ThreeSharp Work Item Rss Feed

Created Issue: How to change ACL in C# [23642]

suryacd — Thu, 22 Nov 2012 07:18:30 GMT

Hi,

I am using AmazonS3 to store files. While uploading, the ACL of the file
is set as Private so that the file is not viewable by giving a url like
this :https://bucketname.s3.amazonaws.com/fileName.
My requirement is I need to change the ACL to PublicRead while a logged in user
clicks a button so as to view this file.
Can this be done in C#?

Thanks,

Surya.

Created Issue: How to restrict access to a file in S3 [23641]

suryacd — Thu, 22 Nov 2012 06:08:42 GMT

Hi,

In AmazonS3, a file can be accessed by giving the url like https://bucketname.s3.amazonaws.com/fileName.

Is there any way to restrict this so that only a logged in user is able to
view the file? I need to do this in my c# project.

Thanks.

Created Issue: Where is the file "DistributionSummary.cs"? [16294]

tvmogul — Wed, 10 Feb 2010 11:50:20 GMT

I noticed that your source code does NOT have the file "DistributionSummary.cs" in the model directory or anywhere else?
I downloded the other project you have, namely, "Affirma.CloudFront.AdminTool" and noticed that if I use your posted source code for the library that that project compiles perfectly !
Your Affirma.CloudFront.AdminTool declares the following:
private SortedList<string, DistributionSummary> distributionList = new SortedList<string,DistributionSummary>();

And DistributionSummary is in "DistributionSummary.cs" withe the namespace "Affirma.ThreeSharp.Model.DistributionSummary" BUT THIS FILE DOES NOT EXIST in your library and yet the code compiles! So I tracked it and found the file located in my "User" directory !!!! Can you explain what is going on here????? How does Affirma.CloudFront.AdminTool compile if the "DistributionSummary.cs" file doesn't exist in your library? Should this file be added? Where is the file "DistributionSummary.cs"? Thanks!

Commented Issue: Half-open connections limit issue? [13120]

gamitech — Fri, 04 Dec 2009 14:41:29 GMT

It seems like that under certain conditions some connections are not closed in a reasonable time and then XP SP2 10 half-open connection limit (as for note described here: http://www.phex.org/mambo/content/view/49/33/) shows up.

My scenario was 15 applications uploading about 20 files one by one (about 50MB in size each). And after 10th file the whole upload process was like suspended right after transfer is added.

It seems like fix in ThreeSharpQuery.Invoke<T> on line 502 to "using (HttpWebResponse httpWebResponse = ..." fixes that issue, at least for my environment.

And isn't it just good practice to make sure all disposable resources are freed after use? That wasn't so essential before SP2 when it was possible to have 65K half-open connections... but with only 10 it seems to be very valuable resource...

Created Issue: StreamResponseToFile doesn't create new file [14249]

optimator — Thu, 24 Sep 2009 15:55:00 GMT

There is an issue using StreamResponseToFile. If a files exists, and if the existing local file contains more bytes than the file that is being downloaded, then the local file's contents will be replaced up to the length of the file from S3, but the remaining existing content of the local file is not replaced.

Two possible solutions: 1) In the user's code, check to see if the file exists, delete it, and then use StreamResponseToFile; or 2) in the Affirma.Threesharp StreamResponseToFile line #86 replace with:

using (FileStream fileStream = new FileStream(localfile, FileMode.Create)

Created Issue: Delete does not work with Mono < 2.4

kkenneth — Sun, 30 Aug 2009 10:05:35 GMT

A DELETE request to S3 always gets a "HTTP/1.1 204 No Content" statuscode as the response.
In the method "ThrowIfErrors" the check for a valid response, only includes "OK" and "TemporaryRedirect" but not "NoContent".

The next lines attempt to read the response stream, which locks up under Mono 2.0 and older.

