tech-pkg archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: sqlite3 and threads

On Fri, 01 Jan 2010 01:51:55 +0900, Greg Troxel <> 

PR pkg/42515 points out that 'svnadmin create' fails because sqlite3 is
not compiled in thread-safe mode.

I have appended a potential patch to sqlite3 taken more or less from the
PR.  I've tried it and subversion seems to work.

--with-readline-incs is removed because sqlite3's configure complains
bout it.  It doesn't actually need to be part of this patch, but it
might as well get cleaned up.

Can anyone comment on how sqlite3 is supposed to be compiled, according
to sqlite3 people, and the consequences of having it compiled with
pthreads on programs that aren't expecting that?  Can anyone foresee
adverse consequences of this.

This is too scary to commit during the freeze, so I'm curious about
opinions if this should be committed post-freeze.

Could you check py-sqlite2 package still works (make test?) after the change?

Back In 2008, I removed CONFIGURE_ARGS+=--disable-threadsafe with comments
"Without it, py-sqlite2 crashes...", because even if removed it, py-sqlite2
still worked.  I didn't added --enable-threadsafe because configure --help
said that it is "yes" by default.
But as your suggestion, it is not exactly enabled,
so it is really enabled, it may break py-sqlite2.
"Of course I love NetBSD":-)
OBATA Akio /

Home | Main Index | Thread Index | Old Index