[Maria-developers] Help request: Contribute to MariaDB 10.3 in Debian efforts
![](https://secure.gravatar.com/avatar/a35ef532b5347ee8f54e236a3c0ba7bd.jpg?s=120&d=mm&r=g)
Hello! I've spent the last 3-4 weeks preparing MariaDB 10.3 for Debian. I am almost done, but there are a bunch of smaller issues I need help with. I would be glad for anybody who can look into the issues and give solution ideas - or even file a merge request on Debian's Gitlab instance where the Debian development happens: https://wiki.debian.org/Teams/MySQL/patches Here is the overall tracker for MariadB 10.3 and it shows various issues: https://tracker.debian.org/pkg/mariadb-10.3 Bugs filed against MariaDB in Debian: https://bugs.debian.org/cgi-bin/pkgreport.cgi?repeatmerged=no&src=mariadb-10.3 Build status in Debian: https://buildd.debian.org/status/package.php?p=mariadb-10.3 Build status on Launchpad (the system Ubuntu uses): https://launchpad.net/~mysql-ubuntu/+archive/ubuntu/mariadb-10.3/+builds?build_text=&build_state=all Build/CI status on Debian Salsa (the development system): https://salsa.debian.org/mariadb-team/mariadb-10.3/-/jobs Git repository for the packaging: https://salsa.debian.org/mariadb-team/mariadb-10.3 Please pick somethng that is close to your expertise and help out if you can! :) - Otto
![](https://secure.gravatar.com/avatar/55725185ac933ae62e4ed7b856f04406.jpg?s=120&d=mm&r=g)
Hello! I did not get any responses to this mail here. If you want to help but don't have the domain expertise, you can still help by giving some publicity and by promoting the tweet https://twitter.com/ottokekalainen/status/1080949471922610176 Thanks! ke 19. jouluk. 2018 klo 19.19 Otto Kekäläinen (otto@mariadb.org) kirjoitti:
Hello!
I've spent the last 3-4 weeks preparing MariaDB 10.3 for Debian. I am almost done, but there are a bunch of smaller issues I need help with. I would be glad for anybody who can look into the issues and give solution ideas - or even file a merge request on Debian's Gitlab instance where the Debian development happens: https://wiki.debian.org/Teams/MySQL/patches
Here is the overall tracker for MariadB 10.3 and it shows various issues: https://tracker.debian.org/pkg/mariadb-10.3
Bugs filed against MariaDB in Debian: https://bugs.debian.org/cgi-bin/pkgreport.cgi?repeatmerged=no&src=mariadb-10.3
Build status in Debian: https://buildd.debian.org/status/package.php?p=mariadb-10.3
Build status on Launchpad (the system Ubuntu uses): https://launchpad.net/~mysql-ubuntu/+archive/ubuntu/mariadb-10.3/+builds?build_text=&build_state=all
Build/CI status on Debian Salsa (the development system): https://salsa.debian.org/mariadb-team/mariadb-10.3/-/jobs
Git repository for the packaging: https://salsa.debian.org/mariadb-team/mariadb-10.3
Please pick somethng that is close to your expertise and help out if you can! :)
- Otto
![](https://secure.gravatar.com/avatar/3bce7f356a04335a99fe8b1a74a04e02.jpg?s=120&d=mm&r=g)
On December 19, 2018, Otto Kekäläinen wrote:
Bugs filed against MariaDB in Debian: https://bugs.debian.org/cgi-bin/pkgreport.cgi?repeatmerged=no&src=mariadb-10.3 [snip] Please pick somethng that is close to your expertise and help out if you can! :)
The only one that caught my eye is this one: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=848603 "Enable innodb_large_prefix" I do not understand "Tags: wontfix", because MariaDB Server fixed this already in version 10.2 by changing the default value of the parameter and by deprecating the parameter. In MariaDB 10.3, I removed this useless parameter along with innodb_file_format. Originally, I did this in MySQL 5.7: https://dev.mysql.com/worklog/task/?id=7703 That said, I do not see any reason to not set innodb_large_prefix=ON, innodb_file_format=barracuda, innodb_default_row_format=dynamic in Debian's sample configuration for MariaDB 10.1. The old default settings of these parameters could have been useful if you wanted to be able to downgrade to something older than MariaDB 5.5 or MySQL 5.5. FWIW, I backported the parameter innodb_default_row_format to MariaDB Server 10.1.32 upon the request of a Debian developer or a fellow Debian user: MDEV-14904 Backport innodb_default_row_format Feel free to copy this information to the bug report, and to address this in the 10.1 packaging. In 10.2 and 10.3, the default values of the InnoDB parameters should be fine. -- Marko Mäkelä, Lead Developer InnoDB MariaDB Corporation
![](https://secure.gravatar.com/avatar/55725185ac933ae62e4ed7b856f04406.jpg?s=120&d=mm&r=g)
Thanks Marko for taking a look!
Marko Mäkelä
On December 19, 2018, Otto Kekäläinen wrote:
Bugs filed against MariaDB in Debian:
https://bugs.debian.org/cgi-bin/pkgreport.cgi?repeatmerged=no&src=mariadb-10.3 [snip]
Please pick somethng that is close to your expertise and help out if you can! :)
The only one that caught my eye is this one: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=848603 "Enable innodb_large_prefix"
I do not understand "Tags: wontfix", because MariaDB Server fixed this already in version 10.2 by changing the default value of the parameter and by deprecating the parameter. In MariaDB 10.3, I removed this useless parameter along with innodb_file_format.
Originally, I did this in MySQL 5.7: https://dev.mysql.com/worklog/task/?id=7703
That said, I do not see any reason to not set innodb_large_prefix=ON, innodb_file_format=barracuda, innodb_default_row_format=dynamic in Debian's sample configuration for MariaDB 10.1. The old default settings of these parameters could have been useful if you wanted to be able to downgrade to something older than MariaDB 5.5 or MySQL 5.5.
FWIW, I backported the parameter innodb_default_row_format to MariaDB Server 10.1.32 upon the request of a Debian developer or a fellow Debian user: MDEV-14904 Backport innodb_default_row_format
Feel free to copy this information to the bug report, and to address this in the 10.1 packaging. In 10.2 and 10.3, the default values of the InnoDB parameters should be fine.
This was fixed 5 months ago https://salsa.debian.org/mariadb-team/mariadb-10.1/commit/89ae638d1d3b5a7157... https://salsa.debian.org/mariadb-team/mariadb-10.1/blob/master/debian/additi... The bug you are looking at was filed against mariadb-10.0, which is a wontfix. For 10.1 (and naturally 10.3) this is already fixed. I will update the metadata for that bug now so that it does not pollute the mariadb-10.3 bug listing. Note that there is a similar kind of bug for myisam at https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=917088 that stem from the same change (default charset utf8mb4 instead of latin1 -> makes keys longer).
![](https://secure.gravatar.com/avatar/2ba23b1ba996c992519115b993f5150a.jpg?s=120&d=mm&r=g)
I see the tradition of collecting a number of non-upstream patches with limited descriptions continues: https://salsa.debian.org/mariadb-team/mariadb-10.3/tree/master/debian/patche... It would be good to see these submitted upstream... some of which are obsolete by outstanding PRs so confirming the suitability of these would be appreciated: https://github.com/MariaDB/server/pull/979 https://github.com/MariaDB/server/pull/597 https://github.com/MariaDB/server/pull/510 There are other PRs that aim to fix other broken aspects of Debian packaging: https://github.com/MariaDB/server/pull/843 https://github.com/MariaDB/server/pull/605 https://github.com/MariaDB/server/pull/530 https://github.com/MariaDB/server/pull/528
![](https://secure.gravatar.com/avatar/55725185ac933ae62e4ed7b856f04406.jpg?s=120&d=mm&r=g)
ma 7. tammik. 2019 klo 3.37 Daniel Black (daniel@linux.ibm.com) kirjoitti: ..
It would be good to see these submitted upstream... some of which are obsolete by outstanding PRs so confirming the suitability of these would be appreciated:
Hello! Yes, Daniel, good point. That is why I nowadays request patch submitters to include a link to upstream PR or MDEV in their patches, and all recently added patches include upstream references. There was also the unfortunate commit last year by Ondrej who descriptions, authors and other metadata from old patches without any good reason: https://salsa.debian.org/mariadb-team/mariadb-10.3/commit/8d21d5a3ff90a48db1...
participants (4)
-
Daniel Black
-
Marko Mäkelä
-
Otto Kekäläinen
-
Otto Kekäläinen