- developers - lists.mariadb.org

Re: [Maria-developers] [Commits] Rev 3966: MDEV-4439 ALTER TABLE .. [ADD|DROP] FOREIGN KEY IF [NOT] EXISTS does not work if constraint name is not used. in file:///home/hf/wmar/10-hf/
by Jan Lindström 04 Feb '14

04 Feb '14

Hi, In my opinion the test case is not complete, you are missing e.g. ALTER TABLE t2 ADD FOREIGN KEY IF NOT EXISTS (id) REFERENCES t1(id); -- should fail on warning ALTER TABLE t2 ADD FOREIGN KEY IF NOT EXISTS (id) REFERENCES t1(id); ALTER TABLE t2 ADD FOREIGN KEY IF NOT EXISTS a(id REFERENCES t1(id); -- should fail, there is already foreign key for id, yes it name is not b or a but anyway ALTER TABLE t2 ADD FOREIGN KEY IF NOT EXISTS b(id) REFERENCES t1(id); R: > At file:///home/hf/wmar/10-hf/ > > ------------------------------------------------------------ > revno: 3966 > revision-id: holyfoot(a)askmonty.org-20140201114304-7bi0ug2yg3eyyhlr > parent: psergey(a)askmonty.org-20140121100700-9g8hl8vx8nmk1e2w > committer: Alexey Botchkov <holyfoot(a)askmonty.org> > branch nick: 10-hf > timestamp: Sat 2014-02-01 15:43:04 +0400 > message: > MDEV-4439 ALTER TABLE .. [ADD|DROP] FOREIGN KEY IF [NOT] EXISTS does not work if constraint name is not used. > Patches for server and the Innodb engine. > Server is fixed so it does nothing if no indexes left to alter. > Innodb parser is fixed so it looks for the IF [NOT] EXISTS option in a string. > Another change is that it uses the index name for the internal dictionary. > Prior to that it only used the CONSTRAINT name for it. > > > _______________________________________________ > commits mailing list > commits(a)mariadb.org > https://lists.askmonty.org/cgi-bin/mailman/listinfo/commits -- -- Jan Lindström Principal Engineer MariaDB | MaxScale | skype: jan_p_lindstrom www.skysql.com <http://www.skysql.com/> Twitter <http://twitter.com/skysql> Blog <http://www.skysql.com/blog/> Facebook <http://www.facebook.com/skysql> LinkedIn <http://www.linkedin.com/company/1214250> Google+ <https://plus.google.com/117544963211695643458/posts>

2 3

[Maria-developers] Extra copying in debian/rules causes troubles
by Elena Stepanova 04 Feb '14

04 Feb '14

Hi all. There is a fragment in MariaDB debian/dist/*/rules files: ... ifeq ($(findstring nocheck,$(DEB_BUILD_OPTIONS)),) # Don't know why the following is necessary... cp unittest/unit.pl $(builddir)/unittest/ cp -r mysql-test/* $(builddir)/mysql-test/ cp -r sql/share/* $(builddir)/sql/share/ cp -r scripts/*sql $(builddir)/scripts/ if [ ! -f testsuite-stamp ] ; then \ ... Recently it started causing a problem. MTR assumes that the tests are run either in a proper source build (be it in-source or out-of-source), or in a binary build. With the extra-copying above, particularly due to copying of mysql-test folder, the builddir by the moment of running tests is in the "neither fish nor fowl" state. So if one runs dpkg-buildpackage with the tests enabled, it fails when it comes to the MTR stage. We had this problem reported at least 3 times over the last 24 hours, so it's not hypothetical. It's currently filed as https://mariadb.atlassian.net/browse/MDEV-5591. I tried to remove these 4 'cp' lines and the build seemed to work all right; but there might also be consequences that I fail to predict, so I would like to have the 2nd (3rd, 4th, ... Nth) opinion. Does anyone have any idea what might go wrong without these 4 lines? Also, *Kristian*, These lines are coming from the initial draft of the rules. I understand from the comment that there is no reason to ask *why* the lines were necessary, but I'm wondering if you remember why you felt they *were* necessary (even if didn't know why). Were you getting some errors without them? Or were they copied from some previous version of the file? I realize it was long time ago so you probably won't remember, but I thought it was worth checking anyway. Regards, Elena

