Hi, Thanks for clarification. You convinced me very well.
I will update maint-guide for wheezy. (But not for squeeze under deep freeze). There will be good amount of translation updates I need to make. But I am convinced that is what is needed. On Thu, Dec 16, 2010 at 08:20:49AM -0500, James Vega wrote: > On Sat, Dec 11, 2010 at 02:26:57AM +0900, Osamu Aoki wrote: > > Hi, maintainer of uscan code: ... > > My thought is "why we make this pedantic useless cosmetic changes with > > uncommon syntax. It adds confusion and creates resource drains." > > I see no practical benefits of change from .* to .+ in the documentation. > > This syntax is not uncommon. It is very normal and only adds confusion > if someone isn't familiar with PCRE syntax. Knowledge of PCRE syntax is > required to create a watch file that properly expresses what the watch > file should be matching. I agree .+ is common for ERE syntax. This was my over reaction. I ment to say: s/with uncommon syntax/over good enough syntax/. But I now know that I was wrong. > > Seriously, if such null matching is the REAL problem, we have many such > > REGEX in uscan code itself. I have seen Perl REGEX ".*" in uscan. I > > have also seen Perl REGEX like ".*?" in uscan which looks even funnier > > than use of ".*" since it is the same as ".*" and not so common one to > > me. > > No, “.*” and “.*?” are not the same. The former is greedy, the latter > is not. This is an important difference and one which we took into > account when writing uscan. Oh, that's a subtle PCRE thing ... I see. (This is not so common for me to use. Now I recall ...) ... > I'd say it's more about having correct regular expressions. In all of > the cases that were fixed, we're expecting at least 1 character to be > matched so “.+” should be used. Osamu -- To unsubscribe, send mail to [email protected].
