Re: Is freebsd-update braindamaged, or I'm using it wrong?

2013-10-01 Thread Tim Kientzle
Another thing I'd like to see:

* Have the tool prompt per-directory first, then per-file.

For example:

   There are 30 changed files in /etc/rc.d.
   Update all?  [y/N]

If you hit 'y' it updates all of them right away.  If you hit 'N", it will 
prompt you for each separate file.

Tim


On Oct 1, 2013, at 2:38 AM, Alexander Yerenkow  wrote:

> To make better tool (than current behaviour of mergemaster regarding
> configs/source files) which could make merge an easy task it *must* have
> such things:
> 
> a) way to get Original configs/files from revision from you are upgrading
> ($Rev1)
> b) way to get Original configs/files from revision to you are upgrading
> ($Rev100)
> c) have ability to ignore differencies in comments
> d) have ability to treat special cases (as $FreeBSD$ - just took newer line)
> 
> Then, your each new file will be  $Rev100 + diