Hey! I have the following issue, when trying to create an environment with EnvCreateFlags, I get this error:

thread '<main>' panicked at 'error opening environment: 22: Invalid argument', src/main.rs:38
An unknown error occurred

Code extract is:

use lmdb::core::{EnvCreateReadOnly};
let env_builder = EnvBuilder::new() 
  .map_size(256 * 1024 * 1024)      
  .max_dbs(5)                       
  .flags(EnvCreateReadOnly);        
let env = env_builder.open(&path, 0o664).unwrap();

Do you find anything wrong with that ? Thanks !

Btw, great lib! I tried to discover the issue by myself but I'm still learning Rust :)

Just a heads up in case somebody else is struggling with the same issue.

After a recent upgrade I've started seeing crashes in LMDB, they looked like this

#0  0x00007fad06517cfc in mdb_cursor_put () from /usr/local/lib/libbyzon.so
#1  0x00007fad0651af53 in mdb_put () from /usr/local/lib/libbyzon.so
#2  0x00007fad06506698 in set_value_with_flags (self=<optimized out>, db=<optimized out>, key=...,
    value=..., flags=<optimized out>)
    at /home/grank/.cargo/registry/src/github.com-1ecc6299db9ec823/lmdb-rs-0.7.2/src/core.rs:1069
#3  set_value (self=<optimized out>, db=<optimized out>, key=..., value=...)
    at /home/grank/.cargo/registry/src/github.com-1ecc6299db9ec823/lmdb-rs-0.7.2/src/core.rs:1061
#4  set (self=<optimized out>, db=2, key=..., value=...)
    at /home/grank/.cargo/registry/src/github.com-1ecc6299db9ec823/lmdb-rs-0.7.2/src/core.rs:1079
#5  lmdb_rs::core::{{impl}}::set (self=<optimized out>, key=..., value=...)
    at /home/grank/.cargo/registry/src/github.com-1ecc6299db9ec823/lmdb-rs-0.7.2/src/core.rs:414

and only happened in the release builds. I've tried a couple of things, like adding global locks around LMDB operations and such, I've upgraded LMDB, but to no avail.

Finally I've lovered the optimization level in order to better debug the issue and voila, the problem vanished.

My conjecture is that LMDB is normally tested with -O2 optimization (cf), but the Rust package uses -O3 (cf), the level which is not supported by the LMDB code base and triggers some kind of a bug.

I've encountered a rather bizarre segfault occurrence that is somehow linked to lmdb-rs. The segfault is odd and manifests in strangle places. In my project that I currently am experiencing issues with this it generally manifests when trying to transform data read from the db or move an Arc to the environment into a new thread. You can check out the code here.

I created a repository purely for reference in helping solve this issue, which you can find here. I'm happy to help in fixing this if you can point me in the right direction. I've really enjoyed using this library and need it to work to complete my senior project.

Hello,

is there a nice way to iterate over keys starting with a given "from" key (including that) while ending the iteration at a given "end" key (this is, excluding that key from the iteration)? While keyrange_from/_to individually provide what I need, it seems that keyrange includes the "end" key on purpose. However, I fail to find a strong reason for this.

P.

Unfortunalty, the libmbd Makefile does not allow to override CC and AR: https://github.com/LMDB/lmdb/blob/ad8488cfac644d7a289e428ab3c403c859d844cb/libraries/liblmdb/Makefile#L21-L22.

I am still a novice to Rust. Maybe it's possible to patch the Makefile with build.rs?

These methods are all thin wrappers around NativeTransaction methods that use dynamic dispatch so monomorphizing is probably a pessimization anyhow, and this lets these functions be called with dynamic dispatch (in our case we'd like to call the function with one of a variety of ToMdbValue concrete implementations, determined at runtime).

Just FYI, I've encountered a persevering early termination problem iterating keys with keyrange_to. With the database defined as follows:

let expiration_idx = try! (env.create_db ("expiration_idx", DbFlags::from_bits_truncate (lmdb_sys::MDB_DUPSORT | lmdb_sys::MDB_INTEGERKEY)));

I had to rewrite the loop as follows:

  let mut hashes_found: Vec<u64> = Vec::new();  // NB: Must not change the index while iterating it.
  let till = time::get_time().sec as u64;
  // NB: Using `iter` because `keyrange_to` doesn't work as expected (misses some keys).
  for rec in try! (expiration_idx.iter()) {
    if rec.get_key::<u64>() > till {break}
    hashes_found.push (rec.get_value::<u64>());}

Because the keyrange_to version terminated early and have been missing some keys.

The problematic keyrange_to version:

  let till = time::get_time().sec as u64 + 1;
  for rec in try! (expiration_idx.keyrange_to (&till)) {
    hashes_found.push (rec.get_value::<u64>());}

there is already a pull request, please press the merge button, currently this crate does not compile on crates.io (apparently a backwards incompatible compiler change) https://github.com/vhbit/lmdb-rs/pull/43

This method should be called periodically to ensure that the crashed readers are removed from the lock table. cf. the caveats at http://symas.com/mdb/doc/index.html

This is a bit icky since it exposes the ffi function header, but since lmdb doesn't pass a void* to the comparison function, I couldn't think of a good way to set up a wrapper function. I guess we could do a macro, but seems like that'd probably be more confusing than just needing to interact with the ffi layer. Thoughts?

Make wip branch build (closing brace, tests not updated for master rust) Add android support (update mdb to a version that correctly builds with ndk, update makefile to support ndk-build).

Makefiles are not my strong point and ndk-build is not supper cooperative, so the update may be pretty naive.

• liblmdb-sys: use cc instead if the deprecated gcc crate.
• liblmdb-sys: suppress innocuous compiler warnings from mdb
• lmdb-rs: update bitflags to version 1.3 with name-spacing [breaking change]
• lmdb-rs: update log to version 0.4
• lmdb-rs: replace try! macro usage with the ? operator
• lmdb-rs: update to Rust 2018 edition and address all clippy warnings

cc crate uses the native C compiler to compile C sources. Since gcc has been removed from android ndk, this commit helps to have a seamless compilation for android with cargo build.

This seems likely to be caused by the bug listed here https://www.openldap.org/lists/openldap-bugs/201604/msg00046.html which has been since resolved. A new version of LMDB should correct it. It looks like lmdb-rs uses LMDB 0.9.18, but this was fixed in 0.9.19 (also visible in this PowerDNS ticket).

lmdb is segment faulting in the function mdb_freelist_save

When debugging with rust-gdb:

[New Thread 0x7ffff47df700 (LWP 18986)]
mdb/libraries/liblmdb/mdb.c:3067: Assertion 'pglast <= env->me_pglast' failed in mdb_freelist_save()

Thread 62 "worker_0" received signal SIGABRT, Aborted.
[Switching to Thread 0x7ffff47df700 (LWP 18986)]
__GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:51
51      ../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb)

anyone had the same problem?