They drop all existing libraries, frameworks and modules into the trashcan, by dropping module and package.seeall from Lua 5.2.
I don't think that this is a good decision, especially as several of them have no maintainer, no documentation, and forks on different websites.
One could also see this as a chance, to throw away all those badly documented libs that are without regression test, and restart a clean repository of modern Lua 5.2.
But I don't think so, when looking at current Lua code.
Lua's draw is not in its libraries (beyond a core set of very awesome GEMS). The stuff that is important is actively developed and will be ported forward if there is indeed an issue. Rarely can you go trolling for a library with a long line down rigger and find a good fish. Many libraries are out of date, the rocks system needs improvement. The addition of the bitops into core is very much welcomed.
These changes look really good. I would love to see someone make a CoffeeScript to Lua compiler.
My main area of concern is when the wonderful Mike Pall will port LuaJit to 5.2 when it is released.
Maybe in Lua 6 they can default to local vars and add list comprehensions.
I would also like to see a CoffeeScript compiler (it has a pretty good Haml compiler already), but more than that I'd like to see an LR parser generator so we can write those compilers more easily.
I have half of one now, it will take an already-built parser table and run it against input to make a tree, but I haven't had the time to finish it and make the actual generator.
6 comments
[ 2.9 ms ] story [ 25.6 ms ] threadAnd by the time I saw this comment, I couldn't edit the title. Apologies to everyone who was misled.
I don't think that this is a good decision, especially as several of them have no maintainer, no documentation, and forks on different websites.
One could also see this as a chance, to throw away all those badly documented libs that are without regression test, and restart a clean repository of modern Lua 5.2. But I don't think so, when looking at current Lua code.
The Lua codebase seems so incredibly stable that anyone married to it can probably live on that older branch.
I don't think that this is a good decision, especially as several of them have no maintainer, no documentation, and forks on different websites.
Aren't you almost obligated to drop support for something like that?
These changes look really good. I would love to see someone make a CoffeeScript to Lua compiler.
My main area of concern is when the wonderful Mike Pall will port LuaJit to 5.2 when it is released.
Maybe in Lua 6 they can default to local vars and add list comprehensions.
Lua ranks up there with SQLite in awesomeness.
I have half of one now, it will take an already-built parser table and run it against input to make a tree, but I haven't had the time to finish it and make the actual generator.