Modifying shell-command-on-region or using call-process-region or both!?

Hi, I've stumbled on an issue when porting the go-mode (go-lang) to XEmacs and needs some advice. In the GNU code shell-command-on-region is called with 6 args. Actually the GNU version takes up to 7 args. (We don't have the last two args. They are an error buffer, where output from stderr goes, and a bool arg whether the error buffer should be displayed or not. And yes, in my current solution I just ignore the separation of stdout and stderr making the XEmacs version just slightly more user unfriendly.) One way of solving the issue for the go-mode is to sync our shell-command-et-al-code with the GNU version. Maybe not that difficult but there will be issues with asynchronous commands since we use the background package for those. I guess we should go for API-compatibility here so it ought to be a good thing, right? On the other hand the functionality that is implemented is a call to the program gofmt to reformat the current buffer and replace it. (Go comes with a program that formats go code for you) This seems ideal to call from call-process-region instead of using shell-command-on-region!? (Am I missing something here, why fire of a new shell just to run that process?) To make things slightly worse it seems like GNUs call-process-region is not compatible with ours so moving to call-process-region is only a solution for XEmacs avoiding the missing features in our version of shell-command-on-region. Views, comments, appreciated! Yours -- %% Mats _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://lists.xemacs.org/mailman/listinfo/xemacs-beta