| --- Log opened Fri Jan 09 00:00:34 2015 |
| 00:30 | | Kindamoody is now known as Kindamoody[zZz] |
| 01:05 | | * Derakon eyes his program. |
| 01:06 | <&Derakon> | For some reason, now when I enable fullscreen mode, the display goes grey instead of showing my UI widgets. |
| 01:06 | <&Derakon> | This is problematic, as it means the only way out is to force-quit the program~ |
| 01:06 | <&Derakon> | (Command-Q / alt-F4 don't work because of Reasons) |
| 01:44 | < Reiv> | HEY VORNICUS |
| 01:44 | < Reiv> | Math, simple math, for you |
| 01:45 | < Reiv> | Roll X dice. 6 of 8 sides are a hit, 2 sides are a miss. You may count one miss side as a hit. What're the odds of getting X damage, X-1 damage, etc? |
| 01:48 | <~Vornicus> | Aside from the "count one miss side as a hit", this is a standard binomial distribution (excel's binomdist) |
| 01:48 | <~Vornicus> | Wait - the "one miss side as a hit" -- do you mean that you have a d8, and either 1s or 8s are a hit, but you pick which one after you roll? |
| 01:49 | <~Vornicus> | or do you mean that one of the missed dice gets flopped to a hit afterwards |
| 01:50 | <~Vornicus> | ('cause that first one is more difficult!) |
| 01:50 | <&McMartin> | If this is what I think it is it cannot be the first. |
| 01:50 | <&Derakon> | The latter should just bias the distribution by 1, shouldn't it? |
| 01:51 | <~Vornicus> | Precisely |
| 01:51 | <~Vornicus> | McM: what do you think it is |
| 01:52 | <&McMartin> | I think it's the X-Wing Miniatures game, again, with Focus enabled. |
| 01:52 | <&McMartin> | Also it doesn't completely bias it because that would be "one free hit", not "turn one miss to hit" |
| 01:52 | <&McMartin> | If you throw all Hits then you don't get a bonus |
| 01:52 | <&McMartin> | (And if it is XWMG then it can't be the first because the miss faces are visually indistinguishable, being as they are entirely blank) |
| 01:53 | <~Vornicus> | Yeah. it's not *exactly* a +1 bias, but it is just like it except that X+1 hits folds down into X hits |
| 02:09 | < Reiv> | One of the missed dice gets flopped to a hit |
| 02:10 | < Reiv> | (I am mostly curious what are the odds of getting 4-for-4 damage on a Concussion Missile with a focus token, McMartin you are correct) |
| 02:11 | < Reiv> | My next curiosity is when there are three sides: Hit, Focus, Blank, and *only* the Blanks qualify for the hit rule. |
| 02:11 | < Reiv> | (4 hits, 2 Focus, 2 Blank.) |
| 02:11 | < Reiv> | So you can in theory roll 4 Focus and deal 0 damage, because there's no hits and nothing to turn into one. |
| 02:13 | < Reiv> | My *next* question is how those answers change when there are 5 dice being rolled instead of 4 dice |
| 02:13 | < Reiv> | Because there's That One Dude coming, and he's going to ruin *everything*. <3 |
| 02:38 | <~Vornicus> | 4*(3/4)^3*(1/4) + (3/4)^4 = 189/256 |
| 02:41 | < Reiv> | huh, that's very good odds indeed |
| 02:41 | < Reiv> | I will keep that in mind |
| 03:05 | <~Vornicus> | (3 hits is 6 * 3/4^2 * 1/4^2 = 54/256; 2 is 12/256; 1 is 1/256) |
| 03:36 | < Reiv> | haha |
| 03:37 | < Reiv> | So 3 hits is almost guranteed, 4 hits is 3 in 4 odds |
| 03:37 | < Reiv> | That's... useful. |
| 03:37 | < Reiv> | IF you can get the double-stack in place, that means they blow the crap outta things |
| 03:37 | < Reiv> | (Of course that's kind of the hard part right there) |
| 03:51 | | Derakon is now known as Derakon[AFK] |
| 04:00 | | AlienPickle [Alien@Nightstar-mmuvu1.res.rr.com] has joined #code |
| 04:17 | | AlienPickle [Alien@Nightstar-mmuvu1.res.rr.com] has quit [[NS] Quit: Leaving] |
| 05:22 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: Leaving] |
| 05:47 | | Checkmate [Z@Nightstar-484uip.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 06:00 | | celticminstrel [celticminst@Nightstar-de4ll6.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 06:34 | | Kindamoody[zZz] is now known as Kindamoody |