A simple fix is to change the line:
if (httpWebResponse.StatusCode != HttpStatusCode.OK && httpWebResponse.StatusCode != HttpStatusCode.TemporaryRedirect)

to:
if (httpWebResponse.StatusCode != HttpStatusCode.OK && httpWebResponse.StatusCode != HttpStatusCode.TemporaryRedirect && httpWebResponse.StatusCode != HttpStatusCode.NoContent)

But actually, any code below 399 is not an error, so the statement could also be changed to:
if ((int)httpWebResponse.StatusCode >= 400)

Commented Issue: Make Model members virtual

mattiasjex — Tue, 28 Jul 2009 15:59:56 GMT

Make model members virtual, please. Without this, it is effectively impossible to mock the IThreeSharp, since I cannot mock the response objects. I am willing to contribute a patch to this if you give me read access to source control.
Comments: ** Comment from web user: mattiasjex **

This is especially important for the response objects.

Commented Issue: Make Model members virtual

mattiasjex — Tue, 28 Jul 2009 15:58:33 GMT

Great software, though!

Created Issue: Make Model members virtual

mattiasjex — Tue, 28 Jul 2009 15:58:04 GMT

Created Issue: Provide a default stream buffer size in ThreeSharpConfig

Crenna — Mon, 20 Jul 2009 01:18:15 GMT

There is currently a hard-coded stream buffer size in ThreeSharpQuery to transfer 1024 bytes at a time which is way too small for some operations. It is easy enough to patch but this should be a part of the ThreeSharpConfig, not a hard-coded int.

Created Issue: compilation error in sample code in Readme.txt

icnocop — Mon, 13 Jul 2009 17:57:10 GMT

Hi.

Thank you for ThreeSharp.

I just wanted to point out a compilation error with the sample code included in the Readme.txt file.

This line of code:
using (ObjectGetResponse response = service.ObjectGet(objectGetRequest)) {
should be this:
using (ObjectGetResponse response = service.ObjectGet(request)) {

:)

Created Issue: LoadStreamWithStream function doesn't work with a new MemoryStream object.

dvela1 — Tue, 30 Jun 2009 01:13:18 GMT

When using a new MemoryStream() to pass to "LoadStreamWithStream".
You need to reset the position to zero or it starts at the end and doesn't read the data.

Fixed code below...

private HttpWebRequest GenerateAndSendHttpWebRequest(Request request)
{
HttpWebRequest httpWebRequest = ConfigureWebRequest(request);

if (httpWebRequest.Method == "PUT" || httpWebRequest.Method == "POST")
{
//httpWebRequest.ContentLength = request.BytesTotal;
using (Stream httpWebRequestStream = httpWebRequest.GetRequestStream())
{
if (request.DataStream != null)
{
// Do some statistical tracking
request.TransferInfo = new TransferInfo(request);
lock (this.statistics) { this.statistics.AddTransferInfo(request.TransferInfo); } // All operations on this.statistics must be thread-safe

byte[] buffer = new byte[1024];
int bytesRead = 0;

//Make sure to reset position to the beginning
request.DataStream.Position = 0;

while (true)
{
bytesRead = request.DataStream.Read(buffer, 0, buffer.Length);
if (bytesRead == 0)
break;
httpWebRequestStream.Write(buffer, 0, bytesRead);
request.TransferInfo.BytesTransferred += bytesRead;
}

request.DataStream.Close();
}
}
}

return httpWebRequest;
}

Created Issue: No method to convert ACLGetResponse to ACLPolicy

BrandonHaynes — Thu, 28 May 2009 18:22:28 GMT

There currently exists no method to convert an ACLGetResponse into an ACLPolicy object. I propose adding this as a method on ACLGetResponse.

