--- Log opened Thu Jul 21 00:00:20 2016 20160721 00:01:12-!- Duthlet [~Duthlet@dslb-188-106-027-209.188.106.pools.vodafone-ip.de] has quit [Quit: leaving] 20160721 00:10:22-!- prkc [~prkc@catv-80-98-46-199.catv.broadband.hu] has quit [Ping timeout: 250 seconds] 20160721 00:13:10-!- Appleman1234 [~Appleman1@KD036012051087.au-net.ne.jp] has joined #wesnoth-dev 20160721 00:23:26-!- prkc [~prkc@46.166.190.177] has joined #wesnoth-dev 20160721 01:32:11-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev 20160721 01:36:49-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 252 seconds] 20160721 02:06:13-!- celmin [~celticmin@unaffiliated/celticminstrel] has joined #wesnoth-dev 20160721 03:13:31-!- celmin [~celticmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The minstrel departs, to spread the music to the masses!] 20160721 03:38:24-!- kool [ae174eeb@gateway/web/freenode/ip.174.23.78.235] has joined #wesnoth-dev 20160721 03:38:32< kool> hi 20160721 03:40:13-!- kool [ae174eeb@gateway/web/freenode/ip.174.23.78.235] has left #wesnoth-dev [] 20160721 03:40:26< celticminstrel> Or not. 20160721 04:02:40< Aginor> celticminstrel: you didn't respond within 1 minute, I'd have lost patience and interest too :D 20160721 04:05:05-!- JyrkiVesterinen [~JyrkiVest@87-100-169-5.bb.dnainternet.fi] has joined #wesnoth-dev 20160721 04:10:25< celticminstrel> Have time to look at PRs now? 20160721 04:12:47< Aginor> celticminstrel: still at work, sorry 20160721 04:13:18< Aginor> celticminstrel: later today though 20160721 04:14:29< celticminstrel> Whoa it's Thursday already. 20160721 04:14:34< celticminstrel> I should sleep. 20160721 04:17:09-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev 20160721 04:21:49-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 252 seconds] 20160721 04:29:18-!- iceiceice [~chris@wesnoth/developer/iceiceice] has joined #wesnoth-dev 20160721 04:29:27< Aginor> celticminstrel: Thursday afternoon if you ask me ;) 20160721 04:29:48 * Aginor makes obligatory joke about living in the future 20160721 04:30:49< iceiceice> loonycyborg, Rhonda, I think you are understating the value 20160721 04:30:56< iceiceice> "< loonycyborg> iirc to make life easier for UMC creators" 20160721 04:31:10< iceiceice> a better way to describe it would be, wesnoth doesn't support translated user made campaigns right now 20160721 04:31:18< iceiceice> and after spirit-po branch, it will 20160721 04:32:04< iceiceice> using mo files isn't an issue for most unix applications 20160721 04:32:21< iceiceice> but for any program that wants to add translated content after the fact, it creates a huge distribution problem 20160721 04:33:08< iceiceice> iirc wescamp has not been active in like 5 years 20160721 04:33:33< iceiceice> it's far simpler just to load po files 20160721 04:38:11-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev 20160721 04:42:18-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 244 seconds] 20160721 04:44:06-!- Appleman1234 [~Appleman1@KD036012051087.au-net.ne.jp] has quit [Ping timeout: 240 seconds] 20160721 05:04:59< Rhonda> iceiceice: How does it not support translated UMC? I fail to see that. Why can't UMC use msgfmt? 20160721 05:05:08< iceiceice> because how do they distribute it? 20160721 05:05:11< iceiceice> mo files are not portable 20160721 05:05:20< Rhonda> In po format. 20160721 05:05:26< iceiceice> so then what do the players do? 20160721 05:05:31< iceiceice> they have to run msgfmt themselves? 20160721 05:05:48< Rhonda> Ah, wait. Not portable. I think I see what you mean there. 20160721 05:06:49< iceiceice> in the past i guess we worked around it by like 20160721 05:06:53< iceiceice> ai0867 had some bash script 20160721 05:07:04< iceiceice> that would generate the mos for each campaigna nd each target architecture 20160721 05:07:11< iceiceice> but no one wants to maintain that anymore 20160721 05:08:10< vultraz> I'm pretty sure some campaigns have translated umc 20160721 05:08:11< vultraz> er 20160721 05:08:14< vultraz> some umc are translated 20160721 05:08:16< vultraz> blah 20160721 05:14:16 * Rhonda asking in #debian-i18n which should contain people working in this area and hopefully have a bit of knowledge on what other projects might do. :) 20160721 05:15:22-!- Kwandulin [~Miranda@p200300760F6062D03D45B796F8FFF196.dip0.t-ipconnect.de] has joined #wesnoth-dev 20160721 05:23:09< celticminstrel> I guess I'll try to remember to check the logs tomorrow. Feel free to remind me if there's something I should be responding to, though. 20160721 05:23:25-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] 20160721 05:29:11-!- atarocch [~atarocch@2.43.162.39] has joined #wesnoth-dev 20160721 05:40:38-!- mjs-de [~mjs-de@x4db59c4d.dyn.telefonica.de] has joined #wesnoth-dev 20160721 05:41:15-!- Appleman1234 [~Appleman1@KD036012041228.au-net.ne.jp] has joined #wesnoth-dev 20160721 05:44:49-!- mjs-de [~mjs-de@x4db59c4d.dyn.telefonica.de] has quit [Remote host closed the connection] 20160721 05:49:51-!- RatArmy [~RatArmy@133.15.175.65] has joined #wesnoth-dev 20160721 05:54:36-!- atarocch [~atarocch@2.43.162.39] has quit [Remote host closed the connection] 20160721 06:03:09< iceiceice> Rhonda, cool, I would be quite interested to hear what they say 20160721 06:03:35-!- RatArmy [~RatArmy@133.15.175.65] has quit [Ping timeout: 258 seconds] 20160721 06:20:04-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has quit [Ping timeout: 258 seconds] 20160721 06:23:54-!- Kwandulin [~Miranda@p200300760F6062D03D45B796F8FFF196.dip0.t-ipconnect.de] has quit [Ping timeout: 258 seconds] 20160721 06:24:43< shadowm> "a better way to describe it [...] doesn't support" 20160721 06:24:59< shadowm> There's a difference between not supporting X and support for X being fundamentally broken in some way. 20160721 06:26:50-!- Kwandulin [~Miranda@p200300760F6062D03D45B796F8FFF196.dip0.t-ipconnect.de] has joined #wesnoth-dev 20160721 06:30:14-!- RatArmy [~RatArmy@133.15.175.65] has joined #wesnoth-dev 20160721 06:30:20< iceiceice> shadowm, i disagree 20160721 06:31:00< shadowm> Then we're probably speaking different languages. 20160721 06:31:12< shadowm> One question: how is the mo format not portable? 20160721 06:32:16< shadowm> Are we talking about gettext being unavailable on specific platforms, or about the mo format being architecture-dependent? 20160721 06:32:18< iceiceice> i didn't write "msgfmt" or "libintl", so i can't speak to it in all the gory detail, but all the documentation says its not portable 20160721 06:32:51< shadowm> I'm asking because I just noticed that the platform-independent libc-l10n package in Debian is full of .mo files. 20160721 06:33:27-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has joined #wesnoth-dev 20160721 06:36:06< shadowm> If that were truly platform-dependent then I'd imagine users of architectures other than the one used in whatever build server builds each revision of it wouldn't have been very pleased for the past 17 years that glibc has been in Debian. 20160721 06:39:42< iceiceice> so what are you proposing exactly? 20160721 06:39:52< shadowm> I asked you a question. 20160721 06:39:59< iceiceice> you want that umc people should run msgfmt and commit the .mo files to their addons? 20160721 06:40:07< iceiceice> and hope that it works on all architectures? 20160721 06:40:24< iceiceice> i have no idea about some debian libc package, you could ask them 20160721 06:40:29< shadowm> That depends on the answer to my question. 20160721 06:40:42< iceiceice> if you think the gnu gettext docs are wrong and .mo files really are portable, 20160721 06:40:50< iceiceice> i guess you could suggest that gnu correct it 20160721 06:41:07< shadowm> Let me rephrase my question then. Where in the documention is their non-portability pointed out? 20160721 06:41:09< iceiceice> i'd rather just take the documentation at face value until someone authoritative says otherwise 20160721 06:41:27< iceiceice> i'll find a link, one sec 20160721 06:41:33< shadowm> Mind you, I'm not questioning your argument, I'm merely curious. 20160721 06:42:54< iceiceice> https://www.gnu.org/software/gettext/manual/html_node/Files.html#Files 20160721 06:42:59< shadowm> Also, to my knowledge WesCamp never produced more than one .mo file per add-on/target language pair. 20160721 06:43:14< shadowm> So WesCamp may have already been broken to begin with. 20160721 06:43:17< iceiceice> i see 20160721 06:43:23< iceiceice> i guess i have only the vaguest ideas how wescamp actually worked 20160721 06:43:32< iceiceice> how did it actually distribute the mo's ? 20160721 06:43:35< iceiceice> was it linked into the add-on server? 20160721 06:43:37< shadowm> It's secondary function, generating the .po files, however, was not broken and was actually quite welcome. 20160721 06:43:48< shadowm> Or primary, rather. 20160721 06:47:00-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev 20160721 06:47:15< shadowm> The last paragraph is actually very vague. 20160721 06:48:07< shadowm> "The format of [the MO files generated by a few systems' native toolchains] is [...] non-portable. [...] Therefore GNU gettext uses its own format for MO files." 20160721 06:50:45-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev 20160721 06:51:17< iceiceice> idk 20160721 06:51:21< iceiceice> it seems to say its not portable 20160721 06:51:30< iceiceice> also there is another section that describes the mo format 20160721 06:51:32-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 260 seconds] 20160721 06:51:39< iceiceice> and it seems to say that you should store integers directly as binary 20160721 06:51:41< shadowm> Re WesCamp, how it was expected to work (you will let me finish) was as follows: add-ons get automatically added to the WesCamp repository, their po catalogues are automatically generated, translators submit updates to those, the mo files are regenerated and the add-ons on the add-ons server get the updated mo files. The content creators are completely passive participants in this process and ... 20160721 06:51:43< Rhonda> It says the original mo is not portable. The one produced by gnu gettext is. 20160721 06:51:47< shadowm> ... nothing is required from them beyond setting up l10n support in their add-ons (#textdomain, [textdomain], etc.). 20160721 06:52:02< iceiceice> Rhonda, so if it is portable, 20160721 06:52:08< iceiceice> then why do we have to build msgfmt into the build system 20160721 06:52:17< iceiceice> why cant pot-update simply also run msgfmt 20160721 06:52:21< iceiceice> and commit all the mos to the repostory 20160721 06:52:22< shadowm> Now I will tell you how it worked before AI0867's takeover, back when I could actually be bothered to deal with translations. 20160721 06:52:24< Rhonda> Because .po isn't .mo? 20160721 06:52:37< Rhonda> That's a quite silly and sidetracking question somehow. 20160721 06:52:43< iceiceice> it significantly complicates the build system that allt he users have to install msgfmt and run it 20160721 06:52:56< iceiceice> most C++ developers that i worked with believed that 20160721 06:53:01< iceiceice> the reason we have to do that is that .mo is not portable 20160721 06:53:10< Rhonda> Because .mo isn't human readable. It's Machine readable. And we need human readable files for tweaking. 20160721 06:53:29< iceiceice> yeah but it would be way simpler to just commit your .mo files to your repository 20160721 06:53:36< Rhonda> It doesn't complicate anything, that's a huge exaggeration. 20160721 06:53:37< iceiceice> and make your pot update script also run msgfmt 20160721 06:53:42< iceiceice> Rhonda, it does 20160721 06:53:54< Rhonda> It's part of basic toolchain. 