On Mon, Jun 25, 2007 at 11:42:23AM +0100, Robert Watson wrote: > > On Sun, 24 Jun 2007, Gabor Kovesdan wrote: > > >Robert Watson escribi?: > >>On Sun, 24 Jun 2007, Stanislav Sedov wrote: > >>> > >>>Personally, I see no point in adding -- style options, since the entire > >>>our world build without them. GNU people can always install gnu- ports, > >>>we just should insure they're posix compatible. > >>> > >>>Furthermore, having -- and - style options in one world effectively > >>>broke BSD's perfect look and feel:-) > >> > >>On the other hand -- intentionally breaking scripts that have worked with > >>FreeBSD for years isn't exactly the best way to make end-users happy. > >> > >How you mean this? The current GNU textproc tools have those -- options. > >How would that break scripts, then? > > > >Personally, I don't need or insist on having -- style options, but > >currently we have those and people might have got used to them. If we > >change the available options by just changing to the BSD-licensed ones > >without a deeper look of their functionality, we might break POLA. > > > >Moreover, the BSD-licensed ones have some of those, too as I wrote before, > >thus we just need to document them in the accompanying manpages. It's > >strange, but the manpages don't cover the existing long options, maybe the > >OpenBSD people didn't want people to use them. > > Ah, OK -- I read the e-mail as stating that the options didn't exit in the > OpenBSD tools, yet proposing moving to them, and hence was concerned about > compatibility with existing scripts. My apologizes for not being clear. With the exception of --help and --version, all the GNU style options are available. (However, they are not documented in the man page). The original authors reason --help is not included is that it doesn't make sense to duplicate the man page in the code. I tend to agree with this. I had a little time on my hands today so I sat down and compared the GNU options to the options avaialble in OpenBSD's version and they are compatible. However, I did notice that --ignore-case was missing, so I added that in. In addition, I got an email from a person asking to include '1' and '2' as alises for 'l' and 'r' when doing merges. (See http://www.nabble.com/sdiff-UI-enhancement-t3877253.html for the reasoning behind it. Its quite interesting) It was a very trivial change so I added it in. Both changes I made are attached to this email as a patch. Finally, as per Doug's request, I did some regression testing with mergemaster and did not run into any issues. Hopefully, if anyone is planning on running mergemaster in the future, please consider doing it with this version of sdiff and let me know if you encounter any problems. Thanks. -- Steven Kreuzer http://www.exit2shell.com/~skreuzer
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:13 UTC