Peter Chen wrote: > > Hi, > > I will probably do more tests this afternoon. Since I am using kylix, it > might be something wrong with kylix configuration. > > I copied a lot of codes from scanimage, except I did not write the image to > stdout like scanimage did. I could not make sane_start work properly if I > called fopen() before sane_start(). Again, it might be something wrong with > my kylix configuration.
Peter, it is a bit difficult to make guesses about possibles causes of the problems you have with your frontend, when you don't give any details. Most if not all Sane backends contain many debug statements. While they are mainly intended to help finding bugs in the backend, they can also be useful to find in bugs in frontends. And it would help if you could provide at least excerpts from your sources, preferably combined with the output of a few fprintf(stderr,...) statements placed near the Sane API calls. With all this information, it will be easier to get a clue, what goes wrong. And I must emphazise Henning's hint to use the test backend. Some months ago, I worked a bit on Andrew Kuchling's Python bindings for Sane. The test backend was incredibly helpful. It is way faster than any real scanner, it provides more types of options than any other Sane backend, it allows to setup a number of uncommon cases, like choosing the number of bytes received. Without all these test cases, my work on the the Python Sane module would have been far less successful and would have taken far more time. Abel
