On Thu, May 25, 2006 at 11:18:44AM -0300, Henrique de Moraes Holschuh wrote:
>On Thu, 25 May 2006, Steve McIntyre wrote:
>> >Notice the file has a broken conflict, which is not detected by cvs status.
>> >Nothing else detects it either, if I commit, I will commit a conflicted file
>> >with conflict hunks.
>> 
>> It seems this is a design decision by CVS upstream - see
>> 
>>   http://www.cygwin.com/ml/cygwin/2006-03/msg00668.html
>
>I did not expect CVS upstream to be actively sabotaging CVS.  Thank you for
>warning me about it.
>
>Please consider adding a very visible warning in the NEWS.Debian file (I'd
>add it as a debconf note of priority high, but to each its own) about it
>ASAP.
>
>> for another description of the same bug. It's rather annoying that
>> they have changed the behaviour so blatantly from what people are used
>> to...
>
>So, after a non-joke merge I have a thousand files I have to manually verify
>for conflicts, because CVS will flag *all* of them the same way in the cvs
>status and cvs update commands (which are almost always the means used to
>verify a working copy before a commit), on top of allowing commits of
>conflicted crap without a warning?
>
>I didn't test a graphical front-end, but people using those expect to have
>those bright-red ! marks on conflicted files after a merge, I guess this
>won't work anymore either.
>
>Need I remind anyone that manually grepping for conflict markers is NOT
>something the majority of CVS users will do, because the tool was, until a
>few days ago, supposed to do it for them?
>
>This change of behaviour is dangerous, and dumb beyond measure.  Doing such
>a change silently just adds to the injury, and it is not acceptable at all.
>
>I hope we can do something about it.

I'm planning on digging into it further over the next couple of days;
I'm very tempted to patch this change back out, if it is this
simple...

-- 
Steve McIntyre, Cambridge, UK.                                [EMAIL PROTECTED]
"...In the UNIX world, people tend to interpret `non-technical user'
 as meaning someone who's only ever written one device driver." -- Daniel Pead

Attachment: signature.asc
Description: Digital signature

Reply via email to