20160721 06:53:57< iceiceice> it means that your build system has to handle finding and running msgfmt 20160721 06:54:02< iceiceice> otherwise it can simply invoke the C++ compiler 20160721 06:54:22< Rhonda> Repeating that .mo isn't portable without proof of that (and also proof of the contrary being done actually) doesn't make it true. 20160721 06:54:39< iceiceice> can you point to anyone authoritative saying that .mo is portable? 20160721 06:54:44< Rhonda> The build system handles *way* more than just a c++ compiler. 20160721 06:54:53< iceiceice> for many simple projects, it doesn't 20160721 06:55:02< Rhonda> wesnoth isn't a simple project. 20160721 06:55:12< iceiceice> the docs seem to say it isn't, and the established practice among developers indicates that it isn't 20160721 06:55:20< Rhonda> Can you please node sidetrack with statements that aren't relevant here? 20160721 06:55:21< iceiceice> also the conventional wisdom is that it is't 20160721 06:55:27< iceiceice> they are relevant 20160721 06:55:31< Rhonda> Your interpretation of the docs say that. 20160721 06:55:54< Rhonda> It's not relevant to speak of simple projects. 20160721 06:56:02< iceiceice> there are many other sections in the docs that talk about it 20160721 06:56:34< shadowm> I'm not finished writing my explanation of WesCamp, but wouldn't the easiest way to answer the question of whether .mo files *as generated by GNU gettext* are portable or not be to ask whatever mailing list there is for GNU gettext? 20160721 06:56:42< iceiceice> again if you can point to anyone authoritative saying ".mo files produces by msgfmt are portable" in unambiguous language i would love to see it 20160721 06:56:46< shadowm> Probaly somethingsomething@gnu.org. 20160721 06:57:50< shadowm> Add-ons get manually added to the WesCamp repository, their po catalogues are automatically generated by a person running make whatever as part of the same step. Thus far it doesn't seem too different from the ideal scenario above, right? Well, now translators submit updates to those, and maybe get in arguments with the mainline translation teams as to whether or not do things a certain way or ... 20160721 06:57:56< shadowm> ... not (or maybe work completely independently of the mainline translation team and produce translations that are inconsistent with the mainline style). Anyway, once translations are committed, someone has to compile them, but that someone in my experience was usually me, the add-on maintainer. There was no mechanism in place to automatically update the add-ons on the add-ons server with the ... 20160721 06:58:02< shadowm> ... updated translations -- or more generally, to automatically update add-ons in the server without the maintainer's intervention. 20160721 06:58:43< shadowm> Then the WesCamp maintainer would have to manually update the add-on on WesCamp the next time it was updated on the add-ons server and the cycle repeats. 20160721 06:58:49< Rhonda> iceiceice: Why do you think Debian ships .mo files in arch:all packages? Why do you think .mo files are stored in /usr/share instead of /usr/lib? Yes, because they are architecture independent. That's why. 20160721 07:00:05< shadowm> After AI0867's takeover I believe the add-to-WesCamp/update-to-WesCamp steps were automated but the rest of my description still applied, as far as I remember (he took over during a period of time when I couldn't keep track of translation updates anymore for unrelated reasons). 20160721 07:00:37 * Rhonda heads to work. 20160721 07:00:43< iceiceice> Rhonda, interesting, thanks 20160721 07:01:00< shadowm> The most important thing I want people to get out of this is the point that WesCamp's focus was the po files, not the mo files. 