- discuss - lists.mariadb.org

[Maria-discuss] Additional promise to the MCA
by Henrik Ingo 08 Oct '09

08 Oct '09

Hi all As Monty blogged some time ago http://monty-says.blogspot.com/2009/08/thoughts-about-dual-licensing-open.h… ...we want to add a promise to our MariaDB contributors signing the MCA, that we will not use their contributions to restrict other licensees ability to contribute to Open Source. This text has now been approved by a lawyer. The question is, is it understandable? ********************************** Additional promise to contributions under the Monty Program Contributor Agreement Monty Program Ab promises to each contributor assigning rights under the Monty Program Ab Contributor Agreement, that when licensing 'products' (software), that include a contribution, to other 3rd parties, the license will be either i) an Open Source license (www.opensource.org) or ii) another license, in which case: * the other license will NOT restrict the licensee from making his own modifications to the product, including adding other suitably licensed code or other 'materials' from other sources (such as code available under suitable Open Source licenses) * the license will NOT restrict the licensee from publishing his own code or other 'materials' including, but not limited to, under an Open Source license. As a special note, it is observed that the Monty Program Ab Contributor Agreement itself fulfills these requirements. Also note that this promise does not not extend as a requirement to 3rd parties so as to restrict their rights to license products containing contributions. ********************************** We are especially interested to get feedback on the understandability of the last paragraph. henrik -- email: henrik.ingo(a)avoinelama.fi tel: +358-40-5697354 www: www.avoinelama.fi/~hingo book: www.openlife.cc

5 17

Re: [Maria-discuss] Information for release notes
by Daniel Bartholomew 28 Sep '09

28 Sep '09

Also posting this message to maria-discuss for those that might be interested in the changelog, but are not subscribed to maria-developers. On Thu, 24 Sep 2009 15:19:19 +0200, Kristian Nielsen <knielsen(a)knielsen-hq.org> wrote: Kristian> Hi Daniel, Kristian> Kristian> Here is a rough changelog written from reading through bzr Kristian> history. Kristian> Kristian> I took a dump of every commit in MariaDB that is not in MySQL Kristian> 5.1.38, and wrote this summary of the changes. Kristian> Kristian> I don't have sufficient knowledge to parse through all of the Kristian> Maria changes from before MariaDB started. Instead I wrote a Kristian> generic entry for the Maria storage engine, and also added Kristian> entries for what is already written (by Monty I think) on Kristian> http://askmonty.org/wiki/index.php/MariaDB_versus_MySQL Kristian> Kristian> So this is probably the best shot we have at a complete list Kristian> of relevant differences between MySQL and MariaDB. Do let me Kristian> know if anyone finds anything missing or any errors. Kristian> Kristian> Daniel, I hope you can use this as the basis for the release Kristian> notes. This list helps a lot. Thanks! Using the contents of the list, I have created a Changelog page on the wiki: http://askmonty.org/wiki/index.php/Manual:MariaDB_5.1_Changelog I organized the page into two sections: Features, and Bug Fixes. I also added links to the pages in the manual which exist for the changelog entries refer to. Kristian> Maybe this list will also be useful for you to fill out the Kristian> Manual on the Wiki. I am thinking you could check that all Kristian> new features are documented, and add pages for those that are Kristian> not. If not trivial, you could just make a stub page and ask Kristian> on maria-developers@ for someone to fill them out? Stub pages for the manual are easy to create: 1. add a link to the new page to http://askmonty.org/wiki/index.php/Manual:Contents 2. click on the new link 3. enter the following text into the new page: {{stub}} 4. save the new stub page To get things going, I created the following new stub pages for the manual: http://askmonty.org/wiki/index.php/Manual:Mysql_client_--abort-source-on-er… http://askmonty.org/wiki/index.php/Manual:XtraDB_storage_engine http://askmonty.org/wiki/index.php/Manual:Character_Set_Conversion_performa… http://askmonty.org/wiki/index.php/Manual:CHECKSUM_TABLE_performance Other stubs may need to be created, but I don't know which of the remaining topics need documentation or not. For example, does the entry about MariaDB NOT supporting NDB need it's own Manual page or not? Any contributions to the above stubs are *very* welcome. Any contributions to any of the other refman pages are likewise welcome: http://askmonty.org/wiki/index.php/Manual:Contents Henrik> Could the list itself be a basis for a changelog document and Henrik> then we make release notes into a more human readable document? Henrik> I thought that was your intention yesterday on IRC? The changelog page is a good compliment to the release notes page. The changelog is brief and is presented as a series of bullet points. The release notes page is longer with items presented as paragraphs in a more readable way. Thanks. -- Daniel Bartholomew