| 07:33 | < abudhabi> | Is it just me, or did Google's meddling with YT's comment system do approximately nothing about their quality? |
| 08:08 | | Syka_ [the@Nightstar-c409v3.vividwireless.net.au] has quit [Connection closed] |
| 08:08 | | Syka [the@Nightstar-c409v3.vividwireless.net.au] has joined #code |
| 09:19 | | Reivles [orthianz@Nightstar-r1b.8ub.160.203.IP] has quit [Ping timeout: 121 seconds] |
| 09:24 | | Orthia [orthianz@Nightstar-ijc.59a.184.203.IP] has joined #code |
| 09:24 | | mode/#code [+o Orthia] by ChanServ |
| 09:46 | | AverageJoe [evil1@Nightstar-2ofrtr.ph.cox.net] has joined #code |
| 09:46 | <@TheWatcher> | abudhabi: not just you. |
| 09:55 | < abudhabi> | Hypothesis: The kind of people deterred by a permanent public link between their identity and their online comments are not the same people who make idiotic comments online. |
| 10:24 | <@TheWatcher> | Potential student has expressed interest in doing the introductory c programming course, I replied with details, just got this reply |
| 10:24 | <@TheWatcher> | "as it's a foundation will this course alone mean further development courses will be needed to make me employable in the discipline?" |
| 10:24 | <@TheWatcher> | I think my brain has broken |
| 10:25 | < abudhabi> | Lacks a little punctuation, but seems a valid concern. |
| 10:25 | <@Tarinaky> | TheWatcher: I can follow the logic. |
| 10:26 | <@Tarinaky> | TheWatcher: The answer is, of course, 'yes' with an explanation. |
| 10:26 | <@TheWatcher> | I'm trying to avoid replying with "No, of coruse not! Those three/four year degree programmes that just barely get students to a state where they can start being employable? They're just for shits&giggles!" |
| 10:27 | <@Tarinaky> | Mind you, further study doesn't have to be a degree program. |
| 10:27 | <@Tarinaky> | Plenty of C programmers come from other disciplines. They torture Derakon all the time~ |
| 10:31 | <@Tarinaky> | But probably best not to tell them that. |
| 10:53 | < abudhabi> | TheWatcher: You should probably point out that (documented) experience counts for more than degrees. |
| 11:08 | <@Tarinaky> | That's true of any profession. |
| 11:08 | < abudhabi> | Of course. |
| 11:09 | < AverageJoe> | its how i keep getting work |
| 11:09 | < AverageJoe> | i skipped college |
| 11:12 | < AverageJoe> | one of my co workers is in debt 120k to private loans at 15% interest |
| 11:12 | < AverageJoe> | dat usury |
| 11:12 | < abudhabi> | Mhm. |
| 11:19 | < abudhabi> | I've never quite understood why people get into so much debt in the US. Are there really no public universities? |
| 11:21 | <@Tarinaky> | The public universities are the one charging so much. |
| 11:22 | <@Tarinaky> | The private universities are cheaper but the qualifications are worth less as well |
| 11:25 | < abudhabi> | Where I come from "public" means "funded by taxpayers". |
| 11:27 | < abudhabi> | It seems ridiculous to demand higher education while also not providing it. |
| 11:28 | <@Tarinaky> | Public means the same thing in the US |
| 11:28 | <@Tarinaky> | They charge twice. |
| 11:34 | < abudhabi> | Right. Sounds like an incentive to study abroad. |
| 11:35 | < abudhabi> | Or, well, do what AverageJoe does and not study. |
| 11:45 | < AverageJoe> | who says i didnt study? |
| 11:45 | < AverageJoe> | i just didnt do it at some campus |
| 11:45 | <@Tarinaky> | abudhabi: Except for the fact that European universities charge almost as much for foreign/non-EU students /and/ you have to deal with a language barrier as well |
| 11:46 | < abudhabi> | Tarinaky: OK. Definitely an incentive to skip university. |
| 11:46 | <@Tarinaky> | Plus there will be less financing available because a lot of the American Universities are the ones who offer those loans |
| 11:46 | <@Tarinaky> | plus scholarships/discounts and all the rest. |
| 11:46 | < AverageJoe> | Here in the states many private colleges leech off of the government for their money. University of Phoenix for example has a stadium but no team to play. |
| 11:47 | < AverageJoe> | And is publicly traded |
| 11:47 | <@Tarinaky> | And it'll be harder to find work to finance yourself through university in Europe because Visas |
| 11:47 | < AverageJoe> | they get about 70% of their money from federal loans and grants |
| 11:47 | <@Tarinaky> | AverageJoe: Publicaly Trades != Public Money |
| 11:48 | < AverageJoe> | meaning i can buy stock in the company and get dividends |
| 11:48 | <@Tarinaky> | Yes, but you don't buy the stocks with public money |
| 11:48 | <@Tarinaky> | You buy the stocks with private money |
| 11:48 | <@Tarinaky> | Public money is money raised through taxes. |
| 11:49 | <@Tarinaky> | (and tarrifs and in principle whatever money-making assets are held by the government) |
| 11:50 | < AverageJoe> | anyways, recently the fed govt made it so that private colleges have to prove the degrees they print are worth a shit in the job market before allowing them to collect fafsa (federal financial aide) which at some point in time will lead to the private college bubble burst here in the states |
| 11:51 | < abudhabi> | Time to short the universities, then? :P |
| 11:51 | < AverageJoe> | Adverstisements for these fly by night colleges are all over the damn place here |
| 11:52 | < AverageJoe> | Bus ads, magazines, news papers, movie theatres, tv, radio internet, etc |
| 11:53 | < AverageJoe> | Whenever I see a degree on a resume from Devry or ITT Tech (2 of the most notorious fly by night tech schools), they go right into the special blue folder holder |
| 11:53 | < AverageJoe> | recycle bin |
| 11:55 | | Orthia [orthianz@Nightstar-ijc.59a.184.203.IP] has quit [Ping timeout: 121 seconds] |
| 11:55 | < abudhabi> | What about degrees from Nowhere, East Sovietistan? |
| 11:57 | < AverageJoe> | We dont get many foreigners. I'd prolly end up giving them an interview / benefit of doubt |
| 11:59 | | AverageJoe [evil1@Nightstar-2ofrtr.ph.cox.net] has quit [[NS] Quit: Leaving] |
| 12:00 | | Orthia [orthianz@Nightstar-1nt.60m.184.203.IP] has joined #code |
| 12:00 | | mode/#code [+o Orthia] by ChanServ |
| 12:01 | | Kindamoody is now known as Kindamoody|afk |
| 12:40 | | Checkmate [Z@Nightstar-484uip.cust.comxnet.dk] has joined #code |
| 12:40 | | mode/#code [+o Checkmate] by ChanServ |
| 12:46 | <@Tarinaky> | "Babies cry so much because they can see the future. *baby scraming* This baby is witnessing the internet and the rise of Al-Qaida... Whatever they are." |
| 12:59 | | Orthia [orthianz@Nightstar-1nt.60m.184.203.IP] has quit [Ping timeout: 121 seconds] |
| 13:03 | | Orthia [orthianz@Nightstar-0l9.til.224.119.IP] has joined #code |
| 13:03 | | mode/#code [+o Orthia] by ChanServ |
| 16:23 | | celticminstrel [celticminst@Nightstar-de4ll6.dsl.bell.ca] has joined #code |
| 16:23 | | mode/#code [+o celticminstrel] by ChanServ |
| 16:29 | <@Tarinaky> | So. Ruby |
| 16:29 | <@Tarinaky> | I have a function receiving a relative path in a unix format (i.e. path/seperators) |
| 16:29 | <@Tarinaky> | And I need to emitt the same relative path in a windows format (i.e. path\seperator) |
| 16:29 | <@Tarinaky> | Because I need to pass it on to another function that won't just 'do the right thing' |
| 16:33 | <@Tarinaky> | Pathname's conversion seems to always be to UNIX |
| 17:28 | < simon_> | how would you treat the following claim: "An NFA tests membership in time proportional with its number of states and input size." |
| 17:29 | < simon_> | I think this is not true. an NFA tests membership as a function of the input size and the number of transitions. |
| 17:30 | | gnolam [lenin@Nightstar-t1tbf0.cust.bahnhof.se] has quit [Ping timeout: 121 seconds] |
| 17:31 | < simon_> | actually, I'm not sure this is even true. first off, by saying "I have an NFA", I am not saying how I choose to evaluate it (by trying all combinations simultaneously or by backtracking) |
| 17:33 | | gnolam [lenin@Nightstar-t1tbf0.cust.bahnhof.se] has joined #code |
| 17:33 | | mode/#code [+o gnolam] by ChanServ |
| 17:43 | <@Alek> | wait, you're not talking about No Fucking Idea? |
| 17:43 | < simon_> | that'd be NFI. no, NFA = Non-deterministic Finite Automaton |
| 18:04 | <&ToxicFrog> | Tarinaky: paths with forward slashes are in windows format. |
| 18:04 | <&ToxicFrog> | There is nothing you need to do here. |
| 18:15 | <@celticminstrel> | I wouldn't say they are in Windows format, but Windows does understand them, I think. |
| 18:16 | <@celticminstrel> | It does in most contexts, at least. |
| 18:39 | | Alek [omegaboot@Nightstar-c8t.a00.36.73.IP] has quit [Ping timeout: 121 seconds] |
| 18:44 | <@celticminstrel> | Has anyone here tried fiddling with WiX for creating Windows installers? |
| 19:02 | < simon_> | what's a good function for practicing graph colouring on? GCD is pretty neat, Fibonacci has an empty interference graph. |
| 19:02 | < simon_> | what other small functions are there that might be useful for analysing? |
| 19:03 | < simon_> | I guess LCM as a variant of GCD. |
| 19:07 | <@Tarinaky> | ToxicFrog: celticminstrel: Well, this program doesn't understand it. |
| 19:09 | | Kindamoody|afk is now known as Kindamoody |
| 19:09 | <&ToxicFrog> | Tarinaky: then it's doing something stupid with the paths internally rather than passing them to windows. Unfortunate. |
| 19:14 | <@celticminstrel> | Maybe something like splitting on '\\' |
| 19:15 | <@celticminstrel> | Would there be something like a "canonocalize path" function? Maybe it'd change slashes to backslashes? |
| 19:15 | <@celticminstrel> | ^canonicalize |
| 19:18 | | Alek [omegaboot@Nightstar-c8t.a00.36.73.IP] has joined #code |
| 19:18 | | mode/#code [+o Alek] by ChanServ |
| 19:19 | <@Alek> | oh d'oh. how'd I miss that? guess I should have had my coffee first. :P |
| 19:20 | | Turaiel[Offline] is now known as Turaiel |
| 19:23 | <&McMartin> | celticminstrel: I have in the past had to contend with WiX |
| 19:24 | <&McMartin> | My first answer is "if it is at all your choice, seriously consider NSIS instead because despite still being a cross between assembly and PHP that's still less painful" |
| 19:24 | <@celticminstrel> | Hm. I sorta was considering that after trying to wrestle with WiX for awhile. |
| 19:24 | <&McMartin> | My second is "make sure you're using a relatively recent version of WiX because as recent;y as a few minor versions ago you had to respect 8.3 constraints" |
| 19:24 | <@celticminstrel> | Pretty sure I was using the latest version. |
| 19:24 | <&McMartin> | WiX produces raw MSI which is a nightmare for consumer-level software. |
| 19:25 | <@celticminstrel> | Oh? |
| 19:25 | <&McMartin> | Among MSI's various glorious properties are the fact that the installer database is consulted on uninstall and not cached |
| 19:25 | <&ToxicFrog> | celticminstrel: probably, but I've never used win32api directly, so I don't know what it might be called |
| 19:25 | <@celticminstrel> | Well, I guess I'll investigate NSIS more closely. |
| 19:25 | <&McMartin> | This means that you cannot move *or raname* the downloaded MSI if you want to be able to uninstall it cleanly later |
| 19:26 | | * McMartin has used/does use both professionally |
| 19:26 | <@celticminstrel> | ...uh. That's odd. |
| 19:26 | <@celticminstrel> | That includes if you're using Add/Remove programs to remove it? |
| 19:26 | <&McMartin> | Yes. |
| 19:26 | <&McMartin> | InstallShield was a commercial product that existed primarily to not have to deal with MSI's bullshit while still being usable with the enterprise-level MSI tools |
| 19:27 | <&McMartin> | MSI's notion of upgrade is also a disaster if you think of it as any end-user does |
| 19:27 | <&McMartin> | (It involves actually unpacking as few files as possible and never deleting anything) |
| 19:27 | <&McMartin> | (And it does it with manually provided UUIDs instead of, say, SHA1s of the file contents) |
