delphi - how to get the progress of socket.SendStream? - delphi

I use dephi7's client/server socket components to send a file with tfilestream from client to server.
I'am able to get the progress of file received at server side,
but at client side, how do i get the progress of the file sent ?
this is how I send the file:
fstream:=tfilestream.Create(opendialog1.FileName,fmOpenRead);
clientsocket1.Socket.SendStream(fstream);
Thanks and appreciate for any help.

It doesn't look like TClientSocket provides any feedback of its progress. I see two alternatives:
Instead of SendStream, use TStream.Read and TClientSocket.Socket.SendBuf in a loop. Read a block of data from the stream and then send it. Repeat until you reach the end of the stream.
Write a TStream descendant class that wraps (or decorates) another stream. Its Read, Write, and Seek methods can simply forward to the wrapped stream, but you can also add some events to the wrapper so you can be notified each time the socket code reads a block of data out of the stream — the SendStream method essentially does the same thing that I described above in the first alternative.

Related

Save OLE Embedded documents in Outlook email to file

I am using late binding to connect to MS Outlook and to open and extract info from outlook emails using the MailItemobject.
I am trying to save attachments to file. This is fairly straightforward in most instances using the Attachment object and its SaveAsFile method.
However, it does not work where the Attachment Type is olOLE. I believe this only relates to documents embedded in emails created in RTF format (hopefully few and far between nowadays).
Via the Attachment object it is possible to access MAPI properties not exposed by the object model using its PropertyAccessor.
The relevant MAPI property for OLE objects is PR_ATTACH_DATA_OBJ, which can be accessed using the PropertyInspector as in the following example:
Function SaveOLEAttachmentToFile(Attachment:Variant; fn:String): boolean;
var
OPA, PropName : Variant;
begin
Result := false;
OPA := Attachment.PropertyAccessor;
PropName := 'http://schemas.microsoft.com/mapi/proptag/0x3701000D '; //PR_ATTACH_DATA_OBJ
?????? := OPA.GetProperty(PropName);
end;
I am stuck at this point as I can't know work out what Delphi type to save the data to and I am not even sure this is possible having read the MS documentation (Click here). PR_ATTACH_DATA_OBJ returns a PT_OBJECT. I am hoping that this object contains the raw data which (if I could work out how to access it in Delphi) can be simply saved to a file. However, the documentation suggests it may not be that simple and it's possible I may have to work with Extended MAPI. I have spent a few hours researching the latter with no concrete result other than a headache. I appreciate I could use Redemption, but I don't want to use a third party tool for something which is fairly minor in the round.
If anyone can advise as to a data type to hold the PT_OBJECT from which it can be simply saved to file that would be my route one.
Failing that, if I need to dig deeper into MAPI, I would be grateful if anyone could clarify/amplify my research so far. I have the following steps:
Initialize MAPI.
Get an IMAPIPROP interface. I think I should be getting the interface from my Attachment object and the following seems to work (ie compiles and executes without problems): MAPIPROP := IUnknown(Attachment.MAPIObject) as IMAPIPROP. Failing that, I would have to cast the parent MailItem to IMAPIPROP interface and work my way down to the attachment via GetAttachmentTable.
Load the attachment data into an IStream: if Succeeded(MAPIPROP.OpenProperty(PR_ATTACH_DATA_OBJ, IStream, STGM_READ, 0, IUnknown(SourceStream)) then
Extract the data from the IStream and save to file
I have failed to get as far as point 3 as something would seem to be wrong with my initial casting to IMAPIPROP albeit it does not cause any violations. I have tried reading a single property from the MailItem cast to IMAPIPROP using the following code:
if (Succeeded(HrGetOneProp(MAPIPROP, PR_SUBJECT, Prop))) then
And I get an access violation. Likewise if I cast the Attachment object and query an attachment property I also get a violation. I don't think the problem lies with the call to HrGetOneProp, I think it has to be the casting to IMAPIPROP.
Any pointers re the above would be greatly appreciated.
Not quite an answer to my question, but I have thought of an alternative solution. What I am ultimately trying to do is convert a msg email as a pdf. To do that I need to extract the body and then somehow insert the embedded images. With an html email this seemed pretty straightforward ((1) extract all the attachments to a folder, (2) parse the html body for references to SRC IMG and update the location of the image to reference the saved files and (3) save the edited html body to file and open it in Word and save as PDF).
RTF emails cannot be handled in this way. However, for my specific problem there is a much easier way to achieve what I need for all email types using Outtlook and Word.
Use the MailItem.SaveAs function and save the email in either html format or mthml. The former format will save all embedded images to a sub-folder (in png and jpg formats) should you need them for any other reason. once you have your html file, open it with Word and save to PDF.
If Office is not a solution then you need to figure Istorage or use one of the Extended MAPI solutions such as Redemption.
For Delphi users there are also the following commercial offerings that I have come across in my recent travels:
IMIBO
Scalabium
Rapware
I did come across one more solution which I can't find at the moment! Will post an update if I do.
PropertyAccessor (and the Outlook Object Model in general) does not handle PT_OBJECT type properties.
What you need to do is open the PR_ATTACH_DATA_OBJ property as IStorage, and then extract the data from there (it depends on the actual type of the attachment). You can see the data in the streams in OutlookSpy (I am its author) - select the message, click IMessage button on the OutlookSpy rubbon, go to the GetAttachmentTable tab, double click on the attachment to open it, select the PR_ATTACH_DATA_OBJ property, right click, select IMAPIProp::OpenProperty, then IStorage.
If using Redemption (I am also its author) is an option, its version of RDOAttachment.SaveAsFile handles OLE attachment for the most popular formats (Word, Excel, bitmap, Power Point, Adobe PDF, etc.) - create an instance of the RDOSession object (using either CrealeOleObject or RedemptionLoader) and use RDOSession.GetRDOObjectFromOutlookObject method (pass either Attachment or MailItem object) to get back RDOMail or RDOAttachment object respectively.

Download email attachment files using Indy IMAP4 in C++ Builder

I am looking for a step by step solution on how to download mail attachments using Indy Imap in C++ Builder (I use C++ Builder XE8). I have read some tutorial in Delphi, but really getting Confused.
For example, what should I do after selecting the mailbox?
ImapClient->UIDRetrieve()
or
ImapClient->RetrieveStructure()
or
ImapClient->RetrievePart()
or
ImapClient->RetrieveEnvelop().
Then, what should I do next to identify the MessagePart no, haveing the attachment file?
The Last One, how to save that file to local drive?
Should I translate the following in C++?
TIdAttachmentFile(mbMsgP.MessageParts.Items[liCount]).SaveToFile(fName);
But I cant create a statement like this
TIdAttachmentFile(IdMessage1->MessageParts->items[no])->SaveToFile("filename");
I have read some tutorial in Delphi, but really getting Confused.
Not surprising, since IMAP is a complex and confusing protocol in general. That is why TIdIMAP4 has so many more methods compared to other mailbox protocols like TIdPOP3 and TIdSMTP (and it doesn't even implement everything IMAP is capable of).
For example, what should I do after selecting the mailbox?
ImapClient->UIDRetrieve() or ImapClient->RetrieveStructure() or ImapClient->RetrievePart() od ImapClient->RetrieveEnvelop().
That really depends on what you intend to do with the emails and their attachments.
(UID)Retrieve() downloads an entire email, parsing it into a TIdMessage and marking it as "read" on the server.
(UID)RetrieveStructure() retrieves the parent/child hierarchy of the various MiME parts within an email, creating an entry for each part in a TIdMessage.MessageParts or TIdImapMessageParts collection, providing some basic descriptive information about each part such as content type and part number. The actual content of each part is not retrieved.
(UID)RetrievePart() retrieves the actual content of a specific MIME part of an email. You do not need to download the entire email. But you do have to download the email's structure first so that you know the part number that you want to retrieve.
(UID)RetrieveEnvelope() retrieves some basic top-level headers for an email: date, subject, from, sender, reply-to, to, cc, bcc, in-reply-to, and message-id.
Then, what should I do next to identify the MessagePart no, haveing the attachment file?
If you download an entire email, you would have to loop through its MessageParts collection looking for a TIdAttachment object containing the desired filename/contenttype that you are interested in.
If you download just a part of an email, you would have to retrieve the email's structure and loop through the resulting collection looking for an entry containing the desired filename/contenttype that you are interested in, then you can request that specific part's content.
The Last One, how to save that file to local drive?
If you download an entire email, then you would call SaveToFile() on the desired TIdAttachment object:
static_cast<TIdAttachment*>(IdMessage1->MessageParts->Items[no])->SaveToFile("filename");
If you download an email's structure, you can use (UID)RetrievePart() to retrieve the attachment's data into a TStream object, such as a TFileStream.

How to check if queue with auto-generated name (amq.gen-*) exists?

In case of non-generated names it's enough to call #'queue.declare' to get newly created queue or existing one with given name. However, when using auto-generated names (beginning with amq.gen- prefix) it's not as trivial. First of all, amq. is restricted prefix, so there is no way to call #'queue.declare'{queue=<<"amq.gen-xxx">>}.
I also tried to play with passive=true option and although I may pass restricted name, I get an exit error when queue does not exists. Following is error report:
** Handler sse_handler terminating in init/3
for the reason exit:{{shutdown,
{server_initiated_close,404,
<<"NOT_FOUND - no queue 'amq.gen-wzPK0nIBPzr-dwtZ5Jy58V' in vhost '/'">>}},
{gen_server,call,
[<0.62.0>,
{call,
{'queue.declare',0,
<<"amq.gen-wzPK0nIBPzr-dwtZ5Jy58V">>,
true,false,false,false,false,[]},
none,<0.269.0>},
infinity]}}
Is there any way to solve this problem?
EDIT: Here is a short story behind this question. Disclaimer: I'm erlang newbie, so maybe there is better way to make it working :)
I have a gen_server based application holding SSE (server-side events) connections with web browsers. Each connection is bound to rabbitmq queue. SSE connection when broken, automatically tries to reconnect after given timeout - this is something that web browser supports out of the box. To reuse previously created queue I'm trying to check if queue of given name (taken from request cookie) already exists. It's all done in init callback.
You can declare a queue with the prefix amq. if the queue already exists. You would get Declare-Ok if the queue exists or access-refused if not. (My question is why would you, though? ;)
Furthermore, you can use the passive option to check if it already exists. According to AMQP reference the server treats it as not-found error if the queue doesn't exist. In order to catch this in your Erlang client you could try something around the lines of this:
try
%% declare queue with passive=true
queue_exists
catch exit:{{shutdown, {server_initiated_close,404,_},_,_} ->
queue_does_not_exist
end

How to do post-processing of video saved by Red5 media server

I have a use case, where in after sending the video stream to the Red5 server, I would like to post process on the video after the video is saved. I would like to add some metadata tags on them.
I found that this can be done in appDisconnect() method in the ApplicationAdapter, but there are other ways via which the video can be saved like by using ClientBroadcastStream also.
Example
ClientBroadcastStream stream = (ClientBroadcastStream) app.getBroadcastStream(
conn.getScope(), "hostStream");
// Stop recording
stream.stopRecording();
I would like to know if there are any events which I can listen on (like which tells me that video is saved and is saved in this location with this filename) to do post-processing of the video. So that I need not put hooks in multiple places.
Thanks
The "ez" way is to implement your own ClientBroadcastStream by extending this base class. Then simply override the "stopRecording()" method. If you would like to take a moment and add an enhancement request on the issue tracker, I would be glad to look into adding scope events for this type of thing. With a scope event you could just listen for them anywhere and handle them appropriately. Red5 issue tracker: http://code.google.com/p/red5/issues/list
Using a custom stream class would be configured in the red5-common.xml like so:
<bean id="clientBroadcastStream" scope="prototype" lazy-init="true" class="com.mypackage.MyClientBroadcastStream">
</bean>

AS3 - how to set VideoPlayer's metadata?

I've set a custom client for a FLVPlayback's netStream, to attach my own functions (onXMPData, onMetaData) to parse the various info myself. However, i'd still like to pass the meta data back to the VideoPlayer. How do I do this? I tried dispatching a METADATA_RECEIVED event with the metadata object (tried dispatching from the client, the netstream, the video player, the flvplayback..), but it does not work.
I gave up, and decided just to open two netstreams for the f4v. The regular, untampered flvplayback one.. and then another just for reading xmpdata.

Resources