00:01:43 <gammafunk> 170 commits for advil...what if he overtakes me eventually :o
00:01:54 <gammafunk> I have to pass elliptic before that happens
00:05:04 <advil> I've still got a ways to go
00:09:29 <Jorgrell> Unstable branch on crawl.jorgrun.rocks updated to: 0.21-a0-344-g2144152 (34)
00:12:13 <Eksell> Unstable branch on CRAWL.XTAHUA.COM updated to: 0.21-a0-344-g2144152 (34)
01:05:24 <nikheizen> you can always console yourself "checkwhite is not a real commit"
01:20:27 <Gretell> Unstable branch on crawl.develz.org updated to: 0.21-a0-344-g2144152 (34)
01:58:28 <Gretell> Windows builds of master branch on crawl.develz.org updated to: 0.21-a0-344-g2144152
02:16:38 <ProzacElf> !tell gammafunk https://www.linkedin.com/pulse/act-philip-holt/
02:16:39 <Sequell> ProzacElf: OK, I'll let gammafunk know.
02:16:49 <ProzacElf> !tell lasty https://www.linkedin.com/pulse/act-philip-holt/
02:16:50 <Sequell> ProzacElf: OK, I'll let lasty know.
02:22:25 <alexjurkiewicz> !tell hellmonk i appreciate your recent commit messages
02:22:26 <Sequell> alexjurkiewicz: OK, I'll let hellmonk know.
02:29:32 -!- Rast- is now known as rast
02:54:02 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.21-a0-344-g2144152
03:15:57 <Rotatell> Unstable branch on crawl.beRotato.org updated to: 0.21-a0-344-g2144152 (34)
03:42:59 -!- amalloy_ is now known as amalloy
03:45:29 -!- amalloy is now known as amalloy_
03:46:20 -!- amalloy_ is now known as amalloy
05:13:11 -!- amalloy is now known as amalloy_
05:15:02 -!- amalloy_ is now known as amalloy
05:36:41 -!- yeeve_ is now known as yeeve
05:47:51 -!- amalloy is now known as amalloy_
09:38:47 <|amethyst> gammafunk:   hm, does get_species_def actually have to be "public" (non-static)?  I would kind of prefer to just have a fixup_gnoll_stats and export that instead
09:41:31 <Cheibriados> 03|amethyst02 07* 0.21-a0-345-ge0b234a: Unbrace 10(16 minutes ago, 1 file, 0+ 2-) 13https://github.com/crawl/crawl/commit/e0b234a320f3
09:41:31 <Cheibriados> 03|amethyst02 07* 0.21-a0-346-g330ab2c: Describe entry layout in species-data.h 10(18 minutes ago, 1 file, 19+ 0-) 13https://github.com/crawl/crawl/commit/330ab2c3618d
09:45:31 <advil> oops
09:45:31 <Sequell> advil: You have 1 message. Use !messages to read it.
09:45:34 <advil> |amethyst: I noticed yesterday that unbrace isn't catching for loops with multi-line bodies (e.g. for if ... sequences)
09:45:40 <advil> !messages
09:45:41 <Sequell> (1/1) gammafunk said (8h 27m 44s ago): mibe says you did a great job on skill targets
09:45:59 <advil> :)
09:46:54 <advil> ...I didn't notice that by actually trying to run unbrace though
09:47:15 <|amethyst> advil:  for if is specifically excluded
09:47:34 <advil> oh, huh...is that supposed to be ok?
09:47:48 <advil> I just manually corrected a bunch of cases
09:48:14 <|amethyst> coding_conventions says "may be" omitted, so that's fine
09:48:22 <|amethyst> When for-loops are nested and the outer loop has no further statements, the
09:48:22 <|amethyst> braces may be omitted.
09:48:27 <|amethyst> The same is true for combined for- and if-conditionals as long as all
09:48:27 <|amethyst> statements fill less than four lines.
09:48:32 <|amethyst> If the order of if- and for-conditionals is reversed, however, place braces.
