Permalink
Commits on Jun 1, 2012
  1. Make sure that SqlString.objectToValues() ignores functions.

    This addresses the problem when trying to update rows received from query call. For example:
    
        conn.query('SELECT * FROM users WHERE id = 1', function(err, users)
        {
            var user = users[0];
    
            user.counter++;
    
            conn.query('REPLACE INTO users SET ?', user, function(err, info) {});
        });
    
    This would cause a problem because RowDataPacket has functions and SqlString.objectToValues() would try to escape them like any other values.
    committed Jun 1, 2012
Commits on May 31, 2012
  1. Implement support for old mysql passwords

    This was available in 0.9.x, but removed in v2.0. Since it is simple
    enough to support, I decided to add it back in, however if at all
    possible, you should upgrade the passwords any user accounts that still
    have old passwords set. They are not secure.
    
    Fixes #204
    felixge committed May 31, 2012
  2. Fix parser benchmark

    I wish MySQL wasn't such a fucked up protocol, this Parser/PacketWriter
    coupling is weird.
    felixge committed May 31, 2012
  3. Add transloadit link

    felixge committed May 31, 2012
  4. Implement support for 'LOAD DATA LOCAL INFILE'

    Closes #182.
    felixge committed May 31, 2012
  5. Link to v0.9 branch

    felixge committed May 31, 2012
  6. Add install instructions

    felixge committed May 31, 2012
  7. 2.0.0-alpha2

    felixge committed May 31, 2012
  8. Improve changes document

    felixge committed May 31, 2012
Commits on May 30, 2012
  1. Add dates to releases

    felixge committed May 30, 2012
  2. Document v2.0.0-alpha2 changes

    felixge committed May 30, 2012
  3. Rename Upgrading.md to Changes.md

    Also merged in the old changelog and added info about v2.0.0-alpha.
    felixge committed May 30, 2012
  4. Final Travis fix for real.

    felixge committed May 30, 2012
  5. Final fix for travis?

    Annoying that we need to re-connect ... but oh well.
    felixge committed May 30, 2012
  6. Travis debugging

    felixge committed May 30, 2012
  7. Test: Adjust max_allowed_packet size if needed

    This should make everything pass on travis which seems to have a MySQL
    server configured for smaller max_allowed_packet values.
    felixge committed May 30, 2012
Commits on May 27, 2012
  1. Fix packet number tracking bug

    felixge committed May 27, 2012
  2. Remove todo

    felixge committed May 27, 2012
  3. Implement support for sending large packets

    Outgoing packets > MAX_PACKET_LENGTH (16 MB) are now split into several
    MySQL packets according to the protocol specification.
    felixge committed May 27, 2012
  4. No longer expose packet headers to Protocol

    Packet headers are not part of the packet, so this is good.
    Unfortunately the MySQL protocol is sufficiently fucked for this to not
    be entirely true, so we still expose the packet length to the Query
    sequence.
    felixge committed May 27, 2012
  5. Redo packet number tracking

    Has to be done in parser now, otherwise the long packet logic would need
    to be exposed to the Sequence object which seems bad.
    felixge committed May 27, 2012
  6. Implement support for receiving large packets

    Large packets in MySQL (> 16 MB) are split into several packets due to
    the 3 byte packet length header field. This patch introduces support for
    it.
    
    This patch also disables packet number verification for now, as this
    will have to be reworked.
    felixge committed May 27, 2012
Commits on May 26, 2012
  1. Handle remaining blob types

    felixge committed May 26, 2012
Commits on May 25, 2012
  1. Switch to UTF8_GENERAL_CI as default charset

    Also make the charset configurable. This fixes #200 and aligns
    node-mysql with most other drivers as far as the default charset is
    concerned.
    
    As for #200, it seems that there is a MySQL bug that will cause certain
    queries to run massively slower when using UTF8_UNICODE_CI, even if the
    conditions used should not involve any string operations. I will file a
    bug for that as soon as I find a little more time (I'm at a conference
    right now).
    felixge committed May 25, 2012
Commits on May 23, 2012
  1. Support MariaDB handshake

    Also added some comments explaining why we have to parse the
    HandshakeInitializationPacket different than suggested by the mysql
    protocol docs.
    
    Fixes #199
    felixge committed May 23, 2012
Commits on May 22, 2012
  1. Detect user error of calling connect() twice

    Enqueuing two handshakes / connects previously had very unexpected
    behavior for the end user, this patch makes the handling of these errors
    consistent with the rest of the error handling.
    
    Fixes #204
    felixge committed May 22, 2012