| 19:27 | <&McMartin> | NSIS, in contrast, is basically a sharchive generator for Windows |
| 19:27 | <&McMartin> | And it's less horrible now that LogicLib exists if you're doing something complicated |
| 19:28 | <&McMartin> | If you're basically just doing a self-extracting EXE that also registers with add/remove programs, there are wizards that will autogenerate NSIS scripts for you |
| 19:28 | <&McMartin> | I believe the one I used back in '03 for stuff like Sable came with NIS Edit. |
| 19:29 | <@celticminstrel> | Alright. |
| 19:29 | <@celticminstrel> | I'd like to at least give the option to choose the install directory. |
| 19:29 | <&ToxicFrog> | McMartin: this...has not been my experience with MSI. |
| 19:30 | <&McMartin> | ToxicFrog: That is interesting, as the Windows Installer and WiX documetnation blares it as a gigantic warning on their pages |
| 19:30 | <&ToxicFrog> | I've successfully uninstalled things that I downloaded as an MSI which was deleted after installing, or which installed from physical media containing an MSI that was not present when it was uninstalled. |
| 19:30 | <&ToxicFrog> | (or installed from network drives that were not mounted when it was uninstalled) |
| 19:30 | <&McMartin> | My understanding is that in those cases the MSI has been wrapped in something that copies the "real" installer into some synthesized directory under %WINDIR% |
| 19:30 | <&McMartin> | This is considered a disadvantage by corp IT because (a) it means a user with admin can mess with stuff and (b) it makes the provisioned image larger |
| 19:31 | <&McMartin> | (Our MSIs at work are wrapped in an NSIS shell to do a similar trick) |
| 19:31 | <&ToxicFrog> | That may have been the case for the stuff on removable media, but for the others it was just "download .msi, double-click" -- maybe it copies itself? I don't know how flexible the msi package format is. |
| 19:31 | <@celticminstrel> | Other optional features I considered were: allowing them to not install the editors; adding a registry key to run the program with a modified PATH (which I've heard to be possible); and I can't remember the third one. |
| 19:31 | <&McMartin> | OK |
| 19:31 | <&McMartin> | Those are all easy - even the second-tier installer generators can do all those things |
| 19:32 | <&McMartin> | (MSI/Windows Installer and NSIS are the only two I consider top-tier) |
| 19:32 | <&McMartin> | TF: MSIs are literally MS Access databases on their own. They can include archive files (".CAB"s). It's possible that it's the CABs that have to be stashed. |
| 19:33 | <@celticminstrel> | Would NSIS also work for cross-building from Linux |
| 19:33 | <&McMartin> | That isn't how the docs I read present it though |
| 19:33 | <&McMartin> | Yes. |
| 19:33 | <&McMartin> | If you want to be able to installer-generate cross-platform, NSIS is your best bet overall, AIUI. |
| 19:33 | <&McMartin> | makensis is fully cross-platform. |
| 19:33 | <@celticminstrel> | I noticed WiX doesn't have non-Windows binaries. |
| 19:33 | < simon_> | I used NSIS some years back. it was nice. |
| 19:34 | <@celticminstrel> | There seems to be something called Visual & Installer that integrates NSIS with Visual Studio, so I guess I'll try that out. |
| 19:34 | <&McMartin> | I don't know of it, but it shouldn't be too awful |
| 19:34 | <&ToxicFrog> | simon_: I used NSIS some years back, and it was a horrifying trainwreck that still managed to be better than the alternatives :( |
| 19:34 | <&McMartin> | ToxicFrog: It still is! |
| 19:35 | <&McMartin> | But sharchives were always horrifying trainwrecks on all platforms so think of it like that~ |
| 19:35 | <@celticminstrel> | On second thoughts, maybe I'll find something else. |
| 19:35 | <@celticminstrel> | It only has a limited trial version. |
| 19:35 | <&McMartin> | Also there's a standard macro package now called LogicLib that gives your installer scripts block-structured if statements and such |
