Re: Fix warnings related to po (issue 551830043 by address@hidden)

[dak]

On 2020/04/25 08:36:02, hahnjo wrote:
> On 2020/04/24 20:14:36, hahnjo wrote:
> > On 2020/04/24 18:25:33, lemzwerg wrote:
> > > LGTM, thanks a lot!
> > > 
> > >
https://codereview.appspot.com/551830043/diff/573740043/po/GNUmakefile
> > > File po/GNUmakefile (right):
> > > 
> > >
> >
>
https://codereview.appspot.com/551830043/diff/573740043/po/GNUmakefile#newcode14
> > > po/GNUmakefile:14: sed-makestuff = -e
's/[a-zA-Z_/]*make\[[0-9]*\].*//'
> > > shouldn't this be rather '[0-9]\+'?
> > > 
> > >
> >
>
https://codereview.appspot.com/551830043/diff/573740043/po/GNUmakefile#newcode15
> > > po/GNUmakefile:15: sed-edstuff = -e 's/[ \.,adic0-9]*//' -e
's/---//' | sort
> > -u
> > > I think that within '[...]' you don't have to escape '.'.
> > > 
> > > And again I would use '\+' instead of '*'.
> > 
> > To be honest, I have no clue what all of this is doing: I copied it
from
> > stepmake/stepmake/podir-targets.make.
> 
> After looking through the code once more, this sed magic is only used
for
> po-changes. This spams the output when running po-update which is
pointless IMO
> because the one updating lilypond.pot is not going to update all
languages at
> the same time. Thus, I'm dropping it in the updated patch set.

At any rate, I don't think the backslash is intended to escape the
period rather than just be literally enclosed in the character list. 
Some people tend to do this in the form of the redundant \\ in order to
not confuse human readers and almost but not quite syntax-aware editors,
but with regard to the actual interpretation, it is not doing anything. 
Not even escaping the backslash itself: something like [A-\\-~] is still
two ranges rather than a single range followed by two isolated
characters, the interpretation that would result from "proper" backslash
escaping.

https://codereview.appspot.com/551830043/