- developers - lists.mariadb.org

Re: [Maria-developers] [Commits] 5f461a9: MDEV-7640: CHANGE MASTER TO doesn't work with prepared statements
by Oleksandr Byelkin 11 Oct '15

11 Oct '15

Hi! On 17.09.15 00:10, Nirbhay Choubey wrote: > revision-id: 5f461a94a5aa2e1a20fa559ca612d96b53201e4e > parent(s): 29ac245dd04c5416d57a90b0ab3c4d08cbc4d723 > committer: Nirbhay Choubey > branch nick: 5.5 > timestamp: 2015-09-16 18:10:06 -0400 > message: > > MDEV-7640: CHANGE MASTER TO doesn't work with prepared statements > > When CHANGE MASTER was executed as a PS, its attributes were wrongly > getting reset toward the end of PREPARE. As a result, the subsequent > executions had no effect. Fixed by making sure that the CHANGE MASTER > attributes are preserved during the lifetime of the PS. It is OK to push but after expanding the comment about what exactly reset in the persistent statement memory (as we agreed) [skip]

1 0

Re: [Maria-developers] [Commits] 06e3c05: MDEV-8842 add group support to pam_user_map module.
by Sergei Golubchik 07 Oct '15

07 Oct '15

Hi, Holyfoot! On Oct 07, holyfoot(a)askmonty.org wrote: > revision-id: 06e3c05f1ad3156b5a3e0a775b07fb13b9ebb1a8 (mariadb-10.1.7-68-g06e3c05) > parent(s): 8afe96f011eb8037a92b4b3aab16118b0771ad50 > committer: Alexey Botchkov > timestamp: 2015-10-07 15:52:26 +0500 > message: > > MDEV-8842 add group support to pam_user_map module. > Added to the pam_user_map module. Great, thanks! Ok to push. I assume you've tested that it still works after all the changes. Regards, Sergei

1 0

Re: [Maria-developers] [Commits] 50dc8c0: MDEV-8842 add group support to pam_user_map module.
by Sergei Golubchik 07 Oct '15

07 Oct '15

Hi, Holyfoot! On Oct 07, holyfoot(a)askmonty.org wrote: > revision-id: 50dc8c0e8a27d18a3d75ff43a87975fcd5e0e7f6 (mariadb-10.1.7-67-g50dc8c0) > parent(s): bed4e847950eef50930b44632eea43416e7b37d1 > committer: Alexey Botchkov > timestamp: 2015-10-07 00:51:33 +0500 > message: > > MDEV-8842 add group support to pam_user_map module. > Added to the pam_user_map module. Looks pretty much ok. Just one suggestion: I'd keep groups in the pam_sm_authenticate and passed the pointer down to user_in_group. Indeed, there may many @group entries in the mapping file, seems like a waste do repopulate group array every time. > --- > plugin/auth_pam/mapper/pam_user_map.c | 59 +++++++++++++++++++++++++++++++++-- > 1 file changed, 57 insertions(+), 2 deletions(-) > > diff --git a/plugin/auth_pam/mapper/pam_user_map.c b/plugin/auth_pam/mapper/pam_user_map.c > index e73ab6d..a3008bd 100644 > --- a/plugin/auth_pam/mapper/pam_user_map.c > +++ b/plugin/auth_pam/mapper/pam_user_map.c > @@ -13,22 +13,71 @@ auth required pam_user_map.so > > And create /etc/security/user_map.conf with the desired mapping > in the format: orig_user_name: mapped_user_name > + @user's_group_name: mapped_user_name > ========================================================= > -#comments and emty lines are ignored > +#comments and emtpy lines are ignored > john: jack > bob: admin > top: accounting > +@group_ro: readonly > ========================================================= > > */ > > +#include <stdlib.h> > #include <stdio.h> > #include <syslog.h> > +#include <grp.h> > +#include <pwd.h> > +#include <limits.h> > #include <security/pam_modules.h> > > #define FILENAME "/etc/security/user_map.conf" > #define skip(what) while (*s && (what)) s++ > > +#define GROUP_BUFFER_SIZE 100 > +int user_in_group(const char *user, const char *group) > +{ > + gid_t group_buffer[GROUP_BUFFER_SIZE]; > + gid_t *groups= group_buffer; > + gid_t group_id; > + gid_t user_group_id; > + int ng, i; > + > + { > + struct passwd *pw= getpwnam(user); > + struct group *g= getgrnam(group); > + if (pw == NULL || g == NULL) > + return 0; > + user_group_id= pw->pw_gid; > + group_id= g->gr_gid; > + } > + > + ng= GROUP_BUFFER_SIZE; > + if (getgrouplist(user, user_group_id, groups, &ng) < 0) > + { > + /* The rare case when the user is present in more than */ > + /* GROUP_BUFFER_SIZE groups. */ > + groups= (gid_t *) malloc(ng * sizeof (gid_t)); > + if (!groups) > + return 0; > + > + (void) getgrouplist(user, user_group_id, groups, &ng); > + } > + > + for (i= 0; i < ng; i++) > + { > + if (groups[i] == group_id) > + break; > + } > + > + if (groups != group_buffer) > + free(groups); > + > + return i < ng; > +} > + > + > int pam_sm_authenticate(pam_handle_t *pamh, int flags, > int argc, const char *argv[]) > { > @@ -51,10 +100,14 @@ int pam_sm_authenticate(pam_handle_t *pamh, int flags, > while (fgets(buf, sizeof(buf), f) != NULL) > { > char *s= buf, *from, *to, *end_from, *end_to; > + int check_group; > + > line++; > > skip(isspace(*s)); > if (*s == '#' || *s == 0) continue; > + if ((check_group= *s == '@')) > + s++; > from= s; > skip(isalnum(*s) || (*s == '_')); > end_from= s; > @@ -67,7 +120,9 @@ int pam_sm_authenticate(pam_handle_t *pamh, int flags, > if (end_to == to) goto syntax_error; > > *end_from= *end_to= 0; > - if (strcmp(username, from) == 0) > + if (check_group ? > + user_in_group(username, from) : > + (strcmp(username, from) == 0)) > { > pam_err= pam_set_item(pamh, PAM_USER, to); > goto ret; Regards, Sergei

2 1

Re: [Maria-developers] [Commits] c934b5e: MDEV-8842 add group support to pam_user_map module.
by Sergei Golubchik 06 Oct '15

06 Oct '15

Hi, Holyfoot! On Oct 06, holyfoot(a)askmonty.org wrote: > revision-id: c934b5e661b19b7934d1a9af40afc70b727c2273 (mariadb-10.1.7-64-gc934b5e) > parent(s): 90f2c822469a4f88f882ba8974e790f0fb0b2702 > committer: Alexey Botchkov > timestamp: 2015-10-06 14:02:29 +0500 > message: > > MDEV-8842 add group support to pam_user_map module. > Added to the pam_user_map module. Right, thanks. A couple of comments, see below. > diff --git a/plugin/auth_pam/mapper/pam_user_map.c b/plugin/auth_pam/mapper/pam_user_map.c > index e73ab6d..0787b1c 100644 > --- a/plugin/auth_pam/mapper/pam_user_map.c > +++ b/plugin/auth_pam/mapper/pam_user_map.c ... > + /* > + So below we get the buffer of the NGROUPS_MAX size > + which can take 256K on some systems. > + It supposed to be working differently - we'd get > + the 'ng' from the getgrouplist() and alloc the respective > + size, but the problem is a bug in getgrouplist(). > + The glibc 2.3.2 implementation of this function is broken: > + it overwrites memory when the actual number of groups is larger > + than *ngroups. > + */ Forget 2.3.2, use it as it was supposed to be used. Our oldest builder uses glibc-2.5 > + (void) getgrouplist(user, user_group_id, groups, &ng); > + for (i= 0; i < ng; i++) > + { > + struct group *g; > + if ((g= getgrgid(groups[i])) == NULL) > + return 0; > + if (g->gr_gid == group_id) > + return 1; Eh? Why don't you compare group_id with groups[i] directly ? Regards, Sergei

2 1

Re: [Maria-developers] [Commits] 7ea735b: MDEV-8431 Feedback plugin needs an option for http proxy.
by Sergei Golubchik 06 Oct '15

06 Oct '15

Hi, Holyfoot! On Oct 05, holyfoot(a)askmonty.org wrote: > revision-id: 7ea735baa2df4f4577cf951bc69ccf87fd91f436 (mariadb-10.1.7-40-g7ea735b) > parent(s): ba0b6685510f4a271da8d386ff75bfd83d3156c2 > committer: Alexey Botchkov > timestamp: 2015-10-05 14:46:12 +0500 > message: > > MDEV-8431 Feedback plugin needs an option for http proxy. > 'feedback_http_proxy' system variable added to specify the > proxy server as host:port. Not a dynamic one. That's a bit more complex that I thought it'll be, but ok. Still - did you test it? How? Regards, Sergei

2 2

Re: [Maria-developers] [Commits] ee656a4: MDEV-7846: postreview fix
by Sergei Golubchik 06 Oct '15

06 Oct '15

Hi, Sanja! I have no idea how this postreview fix relates to any of my review comments. Let me repeat them again: 1. Add comments 2. Add a test case with row items 3. In your test case - if you remove GROUP BY you get a crash again. This might be a different bug, though. On Aug 05, sanja(a)mariadb.com wrote: > revision-id: ee656a4aabb5ebbe973b7e4c99be83eb5dc8c0ce > parent(s): 9f8a458fb2d07298810bb5d9824ce728017bfb32 > committer: Oleksandr Byelkin > branch nick: server > timestamp: 2015-08-05 11:57:35 +0200 > message: > > MDEV-7846: postreview fix > > diff --git a/sql/item_cmpfunc.cc b/sql/item_cmpfunc.cc > index 29fff8e..e96d1bc 100644 > --- a/sql/item_cmpfunc.cc > +++ b/sql/item_cmpfunc.cc > @@ -1455,12 +1455,13 @@ bool Item_in_optimizer::fix_left(THD *thd, Item **ref) > next execution we need to copy args[1]->left_expr again. > */ > ref0= &(((Item_in_subselect *)args[1])->left_expr); > + args[0]= ((Item_in_subselect *)args[1])->left_expr; > } > if ((!(*ref0)->fixed && (*ref0)->fix_fields(thd, ref0)) || > (!cache && !(cache= Item_cache::get_cache(*ref0)))) > DBUG_RETURN(1); > if (args[0] != (*ref0)) > - current_thd->change_item_tree(args, (*ref0)); > + args[0]= (*ref0); > DBUG_PRINT("info", ("actual fix fields")); Regards, Sergei

2 1

[Maria-developers] Problem in compiling Maria-db
by indiavaibhav vaibhavindia 02 Oct '15

02 Oct '15

Hello everyone, My name is Vaibhav.I am new to programming for open source. I am facing problems in trying to compile Maria-Db.As instructed on the website, I cloned the maria db source and installed all the files required for compilation. However on using cmake,I am getting an error,screenshot of which is attached is with this mail. Can someone please help me out? Thanks, Vaibhav Choudhary

2 3

Re: [Maria-developers] crosscompiling in buildroot: cannot find -laio and other link error
by Sylvain Raybaud 02 Oct '15

02 Oct '15

Ouch. I'm really sorry, I don't know what happend with my mua and why all these mails were sent to the list. Only the last one counts. Sincere apologies :'( On 02/10/15 17:10, Sylvain Raybaud wrote: > Hi Sergei! > > Thanks for your help. > > On 01/10/15 16:16, Sergei Golubchik wrote: >> Why would you build host variant for ha_connect.so with the library >> paths for the target architecture? > I don't actually, I don't know why this happened. > >> How do you invoke cmake? > The Buildroot system takes care of most of the maging. But... > >> Did you check >> https://mariadb.com/kb/en/mariadb/cross-compiling-mariadb/ ? Regards, >> Sergei > Ooops. For some reason I completely missed this page. Building only the > host variant of import_executables target eases cross-compilation a lot > indeed... Appart from that I had figured most of this out, or buildroot > was taking care of it. I still have a remark: I don't think I can > hardcode the value of > > HAVE_IB_GCC_ATOMIC_BUILTINS > > as suggested in https://lists.launchpad.net/maria-discuss/msg02911.html > because if I want this package accepted in buildroot it must > -- Sylvain Raybaud www.green-communications.fr

1 0

[Maria-developers] MariaDB Developer Meeting in Amsterdam Oct 13-15
by Rasmus Johansson 02 Oct '15

02 Oct '15

Hi maria-developers, There will be a MariaDB Developer Meeting taking place in Amsterdam on October 13-15. It would be great to see many of you there. The main focus of the event is as the title implies the development aspect of MariaDB, but also aspects like documentation, infrastructure and other community related topics will have their place in the event. Also MaxScale developers will be present. If you are interested please sign up by joining and rsvping the event's meetup page here: http://www.meetup.com/MariaDB-Developers/events/224068714/ The session backlog and the schedule for the three days are found here: https://docs.google.com/a/skysql.com/spreadsheets/d/17J_Jloc8I33nsEIE1KCZct… You'll notice that the schedule is not fixed and the idea is the following: - Everyone attending the meeting can affect the agenda by adding sessions to the session backlog - Everyone attending can/should vote on the session of interest in the backlog by raising the count in the Votes column Please feel free to suggest new sessions and remember to vote on the ones you think are most interesting (no limitation on how many you vote for). Best regards, Rasmus -- Rasmus Johansson, VP Engineering MariaDB | t: +358 50 499 9589 | Skype: ratzpo

2 2

Re: [Maria-developers] [Commits] 09eb231: MDEV-8762: file_key_management plugin should be disabled by default when linked statically
by Sergei Golubchik 01 Oct '15

01 Oct '15

Hi, Jan! That's not a good idea. It'd mean that installing a plugin is not enough, it must be explicitly enabled. I don't see any reason why one should be forced to do it. I'll comment in the bugreport. On Oct 01, Jan Lindström wrote: > revision-id: 09eb23148830c53e594cfb88dcb31f1b9e42ae5e (mariadb-10.1.7-33-g09eb231) > parent(s): 3a7cee05d6bac27c3c2f51afb8986c1581649bbc > committer: Jan Lindström > timestamp: 2015-10-01 14:27:10 +0300 > message: > > MDEV-8762: file_key_management plugin should be disabled by default when linked statically > > --- > sql/sql_plugin.cc | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/sql/sql_plugin.cc b/sql/sql_plugin.cc > index 6d8268e..1d2e2d8 100644 > --- a/sql/sql_plugin.cc > +++ b/sql/sql_plugin.cc > @@ -211,6 +211,7 @@ static struct > > /* we disable few other plugins by default */ > ,{ "feedback", PLUGIN_OFF } > + ,{ "file_key_management", PLUGIN_OFF } > }; Regards, Sergei

2 2