CloudBlockBlob.DownloadText() handles UTF8 BOM incorrectly about azure-storage-net HOT 7 CLOSED

@joefeser 👍 I'm way late to the party, but ...

We could detect and strip out the BOM ourselves, but this is a potentially surprising transformation of the blob's original data

It would have been nice though to have an overload .DownloadText(bool stripByteOrderMark), so that the original method would just do the un-"surprising" thing and the overload would allow the dev to explicitly call for stripping the BOM.

from azure-storage-net.

Seriously, 5 months with no comment?

from azure-storage-net.

I created an extension method DownloadString for CloudBlockBlob and used the same code which is used in WebClient.DownloadString. Here you can find the code: How to get rid of BOM when downloading text from azure blob

from azure-storage-net.

Closed my issue. Found the same issue.

from azure-storage-net.

Apologize for the delay.

DownloadText api is currently a simple utility api that takes the bytes and passes its to corresponding encoder to get a string back. It was not designed to map one to one with File.ReadAllText().

We will evaluate a fix for above issue for our future release. This can only be part of our next breaking change release as it changes the API behavior. For now, you can use the workaround suggested by @rzio

from azure-storage-net.

@vinaysh-msft If it passed it to the correct encoder, it should have dealt with the BOM correctly and not returned it as part of the string.

from azure-storage-net.

We have evaluated this issue. When a blob is downloaded using DownloadText(), we apply an encoding to construct a string (by default, we use .NET's UTF-8 implementation). Applying the encoding to the string does not strip out the BOM. We could detect and strip out the BOM ourselves, but this is a potentially surprising transformation of the blob's original data, especially given that the Content-Encoding header might also be specified on the blob with conflicting info.

See also that the same behavior exists in WebClient.DownloadString: http://stackoverflow.com/questions/1317700/strip-byte-order-mark-from-string-in-c-sharp

Note that you can pass your own Encoding implementation that strips out the BOM in the call to GetString() if this behavior is desired.

I am closing this old issue now; feel free to raise another issue if you have additional concerns related to encoding.

from azure-storage-net.