On Sun, Feb 19, 2017 at 4:20 AM, Tilman Hausherr <[email protected]> wrote:
> Am 18.02.2017 um 17:15 schrieb Thad Humphries: > >> I am processing images, mostly TIFF Group 4, from a repository, and >> outputting them as PDF. The images *always* come to me in a byte array, >> never as a file. In fact, I've no option to get it as a file. The >> repository may or may not be local--it's accessed through its own API. >> >> I would like to convert this buffer directly to a PDImageXObject, but both >> the CCITTFactory.createFromRandomAccess() are deprecated. Why? Looking at >> the source for the CCITTFactory class, they do the same thing as the >> createFromFile() methods, which first create a RandomAccessFile before >> calling createFromRandomAccessImpl(). >> >> Outside of this project, I've tried CCITTFactory.createFromFile() and seen >> that it is significantly faster than creating a BufferedImage and calling >> CCITTFactory.createFromImage(). Possibly it's fast enough to take the hit >> of writing the TIFF byte array to a temporary file, but why take the hit? >> >> So, why are CCITTFactory.createFromRandomAccess() are deprecated? >> >> > That was done in the last commit of > > https://issues.apache.org/jira/browse/PDFBOX-2589 > > There was no discussion, probably nobody bothered about it. I assume it > looked cleaner than this weird RandomAccess thing. > > I'd agree that "The images *always* come to me in a byte array" is a good > argument to un-deprecate the two methods. > > Please create a new issue in JIRA and copy your argument there. I'll > undeprecate the methods but will do this after some wait to give John time > to state his opinion. > > Tilman > Done: https://issues.apache.org/jira/browse/PDFBOX-3696 Thank you. -- "Hell hath no limits, nor is circumscrib'd In one self-place; but where we are is hell, And where hell is, there must we ever be" --Christopher Marlowe, *Doctor Faustus* (v. 121-24)