3 3

[Maria-discuss] License for MariaDB 5.1 Reference Manual
by Daniel Bartholomew 23 Sep '09

23 Sep '09

All, What license should we use for the MariaDB 5.1 Reference Manual? Some of the suggestions I've seen on IRC include: FDL : http://www.gnu.org/copyleft/fdl.html CreativeCommons : http://creativecommons.org My personal preference is to use the Creative Commons BY-SA license: http://creativecommons.org/licenses/by-sa/3.0/ This is the license used by Wikipedia and the share-alike part of the license ensures that the content remains open. It's a common license and is well understood and liked by a lot of people. Henrik mentioned the problem of attribution with CC licenses, but I don't think this is an issue with the mediawiki software we're using because of the edit histories and edit summaries, both of which fulfill this requirement. For example, the history page for the PBXT entry in the manual attributes me and Paul McCullagh for our contributions to the article: http://askmonty.org/wiki/index.php?title=Manual:PBXT_storage_engine&action=… I don't have much experience with the FDL, so if someone has reasons why we should use it, please share. Thanks. -- Daniel Bartholomew

3 2

[Maria-discuss] MariaDB 5.1 Reference Manual Introduction
by Daniel Bartholomew 22 Sep '09

22 Sep '09

All, The MariaDB 5.1 Reference Manual is progressing, but it needs an introduction. If you haven't taken a look at it yet, here is the TOC for the manual: http://askmonty.org/wiki/index.php/Manual:Contents At this stage we do not have a full or complete manual. Eventually we will have one, but creating it will take time. The purpose of the MariaDB 5.1 Reference Manual is to just document the differences between MariaDB 5.1 and MySQL 5.1 with the understanding that users can refer to the MySQL 5.1 manual for anything not covered in the MariaDB 5.1 manual. I'm currently working on an outline for the introduction page. One thing I think the intro should have is a version of the above paragraph to set expectations (we don't want users to think that we think our manual is complete). Beyond that, what other things should the introduction include? Also, if anyone knows of anything else which should be added to the manual, please let me know (or jump in and add it). Thanks. -- Daniel Bartholomew

1 0

Re: [Maria-discuss] [Maria-developers] MariaDB and the SCA
by Paul McCullagh 04 Sep '09

04 Sep '09

Hi Arjen, (Also posted to Maria-Discuss, so postings to Maria-Developers can be dropped on follow-up). On Sep 3, 2009, at 1:21 PM, Arjen Lentz wrote: > Hi Henrik, Kristian, > > On 03/09/2009, at 9:05 PM, Henrik Ingo wrote: >> On Thu, Sep 3, 2009 at 12:51 PM, Kristian >> Nielsen<knielsen(a)knielsen-hq.org> wrote: >>> Kristian Nielsen <knielsen(a)knielsen-hq.org> writes: >>> >>>> Davi Arnaut <Davi.Arnaut(a)Sun.COM> writes: >>>> >>>>> I've been monitoring the maria-developers list and there are some >>>>> patches that we would like to gradually incorporate into our >>>>> branches. For a practical example, i would like to merge into 5.1 >>>>> parts of this (http://tinyurl.com/q2ulgt) patch by Kristian. >>> >>>> Negotiations between Monty Program and Sun have started some time >>>> ago on a >>>> deal which would cover how Sun will get the rights it needs (ie. >>>> SCA) to >>>> include the MariaDB code into the MySQL code, and how Monty >>>> Program will >>>> receive compensation for giving these rights. As I understand it, >>>> these >>>> negotiations are however still in the early phase. >>> >>> Apparently, this is moving nowhere :-(. >>> >>> Davi, I'm sorry there has been no progress on this. We discussed >>> this again in >>> MariaDB, and would like to get things working better for now. >>> >>> What Monty suggested was that it would work for Sun if we submit >>> specific >>> patches under a BSD-new license. >>> >>> I realise that the http://tinyurl.com/q2ulgt patch request is now >>> so old that >>> it may be of no interest to you any longer. But if you have a >>> request for this >>> or other patches, please send them. All developers at Monty >>> Program have the >>> ability to submit their work to Sun at their own discretion, so we >>> should be >>> able to make this work on a low-overhead technician-to-technician >>> level. >> >> Thanks for picking this up. I just wanted to confirm this is true and >> we are happy to cooperate. > > > It's how it works with Drizzle. > However, I've blogged and otherwise written and spoken about this > before, it does not make me particularly happy. > > Here's the layout: > - the GPL parts of the code are owned by Sun. > - BSD can be incorporated non-OSS derivatives. > > And the consequenec of this is that Sun is able to perpetuate the > dual licensing model - a model which perhaps worked once upon a type > and did well, but which is definitely outdated, and only abused by > greedy salespeople. As far as I know, with the advent of the BSD licensed libdrizzle (https://launchpad.net/libdrizzle ), which also runs with MySQL (and I assume MariaDB), the dual licensing model is now irrelevant. My reading of the GPL license (http://www.gnu.org/copyleft/gpl.html) is that only programs that link GPL components become part of the greater work. From the GPL section "1. Source Code.": "For example, Corresponding Source includes interface definition files associated with source files for the work, and the source code for shared libraries and dynamically linked subprograms that the work is specifically designed to require, such as by intimate data communication or control flow between those subprograms and other parts of the work." So linked sub-programs are included. But: "However, it does not include the work's System Libraries, or general- purpose tools or generally available free programs which are used unmodified in performing those activities but which are not part of the work." So an unmodified free program like mysqld (the MySQL daemon), are not part of the work. This means that a programming linking libdrizzle, and then accessing mysqld over a communications _channel_ (like TCP, socket, pipe, etc), does not need to be GPL. > By agreeing to this arrangement, contributors ensure that Sun is > able to continue that. > Do you want that? > > "yes" might be a valid choice, on the basis that it might be more > important to get the changes upstream. > But I think the question needs to be asked explicitly. Hence. -- Paul McCullagh PrimeBase Technologies www.primebase.org www.blobstreaming.org pbxt.blogspot.com

4 7

Re: [Maria-discuss] [Maria-developers] MariaDB and the SCA
by Henrik Ingo 03 Sep '09

03 Sep '09

Hi Giuseppe I'm cross-posting to maria-discuss since this is now a non-technical discussion. Please remove maria-developers when replying. On Thu, Sep 3, 2009 at 3:01 PM, Giuseppe Maxia<g.maxia(a)gmail.com> wrote: > Why it's perfectly OK that MariaDB gets MySQL bugfixes for free and > instead it's a rip off if MySQL gets MariaDB bug fixes? Of course, all of the MariaDB source code is available as GPL, just like for MySQL. Sun is not prohibited from using it under the GPL, just like we are using MySQL code. > MariaDB requires the same kind of legal grants that Sun asks, and the > FAQ even states "The MCA is based on Sun's Contributor Agreement. We > think it is one of the best contributor agreements around! " > http://askmonty.org/wiki/index.php/MCA_FAQ We've made some adjustments, but yes, apparently it is one of the best contributor agreements around, the community seems to think. That's why we use it. > So, where is the catch? If Sun asks for the SCA, it's "greedy > salesmen". If MariaDB asks the same thing, it's "protecting the > community". Actually, it is very important that MariaDB copyrights are pooled to one owner, we could get into trouble if we didn't do that. (I won't give the details, so as not to give any ideas.) Same is true for XtraDB, where we have advised Percona to pool the copryights from any non-Percona contributors. It's not like we (MP) want to own everything, but somebody has to do it. Personally my first motivation in favor of pooling copyrights was however entirely benevolent towards MySQL and its users. By having the copyrights pooled in one place (rather than just going all GPL), we are able to assign the code back to Sun (if we so agree), and I'd really like to do that so that all MySQL users, also the commercial licensees would benefit from the MariaDB work. henrik -- email: henrik.ingo(a)avoinelama.fi tel: +358-40-5697354 www: www.avoinelama.fi/~hingo book: www.openlife.cc

3 2

[Maria-discuss] Compile failure on OS
by Sergey Petrunya 24 Aug '09

24 Aug '09

Hi! We're getting a compile error when trying to compile PBXT+Maria on ppc MacOS X: http://askmonty.org/buildbot/builders/macosx-g5-dbg/builds/12/steps/compile… looking at the bottom there, one sees: if ccache gcc -DHAVE_CONFIG_H -I. -I. -I../../../include -I../../../include -I../../../include -I../../../regex -I../../../storage/innobase/include -I../../../sql -I. -g -Wimplicit -Wreturn-type -Wswitch -Wtrigraphs -Wcomment -W -Wchar-subscripts -Wformat -Wparentheses -Wsign-compare -Wwrite-strings -Wunused-function -Wunused-label -Wunused-value -Wunused-variable -Wreorder -Wctor-dtor-privacy -Wnon-virtual-dtor -felide-constructors -fno-exceptions -fno-rtti -DUNIV_MUST_NOT_INLINE -DEXTRA_DEBUG -DFORCE_INIT_OF_VARS -DSAFEMALLOC -DPEDANTIC_SAFEMALLOC -O0 -g3 -gdwarf-2 -fno-implicit-templates -fno-exceptions -fno-rtti -D_P1003_1B_VISIBLE -DSIGNAL_WITH_VIO_CLOSE -DSIGNALS_DONT_BREAK_READ -DIGNORE_SIGHUP_SIGQUIT -DDONT_DECLARE_CXA_PURE_VIRTUAL -MT libpbxt_a-datalog_xt.o -MD -MP -MF ".deps/libpbxt_a-datalog_xt.Tpo" -c -o libpbxt_a-datalog_xt.o `test -f 'datalog_xt.cc' || echo './'`datalog_xt.cc; \ then mv -f ".deps/libpbxt_a-datalog_xt.Tpo" ".deps/libpbxt_a-datalog_xt.Po"; else rm -f ".deps/libpbxt_a-datalog_xt.Tpo"; exit 1; fi lock_xt.h: In function 'void xt_atomic_inc2(volatile u_int16_t*)': lock_xt.h:194: error: '__sync_fetch_and_add' was not declared in this scope lock_xt.h: In function 'void xt_atomic_dec2(volatile u_int16_t*)': lock_xt.h:209: error: '__sync_fetch_and_sub' was not declared in this scope lock_xt.h: At global scope: lock_xt.h:584: warning: 'struct XTRowLockList' only defines private constructors and has no friends lock_xt.h:640: warning: 'struct XTRowLocks' only defines private constructors and has no friends Is PBXT expected to work on MacOS X G5? BR Sergey -- Sergey Petrunia, Software Developer Monty Program AB, http://askmonty.org Blog: http://s.petrunia.net/blog

3 2

Re: [Maria-discuss] Compile failure on OS
by Hakan Kuecuekyilmaz 21 Aug '09

21 Aug '09

See also, Compile fails on Mac OS X PPC https://bugs.launchpad.net/bugs/415475 Best regards, Hakan Am 21.08.2009 um 15:58 schrieb Sergey Petrunya <psergey(a)askmonty.org>: > Hi! > > We're getting a compile error when trying to compile PBXT+Maria on > ppc MacOS X: > > http://askmonty.org/buildbot/builders/macosx-g5-dbg/builds/12/steps/compile… > > looking at the bottom there, one sees: > > if ccache gcc -DHAVE_CONFIG_H -I. -I. -I../../../include -I../../../ > include > -I../../../include -I../../../regex -I../../../storage/innobase/ > include > -I../../../sql -I. -g -Wimplicit -Wreturn-type -Wswitch - > Wtrigraphs > -Wcomment -W -Wchar-subscripts -Wformat -Wparentheses -Wsign-compare > -Wwrite-strings -Wunused-function -Wunused-label -Wunused-value > -Wunused-variable -Wreorder -Wctor-dtor-privacy -Wnon-virtual-dtor > -felide-constructors -fno-exceptions -fno-rtti -DUNIV_MUST_NOT_INLINE > -DEXTRA_DEBUG -DFORCE_INIT_OF_VARS -DSAFEMALLOC - > DPEDANTIC_SAFEMALLOC -O0 > -g3 -gdwarf-2 -fno-implicit-templates -fno-exceptions -fno-rtti > -D_P1003_1B_VISIBLE -DSIGNAL_WITH_VIO_CLOSE -DSIGNALS_DONT_BREAK_READ > -DIGNORE_SIGHUP_SIGQUIT -DDONT_DECLARE_CXA_PURE_VIRTUAL -MT > libpbxt_a-datalog_xt.o -MD -MP -MF ".deps/libpbxt_a-datalog_xt.Tpo" - > c -o > libpbxt_a-datalog_xt.o `test -f 'datalog_xt.cc' || echo > './'`datalog_xt.cc; \ > then mv -f ".deps/libpbxt_a-datalog_xt.Tpo" > ".deps/libpbxt_a-datalog_xt.Po"; else rm -f > ".deps/libpbxt_a-datalog_xt.Tpo"; exit 1; fi > lock_xt.h: In function 'void xt_atomic_inc2(volatile u_int16_t*)': > lock_xt.h:194: error: '__sync_fetch_and_add' was not declared in > this scope > lock_xt.h: In function 'void xt_atomic_dec2(volatile u_int16_t*)': > lock_xt.h:209: error: '__sync_fetch_and_sub' was not declared in > this scope > lock_xt.h: At global scope: > lock_xt.h:584: warning: 'struct XTRowLockList' only defines private > constructors and has no friends > lock_xt.h:640: warning: 'struct XTRowLocks' only defines private > constructors and has no friends > > Is PBXT expected to work on MacOS X G5? > > BR > Sergey > -- > Sergey Petrunia, Software Developer > Monty Program AB, http://askmonty.org > Blog: http://s.petrunia.net/blog > > _______________________________________________ > Mailing list: https://launchpad.net/~maria-discuss > Post to : maria-discuss(a)lists.launchpad.net > Unsubscribe : https://launchpad.net/~maria-discuss > More help : https://help.launchpad.net/ListHelp

1 0

[Maria-discuss] why does UNION ALL use a temp table?
by MARK CALLAGHAN 03 Aug '09

03 Aug '09

Why does UNION ALL use a temp table? Many users are surprised by this. It would be faster to return rows from each branch of a UNION ALL query directly to the network. -- Mark Callaghan mdcallag(a)gmail.com

3 6

[Maria-discuss] explain for insert/update/delete
by MARK CALLAGHAN 27 Jul '09

27 Jul '09

Does MariaDB have any plans to support explain for insert/update/delete/replace/CTAS? There have been feature requests open for a very long time on official MySQL without any progress. I have had a few problems in production from 'bad' delete statements that had a very different plan than the select statement with the same where clause. The only way to check for a bad plan today is to use SHOW INNODB STATUS to determine whether the statement locks all rows in a table. http://bugs.mysql.com/bug.php?id=14745 http://bugs.mysql.com/bug.php?id=35355 -- Mark Callaghan mdcallag(a)gmail.com

2 1