Re: Japanese, was: Re: [GENERAL] Code of Conduct: Is it time?
On понедельник, 11 января 2016 г. 12:24:37 MSK, Karsten Hilbert wrote: OTOH, there's a whole bunch of words denoting levels and sublevels of politeness for each and every situation. Politeness but not gender differences. Perhaps just for kids (-chan/-kun). -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Building PostgreSQL 9.6devel sources with Microsoft Visual C++ 2015?
Tom Lane wrote:
Somebody will have to research how one is
supposed to get the appropriate locale name now.
It's just a bug:
https://connect.microsoft.com/VisualStudio/feedback/details/1882835/locale-t-compile-issues-with-vs2015
I spied a solution there and made a patch (in attachment).
I came across this error when building Postgres using the CMake for the
MSVC 2015.
Thanks.
PS I do not know whether this patch needs to be added to commitfest?
--
Yury Zhuravlev
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Companydiff --git a/src/include/port.h b/src/include/port.h
index 9fc79f4..98846ea 100644
--- a/src/include/port.h
+++ b/src/include/port.h
@@ -468,4 +468,38 @@ extern char *escape_single_quotes_ascii(const char *src);
/* port/wait_error.c */
extern char *wait_result_to_str(int exit_status);
+#if _MSC_VER >= 1800
+ //From VS2012.
+ typedef struct localerefcount
+ {
+ char *locale;
+ wchar_t *wlocale;
+ int *refcount;
+ int *wrefcount;
+ } locrefcount;
+
+ //From VS2012.
+ typedef struct __crt_locale_data
+ {
+ int refcount;
+ unsigned int lc_codepage;
+ unsigned int lc_collate_cp;
+ unsigned int lc_time_cp;
+ locrefcount lc_category[6];
+ int lc_clike;
+ int mb_cur_max;
+ int * lconv_intl_refcount;
+ int * lconv_num_refcount;
+ int * lconv_mon_refcount;
+ struct lconv * lconv;
+ int * ctype1_refcount;
+ unsigned short * ctype1;
+ const unsigned short * pctype;
+ const unsigned char * pclmap;
+ const unsigned char * pcumap;
+ struct __lc_time_data * lc_time_curr;
+ wchar_t * locale_name[6];
+ } threadlocinfo;
+#endif
+
#endif /* PG_PORT_H */
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Building PostgreSQL 9.6devel sources with Microsoft Visual C++ 2015?
Tom Lane wrote: Ick. Even if that works today, it seems absolutely guaranteed to fail in future, as soon as Microsoft either puts back the visible declaration or changes the struct contents. If they've made a conscious decision to not export the struct anymore, it's likely because they intend to change it ... so I'd put the half-life of this "fix" at no more than one Visual Studio release. Yes. You right. But at the moment, it's better than nothing. In addition, we can then do something like this: #if _MSC_VER >= 1800 && _MSC_VER < 1820 after MS push fix. Hopefully, if they removed the visible declaration intentionally, they provided some other way to get at those locale names. That's what we need to be looking for, not hoping that direct access to undocumented structures will continue to work. It's more like a simple bug after refactoring. But I will try find another way. (I don't like undocumented structures) Thanks. -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Building PostgreSQL 9.6devel sources with Microsoft Visual C++ 2015?
Michael Paquier wrote: How long do you think it would take for MS 1820 to be fixed and out? Maybe never. I wouldn't personally mind telling to people trying to compile with 1800 that we cannot support it because it is buggy. That's one less wart to have forever in the code. For the user, this is a bad response. In addition, many of the new features msvc, due to which the user must use it. Also, I think the current code looks like a hack. It's okay if we for some time to add one more hack. But as I wrote above, I will try to find a better solution. While I was not going to back down: CMake+MSVC2015 Thanks. -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Building PostgreSQL 9.6devel sources with Microsoft Visual C++ 2015?
Please look at the new patch. It is filled with black magic, but it looks
still more true.
He agreed with the internal API.
