A: When it's by design.
If you are a denizen of the forums (I have no choice, I help run ours at eggheadcafe.com) --and the newsgroups, you learn a lot from reading posts.. Of 100 posts indignantly claiming to have found "A BUG" in the .NET Framework, there is rarely more than one that can lay claim to have found a legitimate bug, and usually, if they do, it's one that has already been covered 100 times over via newsgroup posts and KB Articles, had they only taken the time to search first.
I say this once again because I just received an unsolicited email from somebody who had posted his newly found "BUG" at the LadyBug site asking me to vote for it! Good God, it was simply a minor variant of the fact that if you import XML into a DataSet either in the designer or via ReadXml and it has repeating nodes with the same name, the DataSet can't parse it because it would create duplicate table names. Same thing with nested relations.
Er, that's not a bug - its a good thing!
The guy even quoted text from a post I had made some time ago with my suggestions, apparently believing that reminding me what I wrote six months ago would support my remembering the rationale to "vote" for his bug submission:
Peter Bromberg [C# MVP] wrote:
> The dataSet is doing its best to infer a schema when reading your Xml
> document, and it's finding two sets of elements with the same name,
> and doesn't know how to handle it.
> Experiment if you can with changing the name of one of the sets of
> elements and the error will go away.
> This is very common when reading Xml that doesn't correspond to the
> expected Schema for a DataSet.
What am I talking, Greek? Doods, if you think you've found "a bug", please do some research first. It's usually a bug of the genus userstupidius.
Real bugs in the .NET Framework are extremely difficult to find, and they usually ARE NOT found by the type of developer who posts "I found a BUG in...." to newsgroups.
This post has been brought to you by the letters B, U and ... G.