| 19:36 | <&McMartin> | But we wrote UQM's absurdly complicated net installer without one |
| 19:36 | <&McMartin> | Back when we wore onions on our belts, as was the style at the time |
| 19:36 | <&ToxicFrog> | McMartin: sharchives I at least understand the language they're written in, and good documentation exists for that language |
| 19:36 | <&ToxicFrog> | And for the tools that the language can call. |
| 19:37 | <&ToxicFrog> | That was not remotely the case when I use NSIS; there was a lot of "copy this code which no-one understands and wave your hand and then a miracle occurs" in what passed for the official docs. |
| 19:37 | <&ToxicFrog> | ...onions? |
| 19:38 | <&McMartin> | Grampa Simpson reference |
| 19:38 | <&McMartin> | http://nsis.sourceforge.net/Docs/ remains the core reference |
| 19:39 | <&McMartin> | There were also loads of crazy people on the Nullsoft forums, but that is like trying to learn C++ by browsing stackoverflow |
| 19:39 | <&McMartin> | The core reference doesn't include LogicLib yet and may also include the Unicode stuff that I think they don't acutally support in releases yet |
| 19:40 | <&McMartin> | ... oh hey, it looks like it's made it to beta though. |
| 19:41 | <&McMartin> | It also looks based on the changelog that it isn't remotely stable yet. |
| 19:42 | <&McMartin> | celticminstrel's problem doesn't actually require any script instructions other than OutPath and File though (and Delete/RMDir in the unstallers) |
| 19:42 | <&McMartin> | Everything else fun falls out of splitting it up into Sections, listing the appropriate pages, and annotating them accordingly |
| 19:42 | <&McMartin> | MUI's docs are still kind of awful but (a) they're part of the official docs now and (b) MUI is optional anyway |
| 19:45 | <@celticminstrel> | ...I don't need a File for every individual file, do I? |
| 19:45 | <&McMartin> | You do. |
| 19:45 | <@celticminstrel> | It'd be easier if I can just say "copy this whole directory tree". |
| 19:45 | <&McMartin> | that's what things like NIS edit get for you |
| 19:46 | <&McMartin> | Note that this is still simpler than WiX, which requires you to not only specify a file individually, but also components individually, assign each component to a Feature, and then each one of these must be provided with a manually provided UUID. |
| 19:46 | <&McMartin> | (Also: No, it turns out that you don't want that thing - it tends to result in slipping .obj files into the installer you didn't intend~) |
| 19:46 | <&McMartin> | One of the usually-a-misfeatures of NSIS is also that the relationship between the installer and the uninstaller is mostly coincidental. |
| 19:46 | <@celticminstrel> | It's a data folder. |
| 19:47 | <@celticminstrel> | There's no object files in it. |
| 19:47 | <@celticminstrel> | It's mostly XML files. |
| 19:47 | <&McMartin> | That's true until it isn't~ |
| 19:47 | <&McMartin> | If you're building on Linux, you would want to exclude .xml~ files at the very least |
| 19:48 | <@celticminstrel> | ...I guess maybe. |
| 19:48 | <&McMartin> | (The flipside of this is that you also have to individually delete files by name in the uninstaller. While there *exists* an rm -rf, actually using it is considered bad form.) |
| 19:48 | <@celticminstrel> | Though in my Windows build, I've already copied the directory tree into the build output directory, so if a hypothetical Linux build did the same, the .xml~ files would be omitted. |
| 19:49 | <@celticminstrel> | If I can generate the list of files as part of the build process, that would be fine too. |
| 19:50 | <&McMartin> | Yeah, it's very common to have a Perl or Python script scan the directories and build an NSI file or an included NSI file on its own |
| 19:50 | <&McMartin> | Our one at work is a little overengineered, as it actually takes the target exes and walks their dll dependencies to decide what to include there. |
