| CC: | sqlite-users [...] sqlite.org |
| Subject: | Re: [PATCH] Won't compile on archaic OS |
| Date: | Tue, 21 Jun 2011 13:29:41 +0200 |
| To: | "Bugs in DBD-SQLite via RT" <bug-DBD-SQLite [...] rt.cpan.org> |
| From: | "H.Merijn Brand" <h.m.brand [...] xs4all.nl> |
On Mon, 23 May 2011 18:45:14 +0200, "Bugs in DBD-SQLite via RT"
<bug-DBD-SQLite@rt.cpan.org> wrote:
Show quoted text
--
H.Merijn Brand http://tux.nl Perl Monger http://amsterdam.pm.org/
using 5.00307 through 5.14 and porting perl5.15.x on HP-UX 10.20, 11.00,
11.11, 11.23 and 11.31, OpenSuSE 10.1, 11.0 .. 11.4 and AIX 5.2 and 5.3.
http://mirrors.develooper.com/hpux/ http://www.test-smoke.org/
http://qa.perl.org http://www.goldmark.org/jeff/stupid-disclaimers/
> See also
> https://rt.cpan.org/Ticket/Display.html?id=68396
This (rather clear) bug report neither got any acknowledgement, not
does it appear in the RT trackers.
Meanwhile things got even worse.
Show quoted text> Using gcc-3.4.6 on HP-UX 10.20:
HP-UX 10.20 does not have pread at all. -UUSE_PREAD doesn't work
(enough) as the test is (very) wrong. SQLITE_ENABLE_LOCKING_STYLE is
overruled in this section when passed as -USQLITE_ENABLE_LOCKING_STYLE
--8<--- line 23983
#if !defined(SQLITE_ENABLE_LOCKING_STYLE)
# if defined(__APPLE__)
# define SQLITE_ENABLE_LOCKING_STYLE 1
# else
# define SQLITE_ENABLE_LOCKING_STYLE 0
# endif
#endif
-->8---
And thus leaving
--8<--- line 24453
#if defined(USE_PREAD) || defined(SQLITE_ENABLE_LOCKING_STYLE)
{ "pread", (sqlite3_syscall_ptr)pread, 0 },
#else
{ "pread", (sqlite3_syscall_ptr)0, 0 },
#endif
-->8---
a useless and bogus test.
That second test (also for pwrite) should either test for the *content*
of SQLITE_ENABLE_LOCKING_STYLE or not test on it at all
Show quoted text> > make test
> gcc -c -I. -I/opt/perl/lib/site_perl/5.14.0/PA-RISC2.0/auto/DBI -mpa-
> risc-1-1 -DPERL_DONT_CREATE_GVSV -D_HPUX_SOURCE -fno-strict-aliasing -
> pipe -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -O2 -DVERSION=
> \"1.31\" -DXS_VERSION=\"1.31\" -fPIC "-I/opt/perl/lib/5.14.0/PA-RISC2.0/
> CORE" -DSQLITE_ENABLE_FTS3 -DSQLITE_ENABLE_FTS3_PARENTHESIS -
> DSQLITE_ENABLE_COLUMN_METADATA -DNDEBUG=1 -DHAVE_USLEEP=1 -
> DTHREADSAFE=0 -DSQLITE_OMIT_LOAD_EXTENSION sqlite3.c
> sqlite3.c: In function `unixShmMap':
> sqlite3.c:26184: error: `MAP_FAILED' undeclared (first use in this
> function)
> sqlite3.c:26184: error: (Each undeclared identifier is reported only
> once
> sqlite3.c:26184: error: for each function it appears in.)
>
> After applying this patch:
> --8<---
> --- sqlite3.c.org 2011-05-23 15:27:35 +0200
> +++ sqlite3.c 2011-05-23 15:29:55 +0200
> @@ -22661,6 +22661,10 @@ struct unixFile {
> */
> #define SQLITE_WHOLE_FILE_LOCKING 0x0001 /* Use whole-file locking */
>
> +#ifndef MAP_FAILED
> +#define MAP_FAILED ((void *) -1)
> +#endif
> +
> /*
> ** Include code that is common to all os_*.c files
> */
> -->8---
>
> > make test
> :
> All tests successful.
> Files=59, Tests=1308, 87 wallclock secs ( 4.61 usr 0.48 sys + 53.42
> cusr 3.96 csys = 62.47 CPU)
> Result: PASS
> >