On Wednesday 03 May 2006 18:50, Francois Pepin wrote: > How much of the file is generally being read before the guess is made? > I'm thinking very little is needed, especially compared to how much > memory Java usually takes.
Generally not much. Jmol uses 16384 bytes. > It would not be very difficult to save that first part of the stream and > then play it back once the guess is made. See how Jmol does it: http://svn.sourceforge.net/viewcvs.cgi/jmol/trunk/Jmol/src/org/jmol/adapter/smarter/Resolver.java?view=markup > I kind of like the idea of using streams, in cases where you are not > reading from a file. Having to write everything to a temporary file to > satisfy the API isn't a very appealing solution, I think. > > I could code something up if people are interested. An additional advantage is that you get .gz support in one go: BufferedInputStream bis = new BufferedInputStream((InputStream)t, 8192); InputStream is = bis; bis.mark(5); int countRead = 0; countRead = bis.read(abMagic, 0, 4); bis.reset(); if (countRead == 4 && abMagic[0] == (byte)0x1F && abMagic[1] == (byte)0x8B) is = new GZIPInputStream(bis); where t is your InputStream, and is the stream to use after the gzip check/unzip. For the full working code, see again Jmol CVS: http://svn.sourceforge.net/viewcvs.cgi/jmol/trunk/Jmol/src/org/jmol/viewer/FileManager.java?view=markup Egon -- [EMAIL PROTECTED] Cologne University Bioinformatics Center (CUBIC) Blog: http://chem-bla-ics.blogspot.com/ GPG: 1024D/D6336BA6 _______________________________________________ Biojava-l mailing list - [email protected] http://lists.open-bio.org/mailman/listinfo/biojava-l
