 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Java Forum / First Aid / April 2004reading/writing a byte array in java
Hi Gang: byte[] bytes; I have a byte array as declared above that has been filled with the data from a .jpg file. Let's say that I wish to find the index of the fifth occurrence of the hex char 0xff from the start of the byte array. What is wrong with the following? I know its got to be the way I am comparing in the if statement. This works in C but I guess I need something different in Java? Thanks! Drew for (x = 0, ffcounter = 0; x < bytes.length; x++) { if (bytes[x] == 0xff) // counter for ff values ffcounter++; // on the 5th one, exit for loop so // that x has index of 5th ff if (ffcounter == 5) break; } // index of the 5th occurrence of 0xff is value of 'x' > for (x = 0, ffcounter = 0; x < bytes.length; x++) > { [quoted text clipped - 7 lines] > break; > } Several problems for me. 1. the canonical loop counter goes for ( int i=0; i<bytes.length; i++) so i is not known outside the loop. It confuses people when you violate canonical forms. If you want to know i outside the loop, store it in a separate variable to make it very clear what you are doing and make i known only inside the loop. C people will likely disagree with me, Algol people will likely agree. It is a matter of creating least surprise. 2. The for loop has a built in stopping mechanism i<bytes.length. Use that. It is where people expect to find the stopping criteria. You might consider ffcounter < 5 && i < bytes.length 3. Fortran background people would never use x as a loop counter. It is reserved as a float or double temporary. -- Canadian Mind Products, Roedy Green. Coaching, problem solving, economical contract programming. See http://mindprod.com/jgloss/jgloss.html for The Java Glossary. >> for (x = 0, ffcounter = 0; x < bytes.length; x++) >> { [quoted text clipped - 17 lines] > me, Algol people will likely agree. It is a matter of creating least > surprise. Did I miss something? The original posting did not create the variable inside the for loop. > 2. The for loop has a built in stopping mechanism i<bytes.length. Use > that. It is where people expect to find the stopping criteria. You > might consider ffcounter < 5 && i < bytes.length ... unless there is a "break" statement. This is actually quite common usage. > 3. Fortran background people would never use x as a loop counter. It > is reserved as a float or double temporary. But that doesn't really affect the outcome of the program, does it? > -- > Canadian Mind Products, Roedy Green. > Coaching, problem solving, economical contract programming. > See http://mindprod.com/jgloss/jgloss.html for The Java Glossary. >Did I miss something? The original posting did not create the variable >inside the for loop. I know. It is legit C and Java, but illegitimate in other languages to use a loop counter outside the loop. It is not the canonical form of a loop. I think it should be avoided if possible. It is a weird thing to do, it deserves special note by storing the loop counter is a separate variable where you exit the loop prematurely. You just confuse the heck out of people by writing code where loop counters live outside the loop. You need to red flag it in some way. His code with a comment would suffice. -- Canadian Mind Products, Roedy Green. Coaching, problem solving, economical contract programming. See http://mindprod.com/jgloss/jgloss.html for The Java Glossary. >... unless there is a "break" statement. This is actually quite >common usage. The for loop's duty is to stop when an incrementing count hit a value. Break is for other sorts of premature exit. Break is a sort of GOTO that should be avoided if possible. You could write the code as a while loop, but the general principle is to use the most specific tool with the most idiomatic code. -- Canadian Mind Products, Roedy Green. Coaching, problem solving, economical contract programming. See http://mindprod.com/jgloss/jgloss.html for The Java Glossary. >> 3. Fortran background people would never use x as a loop counter. It >> is reserved as a float or double temporary. > >But that doesn't really affect the outcome of the program, does it? No, but it makes you program more confusing since in violates a long standing tradition that other programmers expect. i,j,k are ints. x,y,z are floating point. Similarly you make packages start with lower case, variable lower case, and classes upper case. Code still works otherwise, but it makes the code very confusing to someone