[PATCH 1 of 3 STABLE] i18n: fix "% inside _()" problems

[Pierre-Yves David]

On 04/13/2014 09:32 AM, FUJIWARA Katsunori wrote:
> # HG changeset patch
> # User FUJIWARA Katsunori <foozy at lares.dti.ne.jp>
> # Date 1397394948 -32400
> #      Sun Apr 13 22:15:48 2014 +0900
> # Branch stable
> # Node ID 6c252c9898beaedcf3230741445310b0071e5cb4
> # Parent  469d949a7cb8d20e0fee7464c3b12dcac1e55d4c
> i18n: fix "% inside _()" problems

I started to reviewing this series, but my brain started digging an 
escape tunnel in the back of my head.

Here are a few things that would prevent brain melt down when reading it:

1. split the actual fix in core from the checkcode addition: The flow 
would be:

   a. Fixing this silly issue easy to understand
   b. Adding a new brain melting regexp to catch it
   c. Fixing another silly issue
   d. Another crazy regexp
   e. Fixing a third silly issue
   f. Introduce the mother of all regexp

2. includes examples of undetected bad behavior. Typical message I would 
expect:

   check-code: detect blablabla when % are aligned with pluto and alpha 
centaury

   This changeset add regexp to catch these errors:

     _("lks;kgfd;lksgklds" % ";ldfklgskgf;lkds")
     _("lks;kgfd;lksgklds" % (";ldfklgskgf;lkds", 56))


   New need to add this insane regexp because the previous crazy regexp 
did not take in accout … (rest of the brain melting explanation)


3. add a corresponding test when adding the check code rules (missing in 
patches 1)

Those patch are however much welcome. Looking forward to challenge my 
mind again with V2)

-- 
Pierre-Yves David