3 4

[Maria-developers] MariaDB 5.5.35
by ImperialNetwork 04 Feb '14

04 Feb '14

Hi, I started to notice this on my error log since i upgraded from 5.5.34 to 5.5.35: 140203 5:26:25 InnoDB: Failed to set O_DIRECT on file /dev/shm/#sql7e8f6_1a94bc_10.ibd: CREATE: Invalid argument, continuing anyway 140203 5:26:25 InnoDB: O_DIRECT is known to result in 'Invalid argument' on Linux on tmpfs, see MySQL Bug#26662 I never seen this error/warning until 5.5.35. Should this be normal ? Thank you in advance. Alex

3 3

[Maria-developers] MDEV-5450 Assertion ... mysql_type_to_time_type(cached_field_type) == ltime.time_type' fails ...
by Alexander Barkov 03 Feb '14

03 Feb '14

Hi Sergei, Please review a patch for mdev-5450. Thanks.

2 3

[Maria-developers] MariaDB build dependencies
by Otto Kekäläinen 03 Feb '14

03 Feb '14

Hello, The original MariaDB debian/control in bzr/mariadb-5.5 lists many more dependencies than what are likely to be needed: Build-Depends: libtool (>= 1.4.2-7), procps | hurd, debhelper, file (>= 3.28), libncurses5-dev (>= 5.0-6), perl (>= 5.6.0), libwrap0-dev (>= 7.6-8.3), zlib1g-dev (>= 1:1.1.3-5), ${LIBREADLINE_DEV}, libevent-dev, libssl-dev, libpam0g-dev, psmisc, po-debconf, chrpath, automake1.9, doxygen, texlive-latex-base, ghostscript | gs-gpl, dpatch, gawk, bison, lsb-release, hardening-wrapper, ${CMAKE_DEP}libaio-dev The source builds error free with just: libaio-dev libncurses5-dev perl libwrap0-dev zlib1g-dev libreadline-gplv2-dev po-debconf psmisc chrpath gawk bison lsb-release hardening-wrapper cmake libssl-dev libboost1.48-dev libpam0g-dev Can anybody tell me how in particular these are supposed to affect the build process? - libevent-dev - doxygen, texlive-latex-base, ghostscript | gs-gpl - gawk Debian MySQL team has a page https://wiki.debian.org/Teams/MySQL/BuildDependencies but obiviously MariaDB is different in these aspects, for example MariaDB needs libpam0g-dev for auth_pam.so which does not exist in MySQL. If anybody can explain the above that would be great, thanks!

3 3

Re: [Maria-developers] Rev 3966: MDEV-4439 ALTER TABLE .. [ADD|DROP] FOREIGN KEY IF [NOT] EXISTS does not work if constraint name is not used
by Sergei Golubchik 01 Feb '14

01 Feb '14

Hi, Holyfoot! On Feb 01, holyfoot(a)askmonty.org wrote: > revno: 3966 > revision-id: holyfoot(a)askmonty.org-20140201114304-7bi0ug2yg3eyyhlr > parent: psergey(a)askmonty.org-20140121100700-9g8hl8vx8nmk1e2w > committer: Alexey Botchkov <holyfoot(a)askmonty.org> > branch nick: 10-hf > timestamp: Sat 2014-02-01 15:43:04 +0400 > message: > MDEV-4439 ALTER TABLE .. [ADD|DROP] FOREIGN KEY IF [NOT] EXISTS does not work if constraint name is not used. > Patches for server and the Innodb engine. > Server is fixed so it does nothing if no indexes left to alter. > Innodb parser is fixed so it looks for the IF [NOT] EXISTS option in a string. > Another change is that it uses the index name for the internal dictionary. > Prior to that it only used the CONSTRAINT name for it. ok to push Regards, Sergei

1 0

Re: [Maria-developers] Rev 3886: MDEV-4439 ALTER TABLE .. [ADD|DROP] FOREIGN KEY IF [NOT] EXISTS does not work if constraint name is not used
by Sergei Golubchik 31 Jan '14

