"Mark Olbert" <ChairmanMAO@newsgroups.nospam> escribió en el mensaje
news:946j32ts20ko7f4dg28ppn763pvl9vpdal@4ax.com...
>I ran into a bizarre problem using the NET Framework 2 which I think is a
>bug (and at least is a very subtle gotcha if it isn't
> one).
>
> I have to create "c style" structures from binary data in my application.
> The utility method I wrote to do this creates a
> BinaryReader from the binary data, and then uses the various read methods
> (e.g., ReadByte(), ReadBytes(length)) to initialize the
> structure's fields from the binary data.
>
> One of the fields is a fixed length char array (it's always 512 characters
> long). So my utility method originally did something like
> this:
>
> charArray512 = reader.ReadChars(512);
>
> This worked...sometimes. And sometimes it didn't. After digging into it
> for several hours I noticed that the problem was
> data-related (i.e., some data caused the problem, some didn't). I finally
> noticed that whenever the binary stream contained
> character code 43 (hex 2B) the "code 43" values were being skipped. In
> other words, the following binary stream (hex):
>
> F0
> 9B
> 74
> 81
> 2B
> 12
>
> would be read as: F0 9B 74 81 12. The 2B code was simply skipped.
>
> I have no idea why ReadChars() ignores certain character codes (at least
> including 2B). My workaround is to use ReadBytes() instead
> of ReadChars(), and then convert each byte to a char using
> Convert.ToChar(). This is slower, but doesn't drop the 2B codes.
>
> BTW, the encoding on the binary stream (and the reader) is set to ASCII.
>
> Can someone explain to me why the 2B codes are being dropped? Is this a
> bug in ReadChars()?
>
> - Mark

"Alexander Ubillus" <alxubil***@hotmail.com> escribió en el mensaje
news:e7Oib$DXGHA.5096@TK2MSFTNGP03.phx.gbl...
> Well, you are reading chars, the reader is doing some kind of check or
> conversion. So if you want the bytes themselves, you have to use ReadBytes
> but instead of converting to char each byte use
> System.Text.Encoding.ASCII.GetString(byteArr);
> ------------------------
>
> "Mark Olbert" <ChairmanMAO@newsgroups.nospam> escribió en el mensaje
> news:946j32ts20ko7f4dg28ppn763pvl9vpdal@4ax.com...
>>I ran into a bizarre problem using the NET Framework 2 which I think is a
>>bug (and at least is a very subtle gotcha if it isn't
>> one).
>>
>> I have to create "c style" structures from binary data in my application.
>> The utility method I wrote to do this creates a
>> BinaryReader from the binary data, and then uses the various read methods
>> (e.g., ReadByte(), ReadBytes(length)) to initialize the
>> structure's fields from the binary data.
>>
>> One of the fields is a fixed length char array (it's always 512
>> characters long). So my utility method originally did something like
>> this:
>>
>> charArray512 = reader.ReadChars(512);
>>
>> This worked...sometimes. And sometimes it didn't. After digging into it
>> for several hours I noticed that the problem was
>> data-related (i.e., some data caused the problem, some didn't). I finally
>> noticed that whenever the binary stream contained
>> character code 43 (hex 2B) the "code 43" values were being skipped. In
>> other words, the following binary stream (hex):
>>
>> F0
>> 9B
>> 74
>> 81
>> 2B
>> 12
>>
>> would be read as: F0 9B 74 81 12. The 2B code was simply skipped.
>>
>> I have no idea why ReadChars() ignores certain character codes (at least
>> including 2B). My workaround is to use ReadBytes() instead
>> of ReadChars(), and then convert each byte to a char using
>> Convert.ToChar(). This is slower, but doesn't drop the 2B codes.
>>
>> BTW, the encoding on the binary stream (and the reader) is set to ASCII.
>>
>> Can someone explain to me why the 2B codes are being dropped? Is this a
>> bug in ReadChars()?
>>
>> - Mark
>
>