09:49:13 <advil> huh
09:49:17 <advil> well, these were >4 lines I think
09:49:18 <|amethyst> kind of baroque if you ask me :)
09:49:28 <|amethyst> ah
09:49:31 <|amethyst> then good :)
09:49:52 <|amethyst> I guess unbrace doesn't count lines
09:50:06 <|amethyst> probably the for-if and for-for cases are folded together in unbrace
09:50:14 <|amethyst> and for-for doesn't have the line count restriction
09:50:50 <|amethyst> hm
09:51:12 <|amethyst> looks like unbrace would also permit if-for
09:51:21 <|amethyst> (without braces)
09:52:48 <|amethyst> I think fixing that would approximately double the size/number of nasty regexps
09:53:09 <|amethyst> on the other hand, removing the special case from coding_conventions would simplify the regexps a little
09:55:09 <|amethyst> (maybe more than a little_
09:55:14 <|amethyst> s/_/)/
09:58:39 <advil> I vote for removing the special case
09:59:01 <|amethyst> but think of the millimarvins!
09:59:05 <advil> haha
09:59:42 <advil> sounds like a very old-school exception, like typing that extra key in that one situation just was too annoying to someone who was using ed or whatever
09:59:55 <|amethyst> I think it's more for readability than typing
10:00:17 <|amethyst> for people who prefer to read compact things
10:00:26 <|amethyst> and/or have 24-row editors
10:00:27 <advil> yeah, maybe it's really about for if if sequences
10:00:50 <advil> I do agree that a for if if probably looks better without the braces
10:01:40 <advil> (but so does an if body that is one line that has been line-broken for people who use 80 columns)
10:02:48 <|amethyst> I'm half-tempted to say "curly braces everywhere", but that is a pretty big vertical expansion in some places
10:03:48 <|amethyst> I like the look of more compact code, but I'm more concerned with preventing errors (and to a lesser extent with having simple rules)
10:04:11 <advil> probably you (and now me) are the only one who doesn't realize that it isn't curly braces everywhere >1 line
10:04:17 <advil> er
10:04:20 <advil> who does realize
10:04:29 <advil> too many negations
10:05:52 <|amethyst> also note that, for the actually important stuff (naming), the conventions doc is noncommittal
10:05:58 <|amethyst> but for indentation, whoo boy
10:06:33 <|amethyst> just replace the section with a clang-format invocation and maybe a vim indent function, and be done with it
10:07:01 <|amethyst> (emacs too for people who still use that)
10:09:19 <Jorgrell> Unstable branch on crawl.jorgrun.rocks updated to: 0.21-a0-346-g330ab2c (34)
10:54:23 <trm> we enforce *curly braces everywhere* at our shop. better uniformity overall, no mental overhead (do i put them or not?),  IDE does it automatically so you don't have to manually add them when it finaly become necessary.
10:59:38 <|amethyst> trm:  does everyone get big monitors?  That would help :)
10:59:44 <|amethyst> at least, tall ones :)
11:27:55 <trm> we are on 2 or 3 20+inch monitor
11:56:17 -!- amalloy_ is now known as amalloy
12:23:43 -!- amalloy is now known as amalloy_
13:11:02 -!- amalloy_ is now known as amalloy
13:12:28 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.21-a0-346-g330ab2c (34)
14:53:52 -!- amalloy is now known as amalloy_
15:21:49 <gammafunk> |amethyst: feel free to refactor that, the only code guru around at the time when I asked was advil
15:23:37 <gammafunk> I'm not sure that having a public fixup_gnoll_stats is all that much cleaner, but I guess if we want to minimize external declarations/use of headers it is
15:34:06 -!- yesno_ is now known as yesno
15:44:26 -!- yesno_ is now known as yesno
18:00:47 -!- fazizzz is now known as fazisi
18:06:40 <advil> %git
18:06:40 <Cheibriados> 07|amethyst02 * 0.21-a0-346-g330ab2c: Describe entry layout in species-data.h 10(9 hours ago, 1 file, 19+ 0-) 13https://github.com/crawl/crawl/commit/330ab2c3618d
18:08:20 <Cheibriados> Crash when describing double sword of elec in inventory: ASSERT(xp >= 0) in 'player.cc' at line 2702 failed. 13https://crawl.develz.org/mantis/view.php?id=11245 by Thalfon
18:08:27 <Lantell> Unstable branch on underhound.eu updated to: 0.21-a0-346-g330ab2c (34)
18:11:01 <advil> %git 17ab366a468859063ace431347d7dadcc85a566b
18:11:01 <Cheibriados> 07advil02 * 0.21-a0-319-g17ab366: Remove some doubles, and fix a wrong assumption about shields 10(5 days ago, 2 files, 25+ 17-) 13https://github.com/crawl/crawl/commit/17ab366a4688
18:51:04 -!- amalloy_ is now known as amalloy
21:26:21 -!- amalloy is now known as amalloy_
21:46:17 <nikheizen> If a monster summons something, then you enslave its soul, the summoned monster can persist after you kill the summoner (who becomes a friendly Z)
21:46:45 -!- amalloy_ is now known as amalloy
22:23:45 <advil> |amethyst: wip http://dpaste.com/04Y0594
22:29:14 <advil> uh, version with less stupid indentation: http://dpaste.com/1SQ587X
22:34:51 <gammafunk> so when using this default Scale gets me two decimal points of precision?
