From: David B. <Dav...@mo...> - 2006-05-02 20:25:59
|
John W. Eaton wrote: > I just checked in a largeish change that eliminates all warn_* > built-in variables. Here is the NEWS file entry I wrote for this > change: > > * Previous versions of Octave had a number of built-in variables to > control warnings (for example, warn_divide_by_zero). These > variables have been replaced by warning identifiers that are used > with the warning function to control the state of warnings. Now, > instead of writing > > warn_divide_by_zero = false; > > to disable divide-by-zero warnings, you should write > > warning ("off", "Octave:divide-by-zero"); > > You may use the same technique in your own code to control > warnings. For example, you can use > > warning ("My-package:phase-of-the-moon", > "the phase of the moon could cause trouble today"); > > to allow users to control this warning using the > "My-package:phase-of-the-moon" warning identifier. > > You may also enable or disable all warnings, or turn them into > errors: > > warning ("on", "all"); > warning ("off", "all"); > warning ("error", "Octave:divide-by-zero"); > warning ("error", "all"); > > You can query the state of current warnings using > > warning ("query", ID) > warning ("query") > > (only those warning IDs which have been explicitly set are > returned). > > A partial list and description of warning identifiers is available > using > > help warning_ids > > > The following patch will also adapt Octave Forge to this change and > also removes some old #ifdef cruft. > Want it applied? > BTW, when I updated my copy of Octave Forge, it did not include the > recent changes I posted to handle changes in the structure of the > octave_value classes. Am I missing something? Should I be using CVS > HEAD or a branch to get the latest Octave Forge sources that are > intended to work with the Octave CVS sources? > You have to work against the developer CVS with SSH, as the public CVS is not being synchronized to the developer CVS till the new sourceforge hardware comes on line. > Thanks, > > jwe > |