IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
Dump cling::Value within setNoAlloc routines because this is its lifespan,
otherwise gets deleted if nobody requested it. I.e it was created only for
value printing purposes.
Dump cling::Value outside setWithAlloc because the actual value is not put inside
until the call to ::new finishes, thus we need to do it outside, i.e in EvaluateInternal.
Switch on the ValueExtraction synthesizer even when 'just' value printing. We
depend on it. Must be factored out properly in one class.
Stop attaching the value printing template magic, which didn't work in some cases.
Update the ref file, because now there is better type information.
Adapt the user-defined printout functions (TDatime).
Now it provides an interface which can lookup library names. It can unload a
library and the implementation of loading and resolving a library name is much
simpler.
The extensions are greatly used by the unloading facilities in cling its users.
Iterate over the elements, calling the dtor on each.
Requires two more fields in the AllocatedValue to track the array size and
element offset.
Note that this is only implemented for constant size arrays; all others are expected to not be returned (and thus not stored) by value.
The use of vector forces an allocation for each call, something we don't need to
impose.
SmallVectorImpl doesn't expose the static alloc size; and our interfaces don't
care. Only the actual storage definition needs to define the size, no subsequent
reference.
This is triggered by emitting operator<< 'class std::__1::basic_ostream<char,
struct std::__1::char_traits<char> > &(int)' even though its marked as exported
in general. Now that clang and libc++ agree on the handling of exported
templates we can remove our work-around.
Remove trigger__cxa_atexit(), we will just remap until successful.
Remove unused dso parameter.
Instead of friending just provide a function.
Pass a transaction, the typesafe way.
We keep the FileEntry -> Transaction mapping so that next time there is .L
we can figure out that it was already loaded and unload it and reload it again.
Here we don't want to depend on the JIT runFunction, because of its limitations,
when it comes to return value handling. There it is not clear who provides the
storage and who cleans it up in a platform independent way.
Depending on the type we need to synthesize a call to cling:
0) void : do nothing;
1) enum, integral, float, double, referece, pointer types :
call to cling::internal::setValueNoAlloc(...);
2) object type (alloc on the stack) :
cling::internal::setValueWithAlloc
2.1) constant arrays:
call to cling::runtime::internal::copyArray(...)
We need to synthesize later (see RuntimeUniverse.h)
Wrapper has signature: void w(cling::StoredValueRef SVR)
case 1):
setValueNoAlloc(gCling, &SVR, lastExprTy, lastExpr())
case 2):
new (setValueWithAlloc(gCling, &SVR, lastExprTy)) (lastExpr)
case 2.1):
copyArray(src, placement, N)
The underlying issue is that the caller need to provide argument setup
and cleanup, which for 32bit is different than in 64, for example.
The approach we propose should be independent on the concrete platform
implementation, because we pass in the storage and depending on the
return result, we let cling runtime to deside whether and how to alloc
storage.
Currently only implemented for the STL collection.
This could eventually be used to allow customization
of which default argument to strip and which class template.
For now, it is only used to prevent the addition of template
default template argument to the STL collection ... which
we would any strip.
This fixes the secondary issue in ROOT-6020 (one class template
instance having more than one corresponding TClass due to the
inconsistency in the normalization (not adding vs not stripping)
InclusionDirective() callback tells us that an inclusion directive has been processed, allowing us to try to autoload classes using their header file name.
For example try to autoload TGClient (libGui) when seeing #include "TGClient.h"
This should prevent potential crash when executing macros using globals (e.g. gClient), with the following error:
ExecutionContext: use of undefined symbol 'gClient'!
Tab completion works using the changed (but ROOT 5 compatible) syntax .> A.txt.
.> now uses dup2() that exists on Windows, too, instead of ttyname().
Tnis enables unredirection even during root > out.txt - which did not work in ROOT5.
Multiple ".> X.txt" can be stacked; ".>" unredirects back to the previous one on the stack.
This got lost in 2b6ae0aed / 96a4437c8 where typedefs were handled.
Simplify logic (and thus shorten code).
Fixes error in roottest/root/tree/addresses:
Error in <TStreamerInfo::Build>: Embedded_objects, unknown type: EmbeddedTypedef::Embedded3 m_emb4
*Merge PreVisitDecl and PreVisitMacro and rename to CollectFilesToUncache.
*Add FID to the transaction storing the memory buffer file ID caused the
transaction.
They were once identical but then got bug fixed out of sync.
Moved routines:
GetFullyQualifiedLocalType()
GetFullyQualifiedType()
CreateNestedNameSpecifier()
CreateNestedNameSpecifierForScopeOf()
GetFullyQualifiedTypeNNS()
