This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: couple of fixes for cp-demangle.c
- To: zack at wolery dot cumb dot org, Gabriel Dos Reis <gdr at codesourcery dot com>
- Subject: Re: couple of fixes for cp-demangle.c
- From: Mark Mitchell <mark at codesourcery dot com>
- Date: Thu, 08 Jun 2000 22:44:23 -0700
- Cc: gcc-patches at gcc dot gnu dot org
- Organization: CodeSourcery, LLC
- References: <20000608175341.F16815@wolery.cumb.org>
>>>>> "Zack" == Zack Weinberg <zack@wolery.cumb.org> writes:
Zack> 1) The demangler doesn't know about TF and TJ special
Zack> function types. These names seem reasonable.
Good point. The former will actually go away with the new ABI -- it's
only used for the old ABI. When I (perhaps prematurely) ripped out
the old ABI mangling, but wanted to keep the old ABI layout, I had to
do something with this.
The TJ thing is what Per (thanks, Per!) described in his message about
the Java uses of name mangling. It's the typeinfo-like thingy.
Zack> 2) There was no space printed between "operator" and
Zack> "delete" in the demangling of "_Zdl". For consistency, I've
Zack> made there always be a space after "operator".
Gaby was objecting to the space. (Gaby, did the patch you sent me
handle `operator new' and `operator delete' correctly? I don't mind
spelling `operator==' that way, but `operator delete' definitely needs
the space. We could use isalnum to decide whether or not to add the
space.)
So, I think the consensus is that we should have a space *only* for
operators whose operator name starts with an alpha-numberic
character. Would you mind making this change?
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com