--
Yury Zhuravlev
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Companydiff --git a/src/port/chklocale.c b/src/port/chklocale.c
index a551fdc..6113424 100644
--- a/src/port/chklocale.c
+++ b/src/port/chklocale.c
@@ -203,7 +203,7 @@ win32_langinfo(const char *ctype)
{
char *r = NULL;
-#if (_MSC_VER >= 1700)
+#if (_MSC_VER >= 1700 && _MSC_VER < 1800)
_locale_t loct = NULL;
loct = _create_locale(LC_CTYPE, ctype);
@@ -214,7 +214,20 @@ win32_langinfo(const char *ctype)
sprintf(r, "CP%u", loct->locinfo->lc_codepage);
_free_locale(loct);
}
+#elif (_MSC_VER >= 1800)
+ _locale_t loct = NULL;
+
+ loct = _create_locale(LC_CTYPE, ctype);
+ if (loct != NULL)
+ {
+ __crt_locale_data_public* public_loct = __acrt_get_locale_data_prefix(loct);
+ r = malloc(16); /* excess */
+ if (r != NULL)
+ sprintf(r, "CP%u", public_loct->_locale_lc_codepage);
+ _free_locale(loct);
+ }
#else
+
char *codepage;
/*
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
[GENERAL] Test CMake build
Hello all. Please test build Postgres using cmake. If you are of course interested. Still not everything is ready but most of the work. Assembly instructions as does the repository is on github: https://github.com/stalkerg/postgres_cmake The compilation will be enough (tests even better). I need feedbacks so that create issues on github. Very interesting NetBSD, OpenBSD, Solaris. Thanks! -- Yury Zhuravlev -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Alban Hertroys wrote: I was under the impression that the FreeBSD port already uses cmake? I tested on a 32-bit FreeBSD. All tests passed. -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Andy Colson wrote: the end of CMakeFiles/CMakeError.log shows: Many thanks! I think I understand what the problem is. I will try to fix soon. I write as a corrected. -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Yury Zhuravlev wrote: I will try to fix soon. I write as a corrected. You can try again. Thanks! -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Andy Colson wrote: Is this bad? -- Check size of int64 -- Check size of int64 - failed -- Check size of uint64 -- Check size of uint64 - failed -- Check size of int8 -- Check size of int8 - failed This is normal behavior for linux. -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
I used to use "make install-strip", is that not a thing anymore? I think it's plans for the near future. Please create issue in GitHub. /home/andy/projects/postgres_cmake/build/src/timezone//zic: Cannot create directory /usr/local/pg99: Permission denied I will fix it. sh: /usr/local/pg99/bin/initdb: No such file or directory I have not tested yet with DESTDIR. I think tomorrow will correct it. Yeah, that makes sense, its not actually there yet. Is the installcheck important to you? I can do the install if you like. installcheck is replace "make check" now. Or I did not understand your question. -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Tom Lane wrote: Really? That sure seems misleading as can be, and not something we'd want to be part of a new user's very first impression of Postgres. In configure we have similar messages: checking for int8... no checking for uint8... no checking for int64... no checking for uint64... no -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
On четверг, 11 февраля 2016 г. 22:37:12 MSK, Tom Lane wrote: Hm, well, configure does not use the word "failed" to describe expected cases. Well, it does not confuse CMake users. (MySQL, KDE) Just the other tradition, I do not see the problem here. :) -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Andy Colson wrote: Is the installcheck important to you? Hello! You can try new make check. Also "make install" started support DESTDIR. -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Tom Lane wrote: The other make switch I use all the time is -jN (with varying values of N depending on what machine I'm on). If cmake can't provide an equivalent feature, that would be a large minus, because if you have a decent number of cores -j makes a huge difference in build time. Of course supported. -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Andy Colson wrote: Its not important, but is pretty nice. It's not hard and I think I will do soon. Anyway, thanks for all your work on this. Looking good. Thanks! -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Andy Colson wrote: cmake and make -j2 fine, but then You can try again I removed some features from CMake 3.x . Realy big thanks for testing! -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] Test CMake build
Andy Colson wrote: Still didn't work. 'make check' seems to make it through make install first, but then same sort of error: Now it looks like I can do "make check" only cmake 3.x (I will try to solve it later). But "make installcheck" with DESTDIR should work. I now have no close system with CMake 2.8.x . In order not to clutter up maillist I would prefer to discuss such errors on github. Thanks! -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
[GENERAL] I need testers for incremental backups (similar Oracle)
Hello all. I'm working on incremental backup based on tracking of memory pages (ptrack). Now project in beta stage. And I did not have many feedbacks for catch bugs. I think it is important project for postgres exploitation. This approach can really speed up incremental backup with small overhead. This project append patch to postgres core and add new features to pg_arman. Small HOWTO and all links I place here: https://gist.github.com/stalkerg/dda6fed9ca4bd7cc424ad439bac04303 I will be answered here and at github. Thanks! -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
Re: [GENERAL] I need testers for incremental backups (similar Oracle)
John R Pierce wrote: doesn't a wal archive give you pretty much the same thing? No, because WAL quickly enough may become larger than the database. In the WAL can be any number of records that change only one page of memory. In addition recover from the WAL too long, because it should be "play" by Postgres. In my case pg_arman + ptrack it's almost as simple copy files. Thanks. -- Yury Zhuravlev Postgres Professional: http://www.postgrespro.com The Russian Postgres Company -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
