svn commit: r281129 - head/etc

[Baptiste Daroussin]

On Mon, Apr 06, 2015 at 11:49:37AM +0300, Gleb Smirnoff wrote:
> On Mon, Apr 06, 2015 at 11:41:41AM +0300, Slawa Olhovchenkov wrote:
> S> > S> > >> On Mon, Apr 06, 2015 at 01:32:27AM +0000, Dmitry Marakasov wrote:
> S> > S> > >> 
> S> > S> > >> Author: amdmi3 (ports committer)
> S> > S> > >> Date: Mon Apr  6 01:32:26 2015
> S> > S> > >> New Revision: 281129
> S> > S> > >> URL: https://svnweb.freebsd.org/changeset/base/281129
> S> > S> > >> 
> S> > S> > >> Log:
> S> > S> > >>  Switch russian login class to UTF-8
> S> > S> > > 
> S> > S> > > As discursed before this is may cause data lost.
> S> > S> > 
> S> > S> > Why?
> S> > S> 
> S> > S> Create file in koi8 locale with russian charaters.
> S> > S> Update OS. Now you have utf-8 locale.
> S> > S> Open file in vi.
> S> > S> Save file.
> S> > S> Done. Some data lost.
> S> > S> vi and some other programs (tr and may by other) handle not-utf8 byte secquence in non-safe way.
> S> > 
> S> > Don't save file.
> S> 
> S> You kidding.
> S> vi may inform you about data lost only postfactum (in case no lines
> S> with koi8 characters currently on screen).
> 
> No, I'm not.
> 
> You open an application, the application does something weird. Do you usually
> select "Save" after that? For example, you open Microsoft Word file in OpenOffice,
> and you see that file is definitely mangled. Do you select "Save"?
> 
> If you do, you are the one to blame. Don't blame application developers. The
> application didn't do implicit save.
> 
He is pointing at a vi bug:
See https://lists.freebsd.org/pipermail/freebsd-arch/2015-January/016702.html

Noone seems to have worked on it since :(

Best regards,
Bapt
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://lists.freebsd.org/pipermail/svn-src-all/attachments/20150406/cbbd67c1/attachment.sig>