I am currently implementing this functionality via extension methods; feel free to integrate these into a public method:

public static ACLPolicy ToPolicy(this ACLGetResponse aclResponse)
{
var policyDocument = aclResponse.StreamResponseToXmlDocument();
return new ACLPolicy()
{
OwnerId = policyDocument["AccessControlPolicy"]["Owner"]["ID"].InnerText,
OwnerName = policyDocument["AccessControlPolicy"]["Owner"]["DisplayName"].InnerText,
Grantees = policyDocument["AccessControlPolicy"]["AccessControlList"].GetElementsByTagName("Grant").ToGrantees().ToList()
};
}

private static IEnumerable<ACLGrantee> ToGrantees(this XmlNodeList grantNodes)
{
return grantNodes
.Cast<XmlNode>()
.Select(node => new ACLGrantee()
{
DisplayName = node["Grantee"]["DisplayName"].InnerText,
ID = node["Grantee"]["ID"].InnerText,
Permission = ParsePermission(node["Permission"].InnerText),
Type = ParseGranteeType(node["Grantee"].Attributes["xsi:type"].Value)
});
}

private static ACLGrantee.GrantPermission ParsePermission(string xmlPermission)
{
switch (xmlPermission)
{
case "FULL_CONTROL": return ACLGrantee.GrantPermission.FullControl;
case "READ": return ACLGrantee.GrantPermission.Read;
case "WRITE": return ACLGrantee.GrantPermission.Write;
default: throw new ArgumentException();
}
}

private static ACLGrantee.GranteeType ParseGranteeType(string xmlGranteeType)
{
switch (xmlGranteeType)
{
case "CanonicalUser": return ACLGrantee.GranteeType.User;
case "Group": return ACLGrantee.GranteeType.Group;
case "AmazonCustomerByEmail": return ACLGrantee.GranteeType.Email;
default: throw new ArgumentException();
}
}

Created Issue: MaxKeys in BucketListRequest

Regent — Tue, 28 Apr 2009 15:03:00 GMT

It would be nice to have MaxKeys property within the BucketListRequest. (By defaults results are limited to 1000 items)

Related S3 document:
http://docs.amazonwebservices.com/AmazonS3/2006-03-01/index.html?BucketConfiguration.html

Commented Issue: Half-open connections limit issue?

Regent — Mon, 27 Apr 2009 21:02:43 GMT

In attachment is 3 modified files with some changes for guarantee that HttpWebResponse is disposed not later that Response class.
It is rough workaround but with least possible changes. Imho better to hold HttpWebResponse within the class.

Commented Issue: Half-open connections limit issue?

Regent — Mon, 27 Apr 2009 17:47:29 GMT

I found that this fixes the issue with the uploading but as soon as you try to upload anything after this change you'll get nothing because actual stream referenced by Response.DataStream was disposed with httpWebResponse object.

As for me a neat solution is to store HttpWebResponse object within the Response class and dispose it in its dispose method so they are disposed together.

Created Issue: Half-open connections limit issue?

Regent — Mon, 27 Apr 2009 11:12:35 GMT

Created Issue: Check if file exists

maxfloden — Tue, 24 Feb 2009 22:59:00 GMT

Suggestion to add higtechriders extension to get object header as a method of checking if a file exists in s3.
From this thread: http://www.codeplex.com/Thread/View.aspx?ProjectName=ThreeSharp&ThreadId=38570

Commented Issue: LoadStreamWithBytes() doesn't work

jAffirma — Mon, 16 Feb 2009 23:48:55 GMT

Exception:
"The stream does not support concurrent IO read or write operations." at ThreeSharpQuery.cs: GenerateAndSendHttpWebRequest() line 249
Comments: ** Comment from web user: jAffirma **

EU buckets require using a subdomain calling format.

thanks,
Joel Wetzel
Affirma Consulting

Created Issue: add LoadStreamWithStream() method

DuckMaestro — Sun, 04 Jan 2009 08:25:13 GMT

This is a modification I've made locally, but would be cool if it were added to the library officially. This is the method I made locally (on an older 1.x version):

public void LoadStreamWithStream(Stream sourceStream)
{
this.dataStream = sourceStream;

this.contentType = "application/octet-stream";
this.BytesTotal = sourceStream.Length;
}