[Lazarus] bug report 12594

[ABorka]

On Tue, 30 Mar 2010 08:10:34 -0300
Flávio Etrusco <flavio.etrusco at gmail.com> wrote:

> On Mon, Mar 29, 2010 at 4:39 AM, Mattias Gaertner
> <nc-gaertnma at netcologne.de> wrote:
> > On Mon, 29 Mar 2010 02:03:08 -0300
> > Flávio Etrusco <flavio.etrusco at gmail.com> wrote:
> >
> >> On Sun, Mar 28, 2010 at 5:08 PM, Mattias Gaertner
> >> <nc-gaertnma at netcologne.de> wrote:
> >> > On Sun, 28 Mar 2010 16:34:42 -0300
> >> > Flávio Etrusco <flavio.etrusco at gmail.com> wrote:
> >> >
> >> >>[...]
> >> >> If that's not desirable, shouldn't the .lrs files force a codepage in
> >> >> its "header"? Otherwise we get spurious changes when the files are
> >> >> generated in different systems...
> >> >
> >> > lrs files are plain ASCII. They are the same in any 8bit codepage.
> >> >
> >> > What "spurious changes"?
> >> >
> >> > Mattias
> >>
> >> Do you mean "plain ASCII" or "plain text"? AFAICS they are generate
> >> with sytem's default codepage, which for most modern Linux systems is
> >> UTF8, different to Windows' native/ANSI codepages. At least this is
> >> how a I interpreted the spurious changes that were generated when I
> >> added images to the IDE and needed to (re)generate the .lrs files.
> >
> > The lrs only use characters from #10,#13,#32..#126.
> > The line endings are different, but this is handled by a good version
> > control system automatically.
> >
> > Some months ago the lrs comment was translated, which created
> > differences. Maybe you checked with an old version?
> >
> > Mattias
> >
> 
> I don't know. How do you explain these changes:
> http://svn.freepascal.org/cgi-bin/viewvc.cgi/trunk/images/bookmark.lrs?root=lazarus&r1=24264&r2=24263&pathrev=24264

The special character #127 is now written as constant instead of
string. This was changed a long time ago, but apparently there are
still some even older lrs files.
They need to be updated.

Mattias