[OpenWrt-Devel] [LEDE-DEV] 18.03/4 -> GCC 5 or 7

[Matthias Schiffer]

On 03/09/2018 01:56 PM, Matthias Schiffer wrote:
> On 03/07/2018 06:49 PM, Lucian Cristian wrote:
>> testing r6395-6c194078db on a ramips Ralink RT5350 with 4.14 and GCC 7.3
>> gives this error in luci
>>
>> daemon.err uhttpd[873]: /usr/bin/lua: /usr/lib/lua/luci/debug.lua:6:
>> attempt to call a number value
>> daemon.err uhttpd[873]: stack traceback:
>> daemon.err uhttpd[873]:        /usr/lib/lua/luci/debug.lua:6: in main chunk
>> daemon.err uhttpd[873]:        [C]: in function 'require'
>> daemon.err uhttpd[873]:        /usr/lib/lua/luci/util.lua:8: in main chunk
>> daemon.err uhttpd[873]:        [C]: in function 'require'
>> daemon.err uhttpd[873]:        /usr/lib/lua/luci/config.lua:4: in main chunk
>> daemon.err uhttpd[873]:        [C]: in function 'require'
>> daemon.err uhttpd[873]: /usr/lib/lua/luci/cacheloader.lua:5: in main chunk
>> daemon.err uhttpd[873]:        [C]: in function 'require'
>> daemon.err uhttpd[873]:        /www/cgi-bin/luci:2: in main chunk
>> daemon.err uhttpd[873]:        [C]: ?
> 
> 
> I was able to reproduce this on ar71xx. Minimal testcase:
> 
>> s = 'foo'
>> print(s:sub(2))
> 
> fails, while
> 
>> print(s.sub(s, 2))
> 
> works correctly, even though the former is only syntactic sugar for the
> latter. FWIW, type(s) correctly states that s is a string. Looking into it now.
> 
> Matthias

Bug reported at https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84790

Matthias

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <http://lists.infradead.org/pipermail/openwrt-devel/attachments/20180309/a127f40a/attachment.sig>
-------------- next part --------------
_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel