| Time |
Nick |
Message |
| 00:08 |
|
YuGiOhJCJ joined #minetest-dev |
| 00:37 |
|
Noisytoot joined #minetest-dev |
| 01:11 |
|
Noisytoot joined #minetest-dev |
| 04:00 |
|
MTDiscord joined #minetest-dev |
| 04:21 |
|
fluxionary joined #minetest-dev |
| 05:02 |
|
calcul0n joined #minetest-dev |
| 05:10 |
|
v-rob joined #minetest-dev |
| 05:11 |
v-rob |
A question if anyone's knowledgeable on it: what happens if a formspec string exceeds 65535, the maximum protocol length? Does it get sent as multiple packets, or is it just truncated? |
| 05:11 |
v-rob |
Packets is the wrong word there... |
| 05:24 |
luk3yx |
I think formspecs are sent as "long" strings |
| 05:25 |
luk3yx |
With a 32-bit size instead of 16-bit |
| 05:27 |
luk3yx |
But long strings have a limit of 64MB |
| 06:13 |
|
Thomas-S joined #minetest-dev |
| 08:01 |
|
Noisytoot_ joined #minetest-dev |
| 09:34 |
|
Fixer joined #minetest-dev |
| 10:13 |
|
YuGiOhJCJ joined #minetest-dev |
| 10:17 |
|
YuGiOhJCJ joined #minetest-dev |
| 10:37 |
|
Fleckenstein joined #minetest-dev |
| 10:45 |
|
vampirefrog joined #minetest-dev |
| 11:20 |
|
proller joined #minetest-dev |
| 12:04 |
|
Fleckenstein joined #minetest-dev |
| 12:11 |
Zughy[m] |
reminder that there are 18 one approval PRs and 3 two approval ones |
| 12:19 |
|
TheCoffeMaker joined #minetest-dev |
| 13:35 |
sfan5 |
split packets are automatically used if you try to send a standard reliable packet that's too long |
| 13:36 |
sfan5 |
so as long as the serialization doesn't ruin it? |
| 14:15 |
|
appguru joined #minetest-dev |
| 14:24 |
rubenwardy |
the 16-bit string thing I believe is due to a u16 used to denote the string length - pascal style |
| 14:25 |
rubenwardy |
rather than something to do with packet fragmentation |
| 14:46 |
|
hmmmm joined #minetest-dev |
| 15:57 |
|
Desour joined #minetest-dev |
| 16:13 |
|
Fleckenstein joined #minetest-dev |
| 16:21 |
|
v-rob joined #minetest-dev |
| 16:34 |
v-rob |
As I said, packet was the wrong word. I meant message, or whatever that thing should be called. The entire formspec string is a message, then. I know we do have 65535 limits elsewhere (for instance, modchannels are limited to that for their message size) and hence my wondering about formspecs. |
| 16:40 |
v-rob |
Then again, doesn't LuaJIT limit strings to 65535? In that case, it's unlikely that any mod goes over the limit anyway then. |
| 16:41 |
sfan5 |
it doesn't |
| 16:42 |
v-rob |
I must be confusing it with some other limit. Thought I remembered that. |
| 16:42 |
v-rob |
(It does sound like an unreasonable limit though) |
| 16:48 |
MTDiscord |
<luatic> the constant limit? |
| 16:51 |
Desour |
there might be also a limit in how long the "formspec" node meta field can be |
| 16:54 |
v-rob |
Well, node meta is less important in this case. It's not a good idea to be storing large formspecs in them because changing the formspec becomes hard later when it's sprinkled all throughout the world. |
| 16:54 |
v-rob |
The reason I'm asking is because JSON is less compact than formspec strings, and it's probably easier to hit a 65535 limit, supposing there is one. |
| 16:56 |
Desour |
(just checked btw, nodemeta string values are limited by the same 64MiB LONG_STRING_MAX_LEN limit) |
| 16:57 |
rubenwardy |
could do BJSON |
| 16:57 |
rubenwardy |
but I'd worry about that later |
| 16:59 |
v-rob |
It's definitely something I'm kicking down the road, but I was curious if anyone already knew the answer. |
| 17:07 |
|
appguru joined #minetest-dev |
| 20:07 |
|
Alias joined #minetest-dev |
| 21:06 |
|
Fleckenstein joined #minetest-dev |
| 21:13 |
|
v-rob joined #minetest-dev |
| 22:26 |
|
diceLibrarian joined #minetest-dev |
| 22:33 |
|
panwolfram joined #minetest-dev |
| 22:55 |
|
v-rob joined #minetest-dev |
| 23:11 |
|
AliasAlreadyTake joined #minetest-dev |
