indy | Continuously poll node.js/io.js and time the eventloop delay | Analytics library

TIdMessageBuilderHtml supports adding attachments via streams, as well as via files. However, those streams have to remain alive for the duration that the TIdCustomMessageBuilder.Attachments collection is populated, which is not an option in your case since you are looping through DB records one at a time, thus you would only be able to access 1 DB stream at a time.

You could create a local array/list of TMemoryStream objects, and then populate the TMessageBuilderHtml with those streams, but you will end up wasting a lot of memory that way since TIdMessageBuilderHtml would make its own copy of the TMemoryStream data. And there is no way to have TIdMessageBuilderHtml just use your TMemoryStream data as-is in a read-only mode (hmm, I wonder if I should add that feature!).

The reason why your manual TIdAttachmentMemory objects don't work is simply because TIdCustomMessageBuilder.FillMessage() clears the TIdMessage's body before then re-populating it, thus losing your attachments (and various other properties that you are setting manually beforehand).

You would have to add your DB attachments to the TIdMessage after FillMessage() has done its work first. But, then you risk TIdMessageBuilderHtml not setting up the TIdMessage structure properly since it wouldn't know your DB attachments exist.

On a side note, you are not using TIdAttachmentMemory correctly anyway. Do not call its CloseLoadStream() method if you have not called its OpenLoadStream() method first. Calling its LoadFromStream() method is enough in this case (or, you can even pass the TStream to TIdAttachmentMemory's constructor). Do note, however, that you are leaking the TStream returned by dm.GetDataBlbStrm().

So, in this case, you are probably better off simply populating the TIdMessage manually and not use TIdMessageBuilderHtml at all. Or, you could derive a new class from TIdMessageBuilderHtml (or TIdCustomMessageBuilder directly) and override its virtual FillBody() and FillHeaders() methods to take your DB streams into account.

You have stated that you are submitting the data as Content-Type: application/x-www-form-urlencoded

That format does not allow spaces. You need to properly encode what you are submitting.

Two ways you can do this:

With Indy:

Unless you have an aggregation query (which you don't), this should work:

There is an error on line 16.

incorrect code

You mention 3 pages of data. You are leaking 2 x TJasonArray, and you are freeing one TJasonArray on the last line right after an end; that doesn't have a corresponding begin.

From that I draw the conclusion that you have a loop (that you did not show) that you run three times. On each time you create a TJasonArray but you free only one, after the loop end;, Move the JSonArray.Free; to before the end;.

The message "No config files found in path" indicates that when logstash had finished reading all the configuration files it did not have any configuration. That could because your syslog.conf is empty, or it contains invalid UTF-8 characters (so it was skipped), or possibly logstash was unable to read it.

If you add --log.level debug to the command line you should get a "Reading config file" message for that file. If you do not then logstash was unable to find it (e.g. it does not exist or the directory is not readable by the user running logstash).

The CURL example provided in the Stripe documentation is not sending the data in JSON format at all. It is sending name=value pairs in application/x-www-form-urlencoded format instead, per the CURL documentation:

-d, --data

(HTTP MQTT) Sends the specified data in a POST request to the HTTP server, in the same way that a browser does when a user has filled in an HTML form and presses the submit button. This will cause curl to pass the data to the server using the content-type application/x-www-form-urlencoded. Compare to -F, --form.

Use the TStrings overload of TIdHTTP.Post() when posting an application/x-www-form-urlencoded request, eg:

No, Indy does not provide anything for pooling TCP connections. You will have to implement your own pool of TIdTCPClient objects in your own code.

The TIdMessage.ReplyTo property has no effect on the Return-Path header. The ReplyTo is only used for the Reply-To header, which recipients use to reply back to the sender. The Reply-To header is not used by the receiving SMTP server at all.

The Return-Path header is generated by the receiving SMTP server using the address specified in the sending client's SMTP MAIL FROM command. You can specify that address in the TIdMessage.Sender and TIdMessage.From properties (in that order), or by using one of the other overloaded TIdSMTP.Send() methods that has an AFrom parameter.