Обсуждение: Gist does not build with VC++ anymore
I've updated my VC++ build env with latest CVS, and it no longer builds because of changes to GIST: src\backend\access\gist\gistutil.c(237) : error C2057: expected constant expression src\backend\access\gist\gistutil.c(237) : error C2466: cannot allocate an array of constant size 0 src\backend\access\gist\gistutil.c(237) : error C2133: 'storage' : unknown size The problem appears to come from: #define GEVHDRSZ (offsetof(GistEntryVector, vector[0])) Which can't be used in this context. What would be the proper fix for that? //Magnus
> I've updated my VC++ build env with latest CVS, and it no > longer builds because of changes to GIST: > > src\backend\access\gist\gistutil.c(237) : error C2057: > expected constant expression > src\backend\access\gist\gistutil.c(237) : error C2466: cannot > allocate an array of constant size 0 > src\backend\access\gist\gistutil.c(237) : error C2133: 'storage' : > unknown size > > > The problem appears to come from: > #define GEVHDRSZ (offsetof(GistEntryVector, vector[0])) > > Which can't be used in this context. > > What would be the proper fix for that? Hmm. Now that I look at it more clearly, it seems Hiroshi has a fix for this in his submitted patch (that still had a lot of other problems in the rest of it). I'm not sure if it's the proper fix, but it's there. Comments on it? --- src/backend/access/gist/gistutil.c.orig Thu Jun 22 20:46:55 2006 +++ src/backend/access/gist/gistutil.c Thu Jun 22 20:47:22 2006 @@ -228,6 +228,12 @@/* * makes union of two key */ + +#ifdef _MSC_VER +#undef GEVHDRSZ +#define GEVHDRSZ (offsetof(GistEntryVector, vector)) +#endif +static voidgistMakeUnionKey( GISTSTATE *giststate, int attno, GISTENTRY *entry1, bool isnull1, //Magnus
> > I've updated my VC++ build env with latest CVS, and it no longer > > builds because of changes to GIST: > > > > src\backend\access\gist\gistutil.c(237) : error C2057: > > expected constant expression > > src\backend\access\gist\gistutil.c(237) : error C2466: > cannot allocate > > an array of constant size 0 > > src\backend\access\gist\gistutil.c(237) : error C2133: 'storage' : > > unknown size > > > > > > The problem appears to come from: > > #define GEVHDRSZ (offsetof(GistEntryVector, vector[0])) > > > > Which can't be used in this context. > > > > What would be the proper fix for that? > > Hmm. Now that I look at it more clearly, it seems Hiroshi has > a fix for this in his submitted patch (that still had a lot > of other problems in the rest of it). I'm not sure if it's > the proper fix, but it's there. While I'm spamming everybody anyway, here's another thing that might fix it? This one compiles and tests, and I *think* it does the right thing... If it's correct, I think it looks like a cleaner solution. //Magnus RCS file: /projects/cvsroot/pgsql/src/include/access/gist.h,v retrieving revision 1.52 diff -c -r1.52 gist.h *** src/include/access/gist.h 5 Mar 2006 15:58:53 -0000 1.52 --- src/include/access/gist.h 24 Jun 2006 17:20:28 -0000 *************** *** 142,148 **** GISTENTRY vector[1]; } GistEntryVector; ! #define GEVHDRSZ (offsetof(GistEntryVector, vector[0])) /* * macro to initialize a GISTENTRY --- 142,148 ---- GISTENTRY vector[1]; } GistEntryVector; ! #define GEVHDRSZ (offsetof(GistEntryVector, vector)) /* * macro to initialize a GISTENTRY
Patch applied. It seems offsetof() can only find structure members in MSVC, not the offset of array elements in structures. Anyway, offsetof() was finding the first element, so the "[0]" is not needed. --------------------------------------------------------------------------- Magnus Hagander wrote: > > > I've updated my VC++ build env with latest CVS, and it no longer > > > builds because of changes to GIST: > > > > > > src\backend\access\gist\gistutil.c(237) : error C2057: > > > expected constant expression > > > src\backend\access\gist\gistutil.c(237) : error C2466: > > cannot allocate > > > an array of constant size 0 > > > src\backend\access\gist\gistutil.c(237) : error C2133: 'storage' : > > > unknown size > > > > > > > > > The problem appears to come from: > > > #define GEVHDRSZ (offsetof(GistEntryVector, vector[0])) > > > > > > Which can't be used in this context. > > > > > > What would be the proper fix for that? > > > > Hmm. Now that I look at it more clearly, it seems Hiroshi has > > a fix for this in his submitted patch (that still had a lot > > of other problems in the rest of it). I'm not sure if it's > > the proper fix, but it's there. > > While I'm spamming everybody anyway, here's another thing that might fix > it? This one compiles and tests, and I *think* it does the right > thing... If it's correct, I think it looks like a cleaner solution. > > //Magnus > > > RCS file: /projects/cvsroot/pgsql/src/include/access/gist.h,v > retrieving revision 1.52 > diff -c -r1.52 gist.h > *** src/include/access/gist.h 5 Mar 2006 15:58:53 -0000 1.52 > --- src/include/access/gist.h 24 Jun 2006 17:20:28 -0000 > *************** > *** 142,148 **** > GISTENTRY vector[1]; > } GistEntryVector; > > ! #define GEVHDRSZ (offsetof(GistEntryVector, vector[0])) > > /* > * macro to initialize a GISTENTRY > --- 142,148 ---- > GISTENTRY vector[1]; > } GistEntryVector; > > ! #define GEVHDRSZ (offsetof(GistEntryVector, vector)) > > /* > * macro to initialize a GISTENTRY > > ---------------------------(end of broadcast)--------------------------- > TIP 3: Have you checked our extensive FAQ? > > http://www.postgresql.org/docs/faq > -- Bruce Momjian bruce@momjian.us EnterpriseDB http://www.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +