Em Tue, 17 Jun 2025 18:23:22 +0100 Donald Hunter <donald.hunter@xxxxxxxxx> escreveu: > On Tue, 17 Jun 2025 at 17:00, Mauro Carvalho Chehab > <mchehab+huawei@xxxxxxxxxx> wrote: > > > > > > (2) is cleaner and faster, but (1) is easier to implement on an > > > already-existing code. > > > > The logic below implements (1). This seems to be the easiest way for > > pyyaml. I will submit as 2 separate patches at the end of the next > > version. > > > > Please notice that I didn't check yet for the "quality" of the > > line numbers. Some tweaks could be needed later on. > > Thanks for working on this. I suppose we might be able to work on an > evolution from (1) to (2) in a followup piece of work? Yes, it shouldn't be hard for you to migrate to (2) in the future. Currently, the parser is stateless, but, as there's now a class, IMO the best would be to store the lines as an array of tuples inside the YnlDocGenerator class, like this: self.lines = [ (line_number1, message_string1), (line_number2, message_string2), (line_number3, message_string3), (line_number4, message_string4), ... ] This way, the parse_yaml_file() method would just return self.lines. Thanks, Mauro
