From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Fri, 28 Dec 2018 04:46:50 +0300 From: Alexander Turenko Subject: Re: [PATCH 1/3] Add luaT_iscallable with support of cdata metatype Message-ID: <20181228014649.7v4zm24neusuqelu@tkn_work_nb> References: <20181226183556.gilqpkhj4gmliby4@esperanza> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20181226183556.gilqpkhj4gmliby4@esperanza> To: Vladimir Davydov Cc: tarantool-patches@freelists.org List-ID: Updated Totktonada/gh-3276-on-board-merger and Totktonada/gh-3276-on-board-merger-1.10. On Wed, Dec 26, 2018 at 09:35:56PM +0300, Vladimir Davydov wrote: > On Sun, Dec 16, 2018 at 11:17:24PM +0300, Alexander Turenko wrote: > > Needed for #3276. > > --- > > extra/exports | 1 + > > src/lua/utils.c | 43 ++++++++++++++++ > > src/lua/utils.h | 10 ++++ > > test/app-tap/module_api.c | 10 ++++ > > test/app-tap/module_api.test.lua | 85 +++++++++++++++++++++++++++++++- > > 5 files changed, 147 insertions(+), 2 deletions(-) > > > > diff --git a/extra/exports b/extra/exports > > index 5f69e0730..52f0b2378 100644 > > --- a/extra/exports > > +++ b/extra/exports > > @@ -134,6 +134,7 @@ luaT_call > > luaT_cpcall > > luaT_state > > luaT_tolstring > > +luaT_iscallable > > Do we really need to export it? No, we don't obligated to export it. I'll remove it and rewrite the test case (at least I'll try) if you sure we shouldn't. But I think it is better to leave it as the public function. I think luajit's support of cdata (in C/Lua APIs) could be better and we can invest some effort to make it better for users of our C API and ffi. Also filed #3915. > > > box_txn > > box_txn_begin > > box_txn_commit > > diff --git a/src/lua/utils.c b/src/lua/utils.c > > index 978fe61f1..7a6069fbb 100644 > > --- a/src/lua/utils.c > > +++ b/src/lua/utils.c > > @@ -920,6 +920,49 @@ luaT_tolstring(lua_State *L, int idx, size_t *len) > > return lua_tolstring(L, -1, len); > > } > > > > +/* Based on ffi_meta___call() from luajit/src/lib_ffi.c. */ > > +static int > > +luaT_cdata_iscallable(lua_State *L, int idx) > > I think this function should have prefix luaL_ rather than luaT_. > Other than that, this particular patch looks OK to me. > Fixed. > > +{ > > + /* Calculate absolute value in the stack. */ > > + if (idx < 0) > > + idx = lua_gettop(L) + idx + 1; > > + > > + /* Get cdata from the stack. */ > > + assert(lua_type(L, idx) == LUA_TCDATA); > > + GCcdata *cd = cdataV(L->base + idx - 1); > > + > > + CTState *cts = ctype_cts(L); > > + CTypeID id = cd->ctypeid; > > + CType *ct = ctype_raw(cts, id); > > + if (ctype_isptr(ct->info)) > > + id = ctype_cid(ct->info); > > + > > + /* Get ctype metamethod. */ > > + cTValue *tv = lj_ctype_meta(cts, id, MM_call); > > + > > + return tv != NULL; > > +}