On Thu, 13 Sep 2018, 22:23 Branko Čibej, <br...@apache.org> wrote: > On 13.09.2018 13:08, Branko Čibej wrote: > > On 13.09.2018 13:01, Stefan Sperling wrote: > >> It should not be too hard to automate the process. The release.py > >> script already has some commands which spit out HTML snippets for > >> the news pages so it might as well be parsing CHANGES and produce > >> HTML we can paste into the website. > > +1. I have a bit of time tonight to have a go at the scripting. > > > r1840861 is a very simple first attempt that only does regular > expression replacement. In particular, here are some things it can't do: > > * Convert issue numbers that aren't decorated with a '#' — there are a > number of places where in a list of issues only the first number > starts with #, but others don't. > * Detect revision numbers that don't start with an 'r'. In some places > the 'r' is missing, but especially in early entries we write > 'revision X' instead of 'rX'. > * Understand fancy revision number range shorthands, e.g., 'r1704854, > -88'. > * Convert old, pre-ASF revision numbers to links in the ASF > repository. This is a direct result of doing simple string > replacements instead of trying to interpret the string contents. > > Some of these issues can be solved by editing the CHANGES file to add > the missing 'r' and '#' decorations. Adding a regular expression that > detects 'revision X' is simple. Automagically converting pre-ASF > revision numbers will require a different approach to the conversion, at > least for revision numbers. >
Oh, the script also doesn't html-escape the rest of the text, but that doesn't seem to matter for now.
