On 20 July 2012 22:22, Dennis E. Hamilton <dennis.hamil...@acm.org> wrote: > I've seen the comment about SCMs being sufficient for tracing the provenance > of code and the changes that are made. That puzzles me. > > - History doesn't appear in source-code tarballs. > - It requires the original SCM repository or a history-preserving port of > the SCN to be available to interested parties. > > Basically, it is not a durable form of the information.
Indeed, especially since SVN log messages are not versioned. I think log messages should only be used to inform the reader of the commit message why the commit was done. They should not be used for comments that are useful / necessary to the reader of the code; those should be included as comments in the code itself. [Though of course such comments can go in the log message as well.] However, does the end user of the source need to know provenance and history? > Just sayin' ... DItto. > - Dennis > > -----Original Message----- > From: Pedro Giffuni [mailto:p...@apache.org] > Sent: Friday, July 20, 2012 12:52 > To: ooo-dev@incubator.apache.org > Subject: Re: Coding guideline or common rules > > FWIW; > > > ----- Original Message ----- > ... >> >> Hi, >> >> I just stumbled over a commit message for the new UOF filter. >> >> I think we should agree on a common guideline for our code and how we >> contribute changes and bring them in the code. >> >> SCM's manage the change sets and the information who made the change, >> that means we don't need further comments like this >> >> ///Begin Added by wangyumin for uof2-filter from cs2c >> ... >> /// End Added by wangyumin on 2012-2-22 14:32:18 >> >> It is somewhat redundant and makes the code not really better readable. >> Can we agree on the common understanding that we don't need such >> comments and that we don't want them in the code. We should remove such >> comments wherever we see or find them. >> > > Indeed, I did mention in our local svn tutorial that those comments should be > avoided. SVN does a wonderful job maintaining the origin information. > >> Any opinions? > > > As a side note, I recently found similar prominent begin/end lines in another > project and the culprit on that project was the GPLv2 section 2a: > > "You must cause the modified files to carry prominent notices stating > that you changed the files and the date of any change." > > > It's probable that old code from GPLd derivatives still carry such notes. > > Someone will have to clean them ;). > > Pedro. >