| 19:50 | <@celticminstrel> | I just need to figure out how to tack this onto the end of Visual Studio's build. |
| 19:50 | <@celticminstrel> | Probably some sort of empty project that does nothing but execute a script. |
| 19:50 | <&McMartin> | As a makefile project |
| 19:51 | <@celticminstrel> | Okay, sure. |
| 19:51 | <&McMartin> | And it's not empty; it has the NSI file in it. |
| 19:52 | <&McMartin> | Hm. I don't have a sample script that uses Sections meaningfully |
| 19:52 | <&McMartin> | Ophis only has the one so it skips the component section. |
| 19:52 | <&McMartin> | ...oh right |
| 19:53 | <&McMartin> | What language is this thing written in |
| 19:53 | <&McMartin> | If you're building it with Visual Studio you will probably have to also include the C++ Runtime Redistributables |
| 19:53 | <@celticminstrel> | Yeah, I figured that out. |
| 19:54 | <@celticminstrel> | I wasn't sure which DLLs to distribute, so instead I just tossed the installer into the archive. |
| 19:54 | <&McMartin> | That's the traditional approach, yeah |
| 19:54 | <@celticminstrel> | It's written in C++. |
| 19:54 | <&McMartin> | The installer is smart enough to no-op if it's already been installed. |
| 19:54 | <&McMartin> | (Ophis has to do this to work with py2exe) |
| 19:55 | <@celticminstrel> | It seems that I broke Visual Studio by restarting while an update was in progress... forcing me to manually redownload the update. |
| 19:56 | <@celticminstrel> | Restarting the computer, that is. |
| 19:56 | <&McMartin> | FWIW I also am not sold on the idea of putting the installer generator in the main solution |
| 19:56 | <&McMartin> | I build much more regularly than I release |
| 19:57 | <@celticminstrel> | Any particular reasons to use or not use the "modern interface"? |
| 19:58 | <&ToxicFrog> | McMartin: not only is it smart enough to do that, that is the only supported way of checking if it's already installed. |
| 19:58 | <&McMartin> | The "modern interface" is slightly harder to use, but it is more flexible and it looks like a "Proper Windows Application" |
| 19:59 | <&ToxicFrog> | (which is why you get a "installing directx" and "installing msvs runtime" message every time you install a game on steam) |
| 19:59 | <&McMartin> | (yep) |
| 19:59 | <&McMartin> | celticminstrel: If you've got something that is doing it for you, pick modern interface |
| 19:59 | <&McMartin> | You'll have to use the MUI macros instead of the stuff in Section 3 of the manual, basically |
| 20:00 | <@celticminstrel> | One of the options is "add uninstaller". That sounds ugly. :/ |
| 20:00 | <&McMartin> | Yes |
| 20:00 | <&McMartin> | As I said above... |
| 20:00 | <&McMartin> | NSIS scripts aren't conceptually reversible |
| 20:00 | <&McMartin> | So there's no causal link between installer and uninstaller. |
| 20:00 | <&McMartin> | That's one of its usually-a-misfeatures |
| 20:01 | <@celticminstrel> | Can it be removed with Add/Remove Programs? |
| 20:01 | <&McMartin> | Yes |
| 20:01 | <&McMartin> | If you want to remove it with that, you *must* make an uninstaller |
| 20:01 | <@celticminstrel> | Good enough then, |
| 20:01 | <&McMartin> | Add/Remove programs is literally a registry key that says "when the user asks to uninstall this program, run the following commandline" |
| 20:01 | <@celticminstrel> | Do I even need a license page if it's GPL? |
| 20:01 | <@celticminstrel> | Ah. |
| 20:02 | <&McMartin> | You don't *need* one, but it's customary to use the license page to show it and to replace Accept/Decline with a single "Continue" button |
| 20:02 | <&McMartin> | Which Ophis does, one mome |
| 20:02 | <&McMartin> | nt |
| 20:02 | <@celticminstrel> | I have no idea what "number of component sections" means. |
| 20:02 | <&McMartin> | Remember when you said you wanted to select whether or not to install various things with checkboxes? |
| 20:02 | <&McMartin> | Each of those "things" is an NSIS Component Section. |