used to trusting the convention. -- Canadian Mind Products, Roedy Green. Coaching, problem solving, economical contract programming. See http://mindprod.com/jgloss/jgloss.html for The Java Glossary. >>> 3. Fortran background people would never use x as a loop counter. It >>> is reserved as a float or double temporary. [quoted text clipped - 4 lines] > standing tradition that other programmers expect. i,j,k are ints. > x,y,z are floating point. To you perhaps. However, we're writing Java. In java, there is no such tradition. > Similarly you make packages start with lower case, variable lower > case, and classes upper case. Code still works otherwise, but it makes [quoted text clipped - 4 lines] > Coaching, problem solving, economical contract programming. > See http://mindprod.com/jgloss/jgloss.html for The Java Glossary. >> No, but it makes you program more confusing since in violates a long >> standing tradition that other programmers expect. i,j,k are ints. >> x,y,z are floating point. > > To you perhaps. However, we're writing Java. In java, there is no > such tradition. I'm pretty sure that 99% of all loops out there are using i/j/k as a counter variable.  SignatureKind regards, Christophe Vanfleteren Hi Roedy: Thanks for the thoughts. Really, the problem I'm having is why the logic of the if statement doesn't work. Perhaps there is room for improvement in the style of the code fragment, but that is not the issue here. Thanks Drew >> for (x = 0, ffcounter = 0; x < bytes.length; x++) >> { [quoted text clipped - 24 lines] >3. Fortran background people would never use x as a loop counter. It >is reserved as a float or double temporary. > Hi Gang: > [quoted text clipped - 15 lines] > { > if (bytes[x] == 0xff) // counter for ff values A `byte' has a value in the range -128..+127. You will notice that 0xFF is 255, which is not in that range. Hence, no `byte' value can ever satisfy this `if' statement. Two possible fixes: if (bytes[x] == -1) ... if ((bytes[x] & 0xFF) == 0xFF) ... SignatureEric.Sosman@sun.com >if ((bytes[x] & 0xFF) == 0xFF) ... for more details see http://mindprod.com/jgloss/unsigned.html -- Canadian Mind Products, Roedy Green. Coaching, problem solving, economical contract programming. See http://mindprod.com/jgloss/jgloss.html for The Java Glossary. Eric: Just wanted to thank you for answering my question. You nailed it! I should have realized that! I guess sometimes you just stare at something so long you can't see the obvious. Thanks for the help. THe -1 solution works perfectly! Drew >> Hi Gang: >> [quoted text clipped - 25 lines] > > if ((bytes[x] & 0xFF) == 0xFF) ... >>Hi Gang: >> [quoted text clipped - 25 lines] > > if ((bytes[x] & 0xFF) == 0xFF) ... Or even more readable: if (bytes[x] == (byte)0xff)) SignatureKnute Johnson email s/nospam/knute/ Molon labe... > >>Hi Gang: > >> [quoted text clipped - 27 lines] > > Or even more readable: if (bytes[x] == (byte)0xff)) i thought that a byte had a value from -255 to 255. i might be wrong if the code doesn't work. ben >> Or even more readable: if (bytes[x] == (byte)0xff)) > > i thought that a byte had a value from -255 to 255. i might be wrong > if the code doesn't work. No it doesn't. Since a byte (by eight) consists of 8 bits, you can only use the range from 00000000 to 11111111. When using signed bytes (the only ones available in Java), you lose 1 digit for the sign, so there are only seven left for the "value" itself. With those 7, you can make 128 (2^7) combinations. So that gives us a range from -128 (-2^7) to 127 (2^7-1), when using the first bit as sign bit. when you don't use a sign bit, you can use all of the 8 bits, so that gives you 2^8 possible values, which is the range from 0 to 256. SignatureKind regards, Christophe Vanfleteren >When using signed bytes (the only ones available in Java), you lose 1 digit >for the sign, so there are only seven left for the "value" itself. You don't have to lose it. See http://mindprod.com/jgloss/unsigned.html for how to get it back as a unsigned data bit. -- Canadian Mind Products, Roedy Green. Coaching, problem solving, economical contract programming. See http://mindprod.com/jgloss/jgloss.html for The Java Glossary. Free MagazinesGet these publications absolutely FREE for up to 12 months. There are no hidden fees and no obligation. Simply choose a title, complete the application form and submit it. Read more ...
|
Reply to this Message
|
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2008 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.