Vladislav Vaintroub

My preference would be - no separator option - `--partitions` boolean option that applies to all databases and tables - individual partitions are backed up using "table1#p#part_name1", . "table1#p#part_nameN". That...

@cvicentiu : The ":" separator is fine on the command line . But I still prefer '#' , because it can be used in filenames, and ':' not necessarily ....

> @vaintroub after [MDEV-37483](https://jira.mariadb.org/browse/MDEV-37483), mariadb-dump uses the same table name encoding as the server. This would mean that mariadb-dump could dump partitions in parallel, naming files using the `#p#` pattern,...

please take a look on Windows optimized build. CI won't block that for you, on whatever reason, but [appveyor does actually complain.](https://ci.appveyor.com/project/rasmushoj/server/builds/53053467) `` C:\projects\server\mysys\my_virtual_mem.c(71,5): error C2220: the following warning is...

If this does not happen elsewhere, e.g in the base branch for this one, but happens here, then it is PR issue. If you look at what it complains (destructor...

You are aware that CMAKE_COMPILER_IS_GNU(CC|CXX) and CMAKE_(C|CXX)_COMPILER_ID STREQUAL "GNU" are not the same thing? That CMAKE_COMPILER_IS_GNU(CC|CXX) is either gcc or command-line compatible clang? In some contexts one just wants to...

Indeed, sorry for confusion, heh. this seems not to be the case already since Oct 2013 (https://cmake.org/pipermail/cmake/2013-October/056071.html) and version 2.8.10 or so, but I was sure it is still the...

Hi @csenki, Thank you for taking the time to put together this commit — I appreciate the effort and research you put into it. That said, I'm considering closing this...

closing, because a comprehensive fix dd2982dc333c05c04dac0b3970a0176bd5f809b2 was pushed.

> https://ci.appveyor.com/project/RasmusJohansson/server/builds/52435045 and https://ci.appveyor.com/project/RasmusJohansson/server/builds/52435868 (for the revert [87e2123](https://github.com/MariaDB/server/commit/87e2123ee42e6be2b5ec5e778c7f1f433f4b2051)) prove that @vaintroub’s Windows specific patch indeed has a problem. Several tests were timing out (240 seconds) when the patch is present....