SolidWallOfCode opened a new pull request, #9521:
URL: https://github.com/apache/trafficserver/pull/9521
With commit d07a6acc760350a44be6b47a532a2564080c5926 for #9474 and g++
12.2.2 on Fedora 36 I get compiler errrors, which is basically this repeatedly.
```
In file included from ink_inet.cc:30:
In destructor 'virtual ats::CryptoContext::~CryptoContext()',
inlined from 'virtual ats::CryptoContext::~CryptoContext()' at
../../include/tscore/CryptoHash.h:171:86,
inlined from 'virtual ats::CryptoContext::~CryptoContext()' at
../../include/tscore/CryptoHash.h:171:86,
inlined from 'virtual ats::CryptoContext::~CryptoContext()' at
../../include/tscore/CryptoHash.h:171:86,
inlined from 'virtual ats::CryptoContext::~CryptoContext()' at
../../include/tscore/CryptoHash.h:171:86,
inlined from 'virtual ats::CryptoContext::~CryptoContext()' at
../../include/tscore/CryptoHash.h:171:86,
inlined from 'virtual ats::CryptoContext::~CryptoContext()' at
../../include/tscore/CryptoHash.h:171:86,
inlined from 'virtual ats::CryptoContext::~CryptoContext()' at
../../include/tscore/CryptoHash.h:171:86,
inlined from 'virtual ats::CryptoContext::~CryptoContext()' at
../../include/tscore/CryptoHash.h:171:86,
inlined from 'virtual ats::CryptoContext::~CryptoContext()' at
../../include/tscore/CryptoHash.h:171:86,
inlined from 'uint32_t ats_ip_hash(const sockaddr*)' at
ink_inet.cc:381:5:
../../include/tscore/CryptoHash.h:171:90: error: array subscript
'ats::CryptoContext[0]' is partly outside array bounds of 'ats::CryptoContext
[1]' [-Werror=array-bounds]
171 | ~CryptoContext() { reinterpret_cast<CryptoContextBase
*>(_base)->~CryptoContextBase(); }
|
^
ink_inet.cc: In function 'uint32_t ats_ip_hash(const sockaddr*)':
ink_inet.cc:381:19: note: at offset 144 into object '<anonymous>' of size 272
381 | CryptoContext().hash_immediate(hash, ats_ip_addr8_cast(addr),
TS_IP6_SIZE);
|
```
My opinion is this is a compiler issue, that it's getting confused because
the wrapper type is a subclass of the cast type. This causes repeated inlining
of the same code, while `array bounds of 'ats::CryptoContext [1]` makes no
sense at all, given there is no array of that type.
This fixes the problem by having `CryptoContext` not inherit from the
protocol class. That inheritance is unnecessary and adds an extra virtual
function call for no benefit. At that point it seems better to move the
protocol class inside `CryptoContext` for better scoping.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
