Fix #109 bug transaction concurrency.

[robinbiondi]

When running more than 9 transactions at the same time, we get an error: "pool.release is not a function". at line 74 in lib/transaction.js.
I found out that this.pg is returning an EvenEmitter, containing a pool which contains the function release.
Finally, when I call pool.pool.release instead of pool.release, everything works fine and the connections are released on my pg database.
This modification fixes the issue #109.

[slnode]

Can one of the admins verify this patch? To accept patch and trigger a build add comment ".ok\W+to\W+test."

[slnode]

Can one of the admins verify this patch?

[Elindorath]

When can this PR could be merged? I think it is a very small and highly valuable one! 😉

[robinbiondi]

Hello @raymondfeng. Do you have an idea of when this PR could be processed?
Cheers,

[slnode]

Can one of the admins verify this patch?

[superkhau]

Can you add some tests to verify your changes and prevent regressions in the future?

In addition, can you rebase with master to get the latest changes on this PR?

[cadesalaberry]

Hi there,

What kind of test do you expect for this PR to be merged ?

TLDR: this.pg.release() does not exist. this.pg.pool.release() does.

From what I saw, if you follow the callstack, that's where we start:

function PostgreSQL(postgresql, settings) {
   ...
  this.pg = new postgresql.Pool(this.clientConfig);
  this.settings = settings;
  ...
}

Down the trail, you see that Pool comes from a poolFactory which returns a pg-pool module:

PG = function(clientConstructor) {
  EventEmitter.call(this);
  ...
  this.Client = clientConstructor;
  this.Pool = poolFactory(this.Client);
  ...
};

That's where the naming is confusing, as the pg-pool module also contains a pool object:

Pool = module.exports = function (options, Client) {
  ...
  this.options = objectAssign({}, options)
  ...
  this.pool = new genericPool.Pool(this.options)
  ...
}

here if you want to release the pool, you will have to go through the pool object in which there in an actual pool to release.

[slnode]

Can one of the admins verify this patch?

[slnode]

Can one of the admins verify this patch?

[ssh24]

@robinbiondi Could you please sign the CLA so we can proceed with this PR?https://cla.strongloop.com/agreements/strongloop/loopback-connector-postgresql

@slnode test please

Ignore my comment. This fix has been applied in #197. Closing this PR as a result.

[robinbiondi]

@ssh24 I just did it.

Thanks!