31 Jan '14

Hi, Holyfoot! On Jan 31, holyfoot(a)askmonty.org wrote: > At file:///home/hf/wmar/mdev-4439/ > > ------------------------------------------------------------ > revno: 3886 > revision-id: holyfoot(a)askmonty.org-20140131134044-ej3ftpf240xbn0lc > parent: knielsen(a)knielsen-hq.org-20131107065240-gxxg4ilcs8o2v0q2 > committer: Alexey Botchkov <holyfoot(a)askmonty.org> > branch nick: mdev-4439 > timestamp: Fri 2014-01-31 17:40:44 +0400 > message: > MDEV-4439 ALTER TABLE .. [ADD|DROP] FOREIGN KEY IF [NOT] EXISTS does not work if constraint name is not used. > Patches for server and the Innodb engine. > Server is fixed so it does nothing if no indexes left to alter. > Innodb parser is fixed so it looks for the IF [NOT] EXISTS option, takes into account > the drop->if_not_exists parameter. > Another change is that it uses the index name for the internal dictionary. > Prior to that it only used the CONSTRAINT name for it. That looks pretty much ok, but please add tests that use generated FK names. For example, it doesn't look like you handle the case of ADD FOREIGN KEY IF NOT EXISTS existing_generated_name (...) Regards, Sergei

1 0

[Maria-developers] buildbot for non ~maria-captains branches on launchpad
by Alexander Barkov 31 Jan '14

31 Jan '14

Hello Daniel, Would it be possible to have buildbot watch branches which are on Launchpad, but which are outside of the /~maria-captains/ directory? For example: bzr+ssh://bazaar.launchpad.net/~mroonga/maria/mroonga/ ? Thanks.

3 5

Re: [Maria-developers] Rev 3965: MDEV-5403 - Reduce usage of LOCK_open: tc_count
by Sergei Golubchik 31 Jan '14

