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.
This allows to catch the case where the user generate the dictionary with the top level header (which was the only one recorded before)
and the user #include the lower level header first.
This will also rootcling to tell cling's ForwardDeclPrinter to consider the file included inside a LinkDef file as
being top level headers (rather than listing the linkdef itself\!).
This reverts commit 125a93e0e3562dac5d5a0429ce49a2b625025c26.
It fails in the following case (roottest/root/io/evolution/pragma_read):
dictionary for a/b.h which includes a/c.h as #include "c.h"
This will add a fwd decl with annotation "c.h", but this header cannot
be found. Instead we'd need "a/c.h"...
Writing the template parameters might cause visitation of other decls which need to be in their respective
namespaces. Thus we need to buffer / delay the whole template decl, including its enclosing namespaces.
As we forward e.g. from ClassTemplateInstanceDecl() to RedeclarableTemplateDecl() we need to
selectively write the enclosing namespaces, no just for any Visit(Decl*). Make sure that a template pattern
is not again writing its enclosing scopes - it is just called from its VisitRedeclarableTemplateDecl().
STL contains a lot of forwarding headers. For example map forwards to bits/stl_map.
However we still need to annotate the decls that came from stl_map as if they
came from map, because semantically we would like to autoload/show hints about
map and not some 'interesting' stl implementation.
This is why we need to reconstruct the include chain and write out the second
includer. The first one is the prompt's input_line_N.