20160721 07:01:02< iceiceice> shadowm, sorry, i will read your exposition of wescamp now 20160721 07:05:47< shadowm> And sure, generating po files isn't rocket science if 1) you can use the command line, or write your own Makefile (and use make); or 2) have the patience to figure out the underdocumented wmlxgettext tool and its proper usage in a UMC context; but these aren't things that can be expected from most add-ons creators, are they? Ultimately you'll find that regardless of whether we transition to po ... 20160721 07:05:53< shadowm> ... file parsing, add-on translations are still fundamentally broken. 20160721 07:06:24< iceiceice> i think there was someone working on a new wmlgettext 20160721 07:06:40< iceiceice> Elvish_Hunter repeatedly mentioned that nobun was doing this 20160721 07:06:45< iceiceice> i seem to recall 20160721 07:06:46< shadowm> THe same points would continue to apply to it. 20160721 07:06:53< shadowm> Except for the underdocumented part, granted. 20160721 07:06:57< iceiceice> hehe 20160721 07:06:59< iceiceice> idk i mean 20160721 07:07:29< iceiceice> well 20160721 07:07:36< shadowm> Although if someone wrote an easy-to-use GUI for it that'd probably solve most of the problem. 20160721 07:07:37-!- Kwandulin [~Miranda@p200300760F6062D03D45B796F8FFF196.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer] 20160721 07:07:37< iceiceice> i guess you are right in that sense 20160721 07:07:43< iceiceice> if in fact .mo files are portable, 20160721 07:07:59< iceiceice> then anyone who can handle running wmlgettext over their whole add-on 20160721 07:08:05< iceiceice> probably can also hanlde running msgfmt 20160721 07:08:30< iceiceice> idk 20160721 07:08:43< iceiceice> the early version of spirit-po was actually just a little po parser routine in anura 20160721 07:08:57< shadowm> I belive poedit (a GUI po editor) has or had back in the day (when I was a translator back in 2007) an option to compile the currently open catalogue. 20160721 07:09:06< iceiceice> dave was a fan of the idea, i think he said something like "binary formats are 1995" 20160721 07:09:46< iceiceice> but yeah, i was pretty sure they weren't portable, so i'm surprised to hear that they actually are 20160721 07:10:07< iceiceice> i still think its better to read po at runtime though, for most games 20160721 07:15:37-!- RatArmy [~RatArmy@133.15.175.65] has quit [Ping timeout: 252 seconds] 20160721 07:28:00-!- molt [~molt@46.161.114.253] has joined #wesnoth-dev 20160721 07:43:21-!- Kwandulin [~Miranda@p200300760F6062D09DE390F60816DC77.dip0.t-ipconnect.de] has joined #wesnoth-dev 20160721 07:44:51-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has quit [Ping timeout: 264 seconds] 20160721 07:45:52-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 252 seconds] 20160721 07:50:06-!- Bonobo [~Bonobo@2001:44b8:254:3200:c534:198a:eeb3:5bba] has joined #wesnoth-dev 20160721 07:50:16-!- Kwandulin [~Miranda@p200300760F6062D09DE390F60816DC77.dip0.t-ipconnect.de] has quit [Ping timeout: 272 seconds] 20160721 07:54:01-!- boucman_work [~boucman@229.29.205.77.rev.sfr.net] has joined #wesnoth-dev 20160721 07:56:04-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev 20160721 08:09:08-!- Kwandulin [~Miranda@p200300760F6062D09DE390F60816DC77.dip0.t-ipconnect.de] has joined #wesnoth-dev 20160721 08:30:32-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev 20160721 08:40:19-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 252 seconds] 20160721 08:44:18-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev 20160721 08:49:00-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 276 seconds] 20160721 08:49:00-!- wedge010 is now known as wedge009 20160721 09:02:25-!- Kwandulin [~Miranda@p200300760F6062D09DE390F60816DC77.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer] 20160721 09:12:58-!- molt [~molt@46.161.114.253] has quit [Quit: Leaving] 20160721 09:21:02< zookeeper> what the... 20160721 09:21:27< zookeeper> https://wiki.wesnoth.org/ServerAdministration <- so, there's "/query ban :" there 20160721 09:21:47< zookeeper> but if you go edit the page, the code actually says "'''/query ban