Hardcode

As a follow-up to this post I want to say that about 50% of Google hits to this blog are queries like "delete row from datatable", "how can we remove row from datatable", etc. The Microsoft designers of the .NET Data Layer should better be looking for another job ...

Anyway, here's the solution:

dataTable.Rows[0].Delete(); // marks the row as deleted
dataAdapter.Update(dataTable); // performs the actual delete

Write an SQL query to "stick" two tables together by columns. For example, if you have Table1 and Table2 the following:

Table1          Table 2

A | B           C | D

1 | 2           x | y

3 | 4

The result should be :

A | B | C | D

1 | 2 | x | y

3 | 4 | NULL | NULL 

They charge $12.95 montly for subscribing to their site for viewing solutions to questions posted, take a look here for example. Now only if the solutions were not available on the same page, scroll to a third of the page to see the answers. Now this is utterly stupid, I am seriously wondering if they have any subscribers.

Still, I certainly hope to keep the solutions free (they could hide them better in the HTML source or something :) to filter out the users).

By far the most annoying issue in Windows is the Undo function in Windows Explorer. Suppose you moved or renamed a file or folder 15 minuts ago then you press Control-Z by mistake  - without a warning or something it gets moved back, you have then to remember what you did and go redo the change (there's no Redo in Windows Explorer :) ). To make this even more tricky, Windows Explorer doesn't highlight the object or give you any indication about it.

I seriously hope this was left out of Vista. 

Search for Evensys and check the fifth result. It seems that Google's URL parser got fooled by the following link:

<a href="http://www.umbrire.ro sau toplus.clubafaceri.ro/info_stiri/2006/11/0 ...

It seems that Internet Explorer also parses the URL incorrectly - it goes to toplus.clubafaceri.ro/...

I've ended up rewriting some part of MFC (the OLE Server / Client data transfer code) , using a memory-usage-oriented approach.

The resulting code is extremely memory-efficient, data is never duplicated between the server and the client (however because of shared memory via CreateFileMapping the task manager mem usage appears to be double the data size, even if it isn't).

Well it could have been further more optimized using data compression but I needed to stop at some point.