Re: [bug-gnulib] fwriteerror() interface

2006-04-24 Thread Ben Pfaff
Paul Eggert <[EMAIL PROTECTED]> writes: > In the latest patch, those three gotos confused me. Part of the > confusion is due to uses like 'if (fflush (fp))' which are of course > valid but which I find a bit hard to read at first. Also, the > comment's "contents is" is not quite grammatical, and

Re: [bug-gnulib] fwriteerror() interface

2006-04-24 Thread Paul Eggert
In the latest patch, those three gotos confused me. Part of the confusion is due to uses like 'if (fflush (fp))' which are of course valid but which I find a bit hard to read at first. Also, the comment's "contents is" is not quite grammatical, and the comment is a bit overconfident about getting

Re: [bug-gnulib] fwriteerror() interface

2006-04-24 Thread Bruno Haible
Ben Pfaff wrote on 2006-02-27: > I'm not sure I understand the fwriteerror() interface. The basic idea is to be able to write: output_file = fopen (file_name, "w"); if (output_file == NULL) error (EXIT_FAILURE, errno, _("error while ope

Re: fwriteerror() interface

2006-03-11 Thread Paul Eggert
Ben Pfaff <[EMAIL PROTECTED]> writes: > I still think this is a bug, but no one ever replied and no fix > has been checked into CVS. Can anyone else confirm or deny that > this is in fact a bug? I agree that there is a bug, either in the documentation or in the code. I don't use fwriteerror mys

Re: fwriteerror() interface

2006-03-11 Thread Ben Pfaff
Ben Pfaff <[EMAIL PROTECTED]> writes: > I'm not sure I understand the fwriteerror() interface. As I read > the function comment, fwriteerror() should always close the > stream provided to it (except in the case of stdout). But the > implementation does not close the