>>>> "Jerry" == Jerry James <james(a)xemacs.org&gt; writes:

> ;; #### jj, are C++ modules possible?

>>>>>"Jerry" == Jerry James <james(a)xemacs.org&gt; writes: >>;; #### jj, are C++ modules possible?

Moving to XEmacs Beta. This is not a comment on this patch in particular. "Ben Wing" <ben(a)xemacs.org&gt; wrote: [snip] > cvs server: Diffing lisp > Index: lisp/autoload.el [snip] > - (generate-file-autoloads-1 file funlist)) > + (generate-autoload-ish-1 > + file > + (replace-in-string (file-name-nondirectory file) "\\.elc?$" "") > + nil #'generate-file-autoloads-1 > + funlist)) > ;; #### jj, are C++ modules possible? I don't rant very often, so I'm going to take the opportunity to let one loose.:-) This rant is about the last line above, which is *not* part of Ben's patch, so I'm not ranting in his direction. Here is a comment, apparently directed to me, stuck in the middle of this code. I've only been a reviewer for a very short time, and this was apparently put in prior to that time. I wasn't subscribed to xemacs-patches then, so I didn't see it. Why wasn't it sent to me in email, for heaven's sake? (The answer to the question posed above, by the way, is a definite maybe. We certainly can't deal with mangled C++ names, but wrapping the symbols the module code needs to see in extern "C" might be sufficient. However, I don't know what other beasties might raise their ugly heads. Somebody will just have to try it and see what happens.) Along the same lines I recently stumbled, quite by accident, across a ChangeLog entry that says, in part: * make-docfile.c: * make-docfile.c (read_c_string): * make-docfile.c (scan_c_file): Fix shadowing warnings. NOTE: This was already fixed awhile ago, but reverted by Jerry. Please be careful. Same rant: why was I not sent email about this? There were no comments in the code explaining that the "buf" -> "buff" name change was to prevent a shadowing warning. There were no remarks in the ChangeLog that said anything like this either. There are no shadowing warnings on any of my Unix test platforms, so I can only guess that the warnings occur on Some Other Platform, with developers so braindead that they think defining the symbol "buf" in system headers is a good idea. I sent the patch that made the name change with a ChangeLog entry that explicitly stated that I was making this particular change to reduce the size of the diff against the FSF version, and it was approved. Just how, exactly, was I supposed to "be careful"? I agree that somebody was not careful, but it was not me. It was the person who originally fixed the warning and didn't leave any clues behind to let anybody else know what they had done. But I'm the one being fingered as careless in a file that is now being circulated around the globe. I would have pointed all this out if somebody had bothered to send me email first! So the point of this much-too-long rant is to send email when you have something to say to another developer. Inserting comments into code and ChangeLogs is not effective communication. Okay, I'm done ranting now. Time for a nice relaxing dip in the ocean. Oh wait, I'm in Kansas, darn it....

-- Jerry James, native Californian, currently reaching for the Prozac http://www.ittc.ku.edu/~james/