[BUG] crash in Lua on table insertion

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
26 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Re: [BUG] crash in Lua on table insertion

Viacheslav Usov
On Wed, May 16, 2018 at 6:07 PM, Albert Chan <[hidden email]> wrote:

> Tried on Lua 5.3.4, Lua 5.4.0 work1, LuaJIT 1.1.8.

I am not asking for a confirmation. The bug is real and is confirmed by the Lua team. You can read the entire old email thread if you want details.

Cheers,
V.
Reply | Threaded
Open this post in threaded view
|

Re: [BUG] crash in Lua on table insertion

Roberto Ierusalimschy
In reply to this post by Viacheslav Usov
>  The below has been listed as a bug in 5.3.4
> https://www.lua.org/bugs.html#5.3.4-2 for over a year, and there is no
> official fix. What are the plans for this bug? Another customer has run
> into this bug in the mean time...

Did you try this code with the Lua 5.4 work version?

-- Roberto

Reply | Threaded
Open this post in threaded view
|

Re: [BUG] crash in Lua on table insertion

Viacheslav Usov
On Wed, May 16, 2018 at 10:32 PM, Roberto Ierusalimschy <[hidden email]> wrote:

> Did you try this code with the Lua 5.4 work version?

Lua 5.4.0 (work1)  Copyright (C) 1994-2018 Lua.org, PUC-Rio
> t = {} local t = t for i = 1, 0x7fffffff do t[i] = i end
> print(#t)
2147483647
> t = {} local t = t for i = 1, 0x7ffffffff do t[i] = i end
stdin:1: table overflow
stack traceback:
        stdin:1: in main chunk
        [C]: in ?
> print(#t)
3221225472
>

Cheers,
V.
Reply | Threaded
Open this post in threaded view
|

Re: [BUG] crash in Lua on table insertion

Viacheslav Usov
In reply to this post by Roberto Ierusalimschy
On Wed, May 16, 2018 at 10:32 PM, Roberto Ierusalimschy <[hidden email]> wrote:
>  The below has been listed as a bug in 5.3.4
> https://www.lua.org/bugs.html#5.3.4-2 for over a year, and there is no
> official fix. What are the plans for this bug? Another customer has run
> into this bug in the mean time...

Did you try this code with the Lua 5.4 work version?

I am sorry for being insistent, but I need to decide whether we should wait for 5.3.5 (or 5.4) that will have this bug fixed, or try to roll our own patch from the current 5.4 work sources. Any input will be highly appreciated, but above all an indication of when 5.3.5 or 5.4 are likely to be released would be very helpful.

Thanks,
V.
Reply | Threaded
Open this post in threaded view
|

Re: [BUG] crash in Lua on table insertion

Roberto Ierusalimschy
> > >  The below has been listed as a bug in 5.3.4
> > > https://www.lua.org/bugs.html#5.3.4-2 for over a year, and there is no
> > > official fix. What are the plans for this bug? Another customer has run
> > > into this bug in the mean time...
> >
> > Did you try this code with the Lua 5.4 work version?
> >
>
> I am sorry for being insistent, but I need to decide whether we should wait
> for 5.3.5 (or 5.4) that will have this bug fixed, or try to roll our own
> patch from the current 5.4 work sources. Any input will be highly
> appreciated, but above all an indication of when 5.3.5 or 5.4 are likely to
> be released would be very helpful.

As 5.4 fixes the bug, we will see whether we can backport how it fixed from
it into 5.3. We need a few days to check that.

-- Roberto

Reply | Threaded
Open this post in threaded view
|

Re: [BUG] crash in Lua on table insertion

szbnwer@gmail.com
hi there! :)


what about backporting bugfixes down to 5.2.x and 5.1.x? (i think 5.0
users are already very rare, but do whatever you want :D )

those are still widely used, and cuz lj (with mainly 5.1
compatibility) is a basic entrance for the lua world for so much ppl
out there. if im right, then it is the fastest scripting engine and
jit out there, and i believe right that is the point where much
newbies' interest get caught. mayb this would even make lua a bit less
dreaded, and i think there are even some brave volunteers in the
community who would like to make it, if only the extra work is against
making it. ive recently seen somewhere that linux is developed like
the head is controlled by Linus, and the rest is maintained by some
trusted others. this model could be nice to be seen here, and i think
it wouldnt make any mess, if it is really about bugfix backports. btw
i havent investigate into observing the buglist, so i dont really know
the importance, or if there is anything that cant be fixed without
breaking existing apps/libs. so im just leaving this idea here, as i
think it wouldnt harm anything/anyone to do it. im still not deep
enough in understanding the ecosystem to take this task for myself,
but bugs are kinda much well defined, so i dont think that it would be
basically a huge task.

its fine that anyone can take the opportunity and make an unofficial
fork, but wild forks can lead to mess, and fear to use them. so i
think it would be the best to keep such a fork near to the official
lua, and its enough to make it clear that its community maintained. i
can think about a difference in name, or a jump in the versioning,
like 5.1.100 that must be fully compatible, but none can not see it
without asking 'why is that?'. :D or it can take a different repo like
lua/luabackports or something like, or if there is a single pointer in
the official docs like 'such a beast exists <here>', that is kinda
acceptable for anyone who dont wanna take a deeper look in the
codebase, but still can use it with trust.


btw this importance of older versions is partially about apps/libs and
partially about lj. ive just recently followed up my incoming mails to
present times, so the 5.4-work related topics are kinda fresh for me.
actually i like the new features (above 5.1) that i can think about (i
see them only here, as im using currently only lj), however i always
have a great fear, that what if there will be an abyss in the
language, that cant be handled anymore the way like libs can say 'it
can be used with 5.1-5.3 + lj'. ive left behind back in time the
python world when it made the 2.x 3.x shift with that abyss, with much
tools on the one and much on the other side only, that i was
interested in. lua is running forward related to lj, and i believe it
would be beneficial to talk with Mike about aims, for the best
achievable sync between the paths of the 2 projects. i know well that
both of you (lua team and Mike) have strong visionz about your
roadmap, but i still dont think its impossible to keep the gap between
them as small as possible, even if there can be things one of you
wouldnt like to follow but the other one is pushing it. id really love
to see this, and i think much others would like it so :)


so now i leave these here in the best hope, but its all upon you, so
really, just do whatever you want!

(and sorry for going a bit far from the coast of the original topic,
ive felt like it can fit with a shoehorn, so i just hope u wont hate
me for it :D )


thx for everything, and all the bests for everyone around! :)

12