paweld/fpc
1
0
Fork 0
mirror of https://gitlab.com/freepascal.org/fpc/source.git synced 2025-10-24 16:05:57 +02:00
Code Issues Packages Projects Releases Wiki Activity

* patch by Aleksa Todorovic to solve another issue with nested generics, resolves #17193

Browse Source 
git-svn-id: trunk@16496 -
This commit is contained in:
florian 2010-12-02 17:32:19 +00:00
parent 07e3232fef
commit 9dc584efbf
1 changed files with 17 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
compiler/ptype.pas
View File

@ -338,6 +338,23 @@ implementation
{ Restore symtablestack } { Restore symtablestack }
symtablestack.free; symtablestack.free;
symtablestack:=oldsymtablestack; symtablestack:=oldsymtablestack;
end
else
begin
{ There is comment few lines before ie 200512115
saying "We are parsing the same objectdef, the def index numbers
are the same". This is wrong (index numbers are not same)
in case there is specialization (S2 in this case) inside
specialized generic (G2 in this case) which is equal to
some previous specialization (S1 in this case). In that case,
new symbol is not added to currently specialized type
(S in this case) for that specializations (S2 in this case),
and this results in that specialization and generic definition
don't have same number of elements in their object symbol tables.
This patch adds undefined def to ensure that those
two symbol tables will have same number of elements.
}
tundefineddef.create;
end; end;
generictypelist.free; generictypelist.free;