Forum: Ruby-core minimize regressions for upcoming 2.0.0 release

D469019a435a5dbd5c792016fdf67a0b?d=identicon&s=25 Jon Forums (jonm)
on 2013-02-10 18:35
(Received via mailing list)
With the 2.0.0 release rapidly approaching I ask that each core
contributor closely monitor

  http://ci.rubyinstaller.org/view/Ruby%20Trunk/

to more quickly catch and fix regressions that might affect 2.0.0 on
windows platforms.

While the current ruby_2_0_0 branch is green for both 32 and 64bit

  http://ci.rubyinstaller.org/view/Ruby%202.0.0/

I'm concerned that trunk regressions might make it into the ruby_2_0_0
branch especially if the regressions are related to build tools rather
than core code. For example, this trunk regression:

  http://ci.rubyinstaller.org/view/Ruby%20Trunk/job/...


I haven't looked into the regression yet, but I suspect recent changes
to `configure.in` affecting `config.h`:

--- \rubytrunk\include\ruby-2.0.0\i386-mingw32\ruby\config.h
2013-02-07 21:33:37.000000000 -0500
+++ v:\sandbox\ruby19_build\.ext\include\i386-mingw32\ruby\config.h
2013-02-10 11:25:37.700053200 -0500
@@ -18,6 +18,22 @@
 #define _GNU_SOURCE 1
 #define _POSIX_PTHREAD_SEMANTICS 1
 #define _TANDEM_SOURCE 1
+#define HAVE_DIRENT_H 1
+#define HAVE__BOOL 1
+#define HAVE_STDBOOL_H 1
+#define HAVE_LIMITS_H 1
+#define HAVE_SYS_FILE_H 1
+#define HAVE_FCNTL_H 1
+#define HAVE_SYS_FCNTL_H 1
+#define HAVE_SYS_TIME_H 1
+#define HAVE_SYS_PARAM_H 1
+#define HAVE_UTIME_H 1
+#define HAVE_DIRECT_H 1
+#define HAVE_SYS_UTIME_H 1
+#define HAVE_FLOAT_H 1
+#define HAVE_LOCALE_H 1
+#define HAVE_TIME_H 1
+#define HAVE_PROCESS_H 1
 #define _FILE_OFFSET_BITS 64
 #define HAVE_LONG_LONG 1
 #define HAVE_OFF_T 1
@@ -77,19 +93,6 @@
 #define UNREACHABLE __builtin_unreachable()
 #define RUBY_FUNC_EXPORTED __attribute__ ((visibility("default")))
extern
 #define HAVE_DECL_SYS_NERR 1
-#define HAVE_DIRENT_H 1
-#define HAVE__BOOL 1
-#define HAVE_STDBOOL_H 1
-#define HAVE_LIMITS_H 1
-#define HAVE_SYS_FILE_H 1
-#define HAVE_FCNTL_H 1
-#define HAVE_SYS_FCNTL_H 1
-#define HAVE_DIRECT_H 1
-#define HAVE_SYS_UTIME_H 1
-#define HAVE_FLOAT_H 1
-#define HAVE_LOCALE_H 1
-#define HAVE_TIME_H 1
-#define HAVE_PROCESS_H 1
 #define SIZEOF_SIZE_T 4
 #define SIZEOF_PTRDIFF_T 4
 #define HAVE_STRUCT_STAT_ST_RDEV 1

Jon

---
Fail fast. Fail often. Fail publicly. Learn. Adapt. Repeat.
http://thecodeshop.github.com | http://jonforums.github.com/
twitter: @jonforums
D469019a435a5dbd5c792016fdf67a0b?d=identicon&s=25 Jon Forums (jonm)
on 2013-02-11 05:44
(Received via mailing list)
> I'm concerned that trunk regressions might make it into the ruby_2_0_0 branch
especially if the regressions are related to build tools rather than core code.
For example, this trunk regression:
>
>
http://ci.rubyinstaller.org/view/Ruby%20Trunk/job/...
>
>
> I haven't looked into the regression yet, but I suspect recent changes to
`configure.in` affecting `config.h`:

Thank you @akr for fixing with trunk@39200.  It would have taken me
awhile to spelunk the fix since my knowledge of `configure.in` is very
basic.

The build is OK on my Win7 32bit system, and I see ci.rubyinstaller.org
also `make all test` OK for 32 and 64bit.

Jon
This topic is locked and can not be replied to.