23:37:19-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20170404 23:40:03-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170404 23:40:35-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170404 23:40:54-!- RatArmy_ [~ratarmy@om126234112233.16.openmobile.ne.jp] has joined #wesnoth-dev
20170404 23:50:06-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
--- Log closed Wed Apr 05 00:00:09 2017