IRC logs for #openttd.dev on OFTC at 2014-03-03
#openttd.dev / 2014 / 03 
            

06:20:32  *** Supercheese has quit IRC



14:14:01  *** dihedral has quit IRC



17:06:08  *** Alberth has joined #openttd.dev



17:06:08  *** ChanServ sets mode: +v Alberth



17:22:59  *** frosch123 has joined #openttd.dev



17:22:59  *** ChanServ sets mode: +v frosch123



18:45:21  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r26386 || Logs: http://webster.openttdcoop.org/?channel=openttd.dev || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"



19:32:54  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r26387 || Logs: http://webster.openttdcoop.org/?channel=openttd.dev || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"



20:02:31  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r26388 || Logs: http://webster.openttdcoop.org/?channel=openttd.dev || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"



20:03:13  <fonsinchen> https://github.com/ulfhermann/openttd/commit/cf81b7ddfe09d0d9e4411d760c02a784ffd8d7ce for FS#5885



20:03:23  <fonsinchen> I've tested on Arch linux: it works



20:04:32  *** Rubidium_ is now known as rubidium



20:04:40  *** ChanServ sets mode: +v rubidium



20:04:49  <rubidium> line 297 misses a space



20:05:11  <rubidium> 292 more or less too



20:05:23  <rubidium> what about the case when i == 1?



20:06:05  <rubidium> because then chosen remains 0, right?



20:07:37  <rubidium> oh, never mind... was not seeing the point calculation magic



20:08:03  <rubidium> having said that, what about sizeof(size_t) == 8; then 288 would give a warning/error, wouldn't it?



20:08:17  <fonsinchen> Does that happen?



20:08:37  <rubidium> the warning/error?



20:08:51  <fonsinchen> no sizeof(size_t) == 8



20:09:03  <fonsinchen> what good whould a size_t of size 8 be?



20:10:03  <frosch123> that's the normal case on 64bit systems



20:10:18  <frosch123> but i don't think there would be a warning



20:10:25  <fonsinchen> oh, 8 bytes :/



20:10:35  <frosch123> i have only seen warnings for pointer -> int conversions



20:10:45  <frosch123> but a pointer-difference is no pointer



20:11:18  <rubidium> but sizeof(ptrdiff_t) == sizeof(size_t), right?



20:11:26  <fonsinchen> Where does line 297 miss a space?



20:12:03  <fonsinchen> I've actually compiled it on 64bit and it didn't give an error



20:12:18  <frosch123> rubidium: i do not get a warning when assigning a uint32 to a byte either



20:12:29  <rubidium> fonsinchen: good question; seems like the monospace font in my browser is getting my head messed up



20:12:34  <rubidium> frosch123: I think it's MSVC



20:13:12  <rubidium> but just commit it without, then we'll see it quickly enough



20:13:22  <fonsinchen> num_fixed_sizes - i should do the trick, too



20:13:59  <michi_cc> rubidium: ptrdiff_t *can* be smaller than size_t, usually not for Windows or Linux though. Who knows about stuff like S390 though...



20:14:31  <rubidium> fonsinchen: it looks okay to me



20:15:51  <michi_cc> Hmm, actually, it can even be on Win/Linux, if you use the Intel C++ compiler and compile for x64 with small memory model. 64-bit registers and everything, but program ("near") memory space limited to 32 bit.



20:16:34  <fonsinchen> I'll post a new patch later



21:21:27  <fonsinchen> https://github.com/ulfhermann/openttd/commit/888b006b59e8e2a41656cc49abe6f078643363ae



21:21:40  <fonsinchen> This doesn't do pointer calculation there.



21:23:41  <frosch123> looks fine to me



21:34:36  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r26389 || Logs: http://webster.openttdcoop.org/?channel=openttd.dev || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"



21:54:39  *** frosch123 has quit IRC



22:29:33  *** Alberth has left #openttd.dev