31 Jan '14

Hi, Sergey! Please, see below. On Jan 30, Sergey Vojtovich wrote: > revno: 3965 > revision-id: svoj(a)mariadb.org-20140130123642-wsysj31e1wwlqz0o > parent: psergey(a)askmonty.org-20140120123657-g1z4g11xmsdiw2dc > committer: Sergey Vojtovich <svoj(a)mariadb.org> > branch nick: 10.0 > timestamp: Thu 2014-01-30 16:36:42 +0400 > message: > MDEV-5403 - Reduce usage of LOCK_open: tc_count > > Lock-free tc_count. > === modified file 'sql/table_cache.cc' > --- a/sql/table_cache.cc 2013-12-12 17:49:14 +0000 > +++ b/sql/table_cache.cc 2014-01-30 12:36:42 +0000 > @@ -173,7 +175,9 @@ void tc_purge(void) > { > share->tdc.all_tables.remove(table); > purge_tables.push_front(table); > - tc_count--; > + my_atomic_rwlock_wrlock(&LOCK_tdc_atomics); > + my_atomic_add32(&tc_count, -1); > + my_atomic_rwlock_wrunlock(&LOCK_tdc_atomics); Okay, so in this new code tc_count is always equal or greater than the actual number of tables in the cache, right? (because you first remove the table from the cache, then decrement the counter) > } > } > tdc_it.deinit(); > @@ -245,47 +249,52 @@ static void check_unused(THD *thd) > > void tc_add_table(THD *thd, TABLE *table) > { > + bool need_purge; > DBUG_ASSERT(table->in_use == thd); > mysql_mutex_lock(&LOCK_open); > table->s->tdc.all_tables.push_front(table); > + mysql_mutex_unlock(&LOCK_open); > + > /* If we have too many TABLE instances around, try to get rid of them */ > - if (tc_count == tc_size) > + my_atomic_rwlock_wrlock(&LOCK_tdc_atomics); > + need_purge= my_atomic_add32(&tc_count, 1) >= (int32) tc_size; Hmm, but here you increment after adding a table. I'd suggest to keep tc_count consistently either always no smaller than the number of tables in the cache (that is, decrement after removing, increment before adding), or always no larger than that (derement before removing, increment after adding). > + my_atomic_rwlock_wrunlock(&LOCK_tdc_atomics); > + > + if (need_purge) > { > + TABLE *purge_table= 0; > + TABLE_SHARE *share; > TDC_iterator tdc_it; > - mysql_mutex_unlock(&LOCK_open); > > tdc_it.init(); > mysql_mutex_lock(&LOCK_open); > - if (tc_count == tc_size) > - { > - TABLE *purge_table= 0; > - TABLE_SHARE *share; > while ((share= tdc_it.next())) > { > TABLE_SHARE::TABLE_list::Iterator it(share->tdc.free_tables); > TABLE *entry; > while ((entry= it++)) > if (!purge_table || entry->tc_time < purge_table->tc_time) > purge_table= entry; > } > + tdc_it.deinit(); > + > if (purge_table) > { > - tdc_it.deinit(); > purge_table->s->tdc.free_tables.remove(purge_table); > purge_table->s->tdc.all_tables.remove(purge_table); > mysql_rwlock_rdlock(&LOCK_flush); > mysql_mutex_unlock(&LOCK_open); > intern_close_table(purge_table); > mysql_rwlock_unlock(&LOCK_flush); > + > + my_atomic_rwlock_wrlock(&LOCK_tdc_atomics); > + my_atomic_add32(&tc_count, -1); > + my_atomic_rwlock_wrunlock(&LOCK_tdc_atomics); > check_unused(thd); > - return; > } > - } > - tdc_it.deinit(); > - } > - /* Nothing to evict, increment tc_count. */ > - tc_count++; > + else > mysql_mutex_unlock(&LOCK_open); > + } > } > > > @@ -357,25 +366,35 @@ bool tc_release_table(TABLE *table) > DBUG_ASSERT(table->in_use); > DBUG_ASSERT(table->file); > > + if (table->needs_reopen() || tc_records() > tc_size) > + { > + mysql_mutex_lock(&LOCK_open); > + table->in_use= 0; > + goto purge; > + } Why did you move this up? > table->tc_time= my_interval_timer(); > > mysql_mutex_lock(&LOCK_open); > table->in_use= 0; > - if (table->s->has_old_version() || table->needs_reopen() || tc_count > tc_size) > - { > - tc_count--; > + if (table->s->has_old_version()) > + goto purge; > + /* Add table to the list of unused TABLE objects for this share. */ > + table->s->tdc.free_tables.push_front(table); > + mysql_mutex_unlock(&LOCK_open); > + check_unused(thd); > + return false; > + > +purge: > + my_atomic_rwlock_wrlock(&LOCK_tdc_atomics); > + my_atomic_add32(&tc_count, -1); > + my_atomic_rwlock_wrunlock(&LOCK_tdc_atomics); > table->s->tdc.all_tables.remove(table); and now you decrement before removal > mysql_rwlock_rdlock(&LOCK_flush); > mysql_mutex_unlock(&LOCK_open); > intern_close_table(table); > mysql_rwlock_unlock(&LOCK_flush); > return true; > - } > - /* Add table to the list of unused TABLE objects for this share. */ > - table->s->tdc.free_tables.push_front(table); > - mysql_mutex_unlock(&LOCK_open); > - check_unused(thd); > - return false; > } > > Regards, Sergei

2 1

[Maria-developers] FreeBSD compilation problems and a fix
by Alexander Barkov 31 Jan '14

31 Jan '14

Hi guys, I just tried to compile 10.0 on FreeBSD-9.1 RELEASE and noticed two problems: 1. It printed this warning for every source file: /usr/include/sys/timeb.h:42:2: warning: #warning "this file includes <sys/timeb.h> which is deprecated" This diff fixes the problem: -#ifdef HAVE_SYS_TIMEB_H +#if defined(HAVE_SYS_TIMEB_H) && !defined(__FreeBSD__) 2. It failed to compile because of pcre (both libraries and includes) installed in the system. Compiler used the *system* header pcre.h together with the *bundled* library libpcre.a. I had a similar complaint from the MacPorts maintainers. This diff fixes the problem: # Common defines and includes +SET(CMAKE_INCLUDE_DIRECTORIES_PROJECT_BEFORE ON) ADD_DEFINITIONS(-DHAVE_CONFIG_H) INCLUDE_DIRECTORIES(${CMAKE_CURRENT_BINARY_DIR}/include Does this look OK? If so, which version should I push it in? Thanks.

2 3