| 20:03 | <&McMartin> | The macros to modify the license page for GPL-friendliness are: |
| 20:03 | <&McMartin> | ; License page |
| 20:03 | <&McMartin> | !define MUI_LICENSEPAGE_BUTTON "Install" |
| 20:03 | <&McMartin> | !define MUI_LICENSEPAGE_TEXT_BOTTOM "Press the Install button to continue." |
| 20:03 | <&McMartin> | !insertmacro MUI_PAGE_LICENSE "..\..\README" |
| 20:03 | <@celticminstrel> | Huh, apparently no-one ever bothered to commit a license file to the repository. |
| 20:03 | | Turaiel is now known as Turaiel[Offline] |
| 20:03 | <&McMartin> | (Replace filenames and words appropriately) |
| 20:08 | <&McMartin> | Rust-lang has officially hit alpha |
| 20:19 | <@Tarinaky> | In other news: apparently my porformance exeeds expectations. |
| 20:19 | <@Tarinaky> | I shudder to think how low the bar is :P |
| 20:28 | | Turaiel[Offline] is now known as Turaiel |
| 20:43 | <&McMartin> | Reminder that programmer productivity spans three orders of magnitude |
| 21:00 | <@celticminstrel> | There's some way to make nmake do nothing if the source file hasn't changed, isn't there? |
| 21:06 | | Turaiel is now known as Turaiel[Offline] |
| 21:07 | <@celticminstrel> | Ah, found it. The target name needs to be the same as the output file. |
| 21:08 | <&McMartin> | Yeah, I think that's true with makefiles in general |
| 21:09 | <@celticminstrel> | I keep forgetting that. |
| 21:09 | <&McMartin> | If the target name is missing that overrides "have the sourcefiles change" |
| 21:09 | <&McMartin> | *changed |
| 21:09 | <@celticminstrel> | Now I just have to figure out how to generate the list of data files. |
| 21:09 | | Checkmate [Z@Nightstar-484uip.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 21:11 | | * McMartin mostly uses Python for cross-platform directory-walking shenanigans. |
| 21:14 | <@celticminstrel> | Eh, why not. |
| 21:15 | | * iospace sighs |
| 21:15 | <@iospace> | i hate magic numbers |
| 21:15 | <@iospace> | sooooooooooooooo much |
| 21:18 | <@iospace> | Stoner jokes aside, 420 is /not/ a magic number |
| 21:18 | <@iospace> | *acceptable as a magic number |
| 21:18 | <@iospace> | *not acceptable |
| 21:23 | <&McMartin> | Well |
| 21:23 | <&McMartin> | It's OK as an entry in a 500-entry enum~ |
| 21:38 | <@iospace> | ok, as an enum, fine |
| 21:38 | <@iospace> | THIS ISN'T AN ENUM |
| 21:40 | <@Alek> | somebody's hardcoded password hash? :P |
| 21:40 | <@Alek> | </badjoke> |
| 21:42 | <&McMartin> | TWISTER WALRUS |
| 22:02 | | Checkmate [Z@Nightstar-pdi1tp.customer.tdc.net] has joined #code |
| 22:02 | | mode/#code [+o Checkmate] by ChanServ |
| 22:12 | | Kindamoody is now known as Kindamoody[zZz] |
| 22:20 | <@celticminstrel> | ...now nmake doesn't want to rebuild when a source file has changed. |
| 22:29 | <@celticminstrel> | That aside, I now have a working installer. |
| 22:30 | <@celticminstrel> | I suppose there should be some sort of conditional thing so that MSVC redistributables aren't included if it's a non-VS build (eg MinGW), but that can wait until someone decides to build it that way. |
| 22:32 | <@celticminstrel> | Just need to fix the makefile so that it works properly if something is changed. |
| 22:47 | <@celticminstrel> | The source file is listed as a dependency, so I dunno why it's not working... |
| 23:21 | | Thalass [thalass@Nightstar-1dh83p.bigpond.net.au] has quit [Ping timeout: 121 seconds] |
| 23:24 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 23:24 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 23:26 | <@celticminstrel> | Fixed, yay. |
| 23:40 | <&McMartin> | The source file may be a dependency but the things the source file refers to won't be. |
| 23:40 | <&McMartin> | Make is Intrinsically Kind Of Bad At That though custom tools can work around it |
| 23:41 | <@celticminstrel> | The actual issue was that the output file was in a subfolder. |
| 23:46 | | * McMartin installs the Rust alpha. |
| --- Log closed Sat Jan 10 00:00:49 2015 |