| Time |
Nick |
Message |
| 00:16 |
Zughy |
hey there, working on a PR with a friend, then we saw this: https://github.com/minetest/minetest/blob/master/src/gui/guiFormSpecMenu.cpp#L4442 <= infostream. https://github.com/minetest/minetest/blob/master/src/gui/guiVolumeChange.cpp#L174 <= dstream. Is it normal? |
| 00:17 |
Zughy |
or should we change one of them (guiKeyChangeMenu uses dstream too)? |
| 00:51 |
|
Ritchie_ joined #minetest-dev |
| 01:54 |
|
Lunatrius joined #minetest-dev |
| 03:27 |
|
oil_boi joined #minetest-dev |
| 04:10 |
oil_boi |
For months I've been having multiple users tell me that my client mod was running on servers because "Issued command:" is loading into the client even when the server disallows it. Also it's extremely redundant and clutters the chat dialogue. |
| 04:10 |
oil_boi |
#10123 |
| 04:10 |
ShadowBot |
https://github.com/minetest/minetest/issues/10123 -- Stop sending "Issued command:" by oilboi |
| 04:10 |
oil_boi |
If a command doesn't exist it tells you anyways |
| 04:12 |
oil_boi |
Source, literally days of trying to figure out why users were telling me that the CSM was loading even though it does a double check to not load if not on the specific game mode |
| 06:39 |
|
absurb joined #minetest-dev |
| 07:40 |
|
calcul0n joined #minetest-dev |
| 08:00 |
|
ShadowNinja joined #minetest-dev |
| 08:08 |
|
NetherEran joined #minetest-dev |
| 08:14 |
|
NetherEran joined #minetest-dev |
| 08:17 |
|
proller joined #minetest-dev |
| 08:50 |
|
Qiangong2[m] joined #minetest-dev |
| 09:05 |
|
texmex joined #minetest-dev |
| 09:05 |
|
kb1000 joined #minetest-dev |
| 09:05 |
|
Awkor[m] joined #minetest-dev |
| 10:31 |
|
Zughy joined #minetest-dev |
| 10:40 |
|
Fixer joined #minetest-dev |
| 10:45 |
|
Fixer_ joined #minetest-dev |
| 11:31 |
|
appguru joined #minetest-dev |
| 11:51 |
|
sofar joined #minetest-dev |
| 13:50 |
|
proller joined #minetest-dev |
| 14:06 |
appguru |
How is set_eye_offset supposed to work? |
| 14:06 |
appguru |
I assumed that it was an eye offset in world coordinates, but this appears to not be the case. |
| 14:14 |
rubenwardy |
Probably *10 |
| 14:59 |
|
absurb joined #minetest-dev |
| 15:26 |
|
fluxflux joined #minetest-dev |
| 15:33 |
|
oiaohm joined #minetest-dev |
| 16:22 |
|
_Zaizen_ joined #minetest-dev |
| 16:45 |
|
oil_boi joined #minetest-dev |
| 17:48 |
|
NetherEran joined #minetest-dev |
| 18:04 |
|
appguru joined #minetest-dev |
| 18:59 |
|
lisac joined #minetest-dev |
| 19:30 |
|
proller joined #minetest-dev |
| 20:30 |
|
Anulo2 joined #minetest-dev |
| 20:31 |
sfan5 |
seems like -ffinite-math-only is what breaks std::isnan and consequently #10121 |
| 20:31 |
ShadowBot |
https://github.com/minetest/minetest/issues/10121 -- 'minetest.is_nan' is broken! |
| 20:32 |
sfan5 |
sad that you can't specify "assume there are no NaNs inside my calculations but allow me to test for NaN" to the compiler |
| 20:34 |
rubenwardy |
Could you check for specific float encodings? |
| 20:34 |
rubenwardy |
I forget how nan is encoded |
| 20:35 |
sfan5 |
there's different ones and doing that is also not very portable |
| 20:37 |
rubenwardy |
Thought that would be the case :/ |
| 20:56 |
pgimeno |
maybe function minetest.is_nan(arg) return arg ~= arg end ? |
| 20:57 |
pgimeno |
I'd check with PUC Lua too to be sure, though. And with JIT off. |
| 20:58 |
pgimeno |
NaN is encoded as exponent = max, mantissa = nonzero (one bit of the mantissa specifies if it's a signaling NaN; the sign bit specifies if it's an indeterminate form, though LuaJIT ignores the sign) |
| 21:03 |
sfan5 |
moving it to lua would work but the isNaN() function in C++ still wouldn't work |
| 21:03 |
sfan5 |
so you could smuggle a NaN into the calculations (despite checks) |
| 21:09 |
pgimeno |
yeah, -ffinite-math-only is too broad in scope |
| 21:17 |
appguru |
arg ~= arg should also apply in C++? |
| 21:17 |
appguru |
s/~=/!= |
| 21:17 |
rubenwardy |
not when -ffinite-math-only is enabled |
| 21:18 |
rubenwardy |
the compiler will optimise that out |
| 21:18 |
rubenwardy |
because it assumes that NaN will never happen when that flag is present |
| 21:19 |
appguru |
https://en.wikipedia.org/wiki/NaN |
| 21:20 |
appguru |
We could try other comparisons |
| 21:21 |
appguru |
How about "!(n <= std::numeric_limits<double>::min)" ? |
| 21:23 |
appguru |
s/::min/::min() |
| 21:52 |
|
erlehmann joined #minetest-dev |
| 22:21 |
|
skyliner_369 joined #minetest-dev |
| 23:49 |
|
Zughy joined #minetest-dev |
