[MDEV-10472] Compliation failed on bld-dan-debug Created: 2016-07-31  Updated: 2019-07-07  Resolved: 2019-07-07

Status: Closed
Project: MariaDB Server
Component/s: Compiling
Affects Version/s: 10.2
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Elena Stepanova Assignee: Unassigned
Resolution: Fixed Votes: 0
Labels: None
Environment:

cmake 2.8.12.2 / GNU 4.8.4

Issue Links:
Blocks
blocks MDEV-7069 Fix buildbot failures in main server ... Stalled

 Description   

http://buildbot.askmonty.org/buildbot/builders/bld-dan-debug/builds/3511/steps/compile/logs/stdio

In file included from /opt/buildbot-slave/mariadb/dan_demeter1/build/sql/procedure.h:30:0,
 
                 from /opt/buildbot-slave/mariadb/dan_demeter1/build/sql/sql_select.h:31,
 
                 from /opt/buildbot-slave/mariadb/dan_demeter1/build/sql/field.cc:32:
 
/opt/buildbot-slave/mariadb/dan_demeter1/build/sql/sql_class.h:4650:14: warning: missing terminating " character [enabled by default]
 
   TABLE_LIST *select_tables;
 
              ^
 
/opt/buildbot-slave/mariadb/dan_demeter1/build/sql/sql_class.h:4650:3: error: missing terminating " character
 
   TABLE_LIST *select_tables;
 
   ^


/opt/buildbot-slave/mariadb/dan_demeter1/build/sql/sql_class.h:4651:3: error: expected ';' at end of member declaration
 
   Alter_info *alter_info;
 
   ^
 
cd /opt/buildbot-slave/mariadb/dan_demeter1/build/storage/heap && /usr/bin/cmake -E cmake_link_script CMakeFiles/heap.dir/link.txt --verbose=1
 
/opt/buildbot-slave/mariadb/dan_demeter1/build/sql/sql_class.h:4651:15: error: ISO C++ forbids declaration of 'alter_info' with no type [-fpermissive]
 
   Alter_info *alter_info;
 
               ^
 
/opt/buildbot-slave/mariadb/dan_demeter1/build/sql/sql_class.h:4663:17: error: 'Alter_info' is not a type
 
                 Alter_info *alter_info_arg,
 
                 ^
 
/opt/buildbot-slave/mariadb/dan_demeter1/build/sql/sql_class.h: In constructor 'select_create::select_create(THD*, TABLE_LIST*, Table_specification_st*, int*, List<Item>&, enum_duplicates, bool, TABLE_LIST*)':
 
/opt/buildbot-slave/mariadb/dan_demeter1/build/sql/sql_class.h:4670:5: error: class 'select_create' does not have any field named 'select_tables'
 
     select_tables(select_tables_arg),
 
     ^

It also sometimes crashes with segfault:
http://buildbot.askmonty.org/buildbot/builders/bld-dan-debug/builds/3441/steps/compile/logs/stdio

 Comments   
Comment by Sergey Vojtovich [ 2016-08-04 ]

elenst, did this happens only ones or is it recurring failure? This looks like a compiler bug, is it possible to upgrade compiler on this host?

Comment by Elena Stepanova [ 2016-08-04 ]

Segmentation fault happened several times:
http://buildbot.askmonty.org/buildbot/builders/bld-dan-debug/builds/3518/steps/compile/logs/stdio
http://buildbot.askmonty.org/buildbot/builders/bld-dan-release/builds/3765/steps/compile/logs/stdio
http://buildbot.askmonty.org/buildbot/builders/bld-dan-debug/builds/3441/steps/compile/logs/stdio
http://buildbot.askmonty.org/buildbot/builders/bld-dan-debug/builds/3436/steps/compile/logs/stdio
http://buildbot.askmonty.org/buildbot/builders/bld-dan-debug/builds/3435/steps/compile/logs/stdio

For the failure described in this report, I haven't seen it before.

I suppose we can ask about the machine owner to upgrade. Based on previous experience, it might take a while. Please notice however, that it's old, but not that old – it's running 4.8.4, we have older compilers in buildbot, e.g. Quantal bintar is running on 4.7.2, not mentioning the ancient CentOS 5 bintar; trusty Debs runs 4.8.2, Wheezy has 4.7.1 etc. Of course, maybe 4.8.4 was a particularly unfortunate version, I don't know if we have it on any other builders.

Comment by Sergey Vojtovich [ 2016-08-05 ]

elenst, I don't see a bug in the code here. Subsequent build succeeded. I leave it up to you to decide how to proceed with this bug further. Feel free to assign back to me.

Comment by Elena Stepanova [ 2016-08-14 ]

Segmentation fault again:
http://buildbot.askmonty.org/buildbot/builders/bld-dan-release/builds/3825/steps/compile/logs/stdio

Comment by Elena Stepanova [ 2016-08-14 ]

I've checked some history – while it's not definitive (and error-prone, since it has to be done manually), it appears that the segfaults started happening from ~June 30. GCC was the same 4.8.4 long enough time before that, so it's not the major suspect.

I've sent email to the box owner Dan, we'll need to wait for his response.

Comment by Elena Stepanova [ 2019-07-07 ]

I don't see it happen anymore, probably some upgrades did it.

Generated at Thu Feb 08 07:42:28 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.