22:35:04 <advil> yeah
22:35:20 <advil> I set that because that's what crawl already mostly uses
22:35:40 <advil> still not really sure whether integrating this into crawl is a good idea or not
22:36:00 <gammafunk> this is a bit interesting, as in how this defined postfix ++: http://dpaste.com/1SQ587X#line-128
22:36:07 <gammafunk> taking an int argument?
22:36:23 <advil> yeah, it's pretty lame
22:36:53 <advil> the int is just a dummy
22:37:05 <gammafunk> yeah, ok
22:37:25 <advil> https://stackoverflow.com/questions/4421706/what-are-the-basic-rules-and-idioms-for-operator-overloading was super useful for doing this
22:39:30 <advil> once I've tested this more I'll create a branch to demo converting a few things in crawl to this, and see if anyone actually likes it
22:43:55 <amalloy> advil: i think the conversions should be explicit rather than implicit, at the floating-point ones if you keep them
22:44:43 <advil> aren't they already all explicit?
22:45:11 <amalloy> i thought single-argument constructors were implicit unless marked explicit
22:45:11 <advil> or maybe you mean the constructor
22:45:11 <amalloy> yes
22:45:21 <advil> I see
22:45:52 <amalloy> is it really kosher to put a using std::swap in the middle of a function definition? what does that do?
22:45:53 <advil> I have them like that right now because if they are implicit, stuff like "fixedp<>(2) + 3.9" autoatically works. but yeah, I guess you could be concerned about that.
22:46:41 <advil> well, it works, but I'm not sure why it's that way (I got that idiom from elsewhere)
22:47:13 <amalloy> https://stackoverflow.com/q/4782692/625403
22:47:40 <amalloy> tldr yes, the way you wrote it is best
22:48:39 <advil> ah, interesting
22:48:41 <amalloy> although the answer points out that for a simple argument type like an int it's almost surely useless
22:49:54 <advil> yeah...in principle BaseType could be anything numeric-y (though that's more because I wanted to get some experience writing template code than I think crawl needs that)
22:50:28 <amalloy> your hard work will pay off when someone realizes that what they really need is a fixedp<string>
22:51:13 <gammafunk> fixedp<bolt> Kreygasm
22:51:13 <advil> I got that somewhere in https://stackoverflow.com/questions/3279543/what-is-the-copy-and-swap-idiom
22:51:14 <advil> heh
22:54:57 <amalloy> does fixedp<float, 5> compile? it seems like it should have duplicate constructors
22:55:06 <advil> no, it won't
22:55:13 <amalloy> which like, okay, you shouldn't use floats as the base
22:55:21 <advil> I've been tweaking that, in the current version I don't have a BaseType constructor
22:55:40 <advil> it's sort of annoying, do you know a good way to deal with that?
22:56:16 <amalloy> well, what do you want fixedp<int, 100>(100) to represent? 1 * 100, or 100 * 100?
22:56:47 <advil> 1.00
22:56:55 <advil> er
22:56:57 <advil> 100.00
22:57:07 <advil> I see what you mean
22:57:10 <amalloy> so how am i supposed to construct 0.05?
22:57:37 <advil> ah, well, the float constructors will work fine for that
22:57:40 <advil> I'll think about that
22:58:15 <advil> the int constructor really needs to deal in scaled values in order for arithmetic to work, in the current version
22:58:26 <amalloy> i agree
22:58:42 <amalloy> but then your int constructor and float constructors are inconsistent. one take scaled values, the other takes values and then scales them
22:58:49 <amalloy> the type will be easy to use incorrectly
22:59:22 <advil> in the paste version both take a value and scale them?
23:00:32 <amalloy> i guess you're right, that's not the problem. i mistunderstood what you mean by "scaled values"
23:00:51 <amalloy> the problem is that you have to use the float constructor to represent a non-integral value
23:01:12 <amalloy> right? which means you have to deal with floating point in order to use fixed point
23:01:27 <advil> yeah
23:01:44 <advil> well, only if you need the extra precision at the start
23:01:55 <advil> a standard crawl idiom is to start the function by multiplying everything by 100
23:02:01 <amalloy> i think it would be better to remove the implicit floating conversions, and add a two-argument constructor or something that says "here is an int and its scale"
23:02:18 <amalloy> but i guess that is sorta the same thing as our current fixed-point hacks
23:02:44 <advil> having a two-arg constructor seems reasonable
23:03:01 <amalloy> well, maybe only have one constructor, which takes a non-scaled int (ie, 1 represents 0.01)
23:03:16 <amalloy> and have a different convert function of some sort, which lets 1 represent 1.00
23:03:33 <amalloy> or maybe the opposite of that, so that implicit conversions work reasonable?
23:03:58 <advil> to have integer arithmetic / comparison work (e.g. fixedp<>(10) + 2 == 12) I'd rather the int constructor is taking ints that will be scaled
23:04:03 <amalloy> right
23:04:32 <advil> I guess c++ frowns on factory functions?
23:04:43 <amalloy> it frowns on putting them in the class, i suspect
23:04:54 <advil> that's what I'd do in python for this kind of situation
23:04:55 <amalloy> but functions that return values is like...programming
23:05:06 <advil> heh
23:05:11 <amalloy> i guess you still need a constructor for it
23:05:25 <advil> having that two-value constructor is a good idea anyways
23:06:04 <amalloy> operator bool btw, to test for zero?
23:07:16 <advil> yeah, good catch
23:07:45 <advil> my assumptions about what might happen automatically in terms of overloading are all skewed by python
23:09:04 <amalloy> operator- makes me worry about what will happen if someone makes a fixedp<unsigned, 10>. does that behave well?
23:10:48 <advil> I have only cursorily tested unsigned so far
23:11:18 <advil> I saw advice that defining unary - was just a bad idea, so I could drop that entirely
23:13:56 <amalloy> allowing -x doesn't seem any worse than allowing 0 - x
23:13:59 <amalloy> but what do i know
23:16:44 <advil> heh - on an unsigned leads to overflow issues here
23:17:09 <advil> "- (unsigned int) 10" actually gets -10, not sure why
23:17:23 <advil> I guess - casts it back?
23:21:09 <alexjurkiewicz> did anyone talk about this ijyb thread https://www.reddit.com/r/dcss/comments/74brrj/ijyb/
23:21:31 <amalloy> i upvoted it to be a troublemaker
23:21:45 <amalloy> if that counts
23:22:00 <alexjurkiewicz> i do think one shot kills from ijyb are bad
23:22:11 <alexjurkiewicz> and that without dangerous wands ijyb is really boring
23:23:33 <alexjurkiewicz> i wonder if you could give ijyb hex wands and a loud shout
23:23:44 <alexjurkiewicz> so she polymorphs you and then attracks the level
23:25:00 <amalloy> attrack. gell's gravitas is the only attrack in the game now that singularity is gone
23:25:59 <amalloy> advil: (int)-(unsigned int)10 is implementation-defined behaviour afaict, so it will depend on what system you target
23:26:18 <amalloy> (the cast to int is implied, because -foo can't possibly be an unsigned value)
23:26:58 <advil> ah, interesting
23:59:30 <Eksell> Unstable branch on CRAWL.XTAHUA.COM updated to: 0.21-a0-346-g330ab2c (34)