[MDEV-5053] Build fails with CMake when using the Ninja backend Created: 2013-09-23 Updated: 2013-09-24 Resolved: 2013-09-24 |
|
| Status: | Closed |
| Project: | MariaDB Server |
| Component/s: | None |
| Affects Version/s: | 5.5.33 |
| Fix Version/s: | 5.5.34 |
| Type: | Bug | Priority: | Trivial |
| Reporter: | Brad Smith | Assignee: | Vladislav Vaintroub |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | None | ||
| Attachments: |
|
| Description |
|
The build fails with CMake when using the Ninja backend as follows.. ninja: error: dependency cycle: sql/GenServerSource -> sql/CMakeFiles/GenServerSource -> sql/sql_builtin.cc -> sql/GenServerSource Attached is a patch which seems to fix the issue. |
| Comments |
| Comment by Brad Smith [ 2013-09-23 ] |
|
Actually the initial ticket was for 5.5.32, but trying with 5.5.33a I see there is a new build issue. ninja: error: 'extra/jemalloc/build/lib/libjemalloc_pic.a', needed by 'unittest/mysys/explain_filename-t', missing and no known rule to make it The build was also broken even with CMake's make backend as well. |
| Comment by Sergei Golubchik [ 2013-09-23 ] |
|
That's BSD again, right? |
| Comment by Vladislav Vaintroub [ 2013-09-24 ] |
|
Cyclic dependencies is a real bug (thanks for the patch). The other thing is a CMake bug is described here http://www.cmake.org/Bug/view.php?id=13574, (target for the fix is the next CMake version, 2.8.12) Until CMake bug is fixed, you can workaround by building like this ninja jemalloc && ninja build_lzma && ninja This builds external targets first, then the rest of the project |
| Comment by Vladislav Vaintroub [ 2013-09-24 ] |
|
http://lists.askmonty.org/pipermail/commits/2013-September/005424.html |
| Comment by Vladislav Vaintroub [ 2013-09-24 ] |
|
The 'build_lzma' target mentioned in the last entry, only applies if you build tokudb. If you install jemalloc yourself and use WITH_JEMALLOC=system, or if you use WITH_JEMALLOC=no, you do not need "ninja jemalloc" either. |
| Comment by Brad Smith [ 2013-09-24 ] |
|
I am using a BSD OS but these issues are not OS specific. I'll be sticking with MariaDB 5.5.32 until these bugs are fixed. Trying to move forward with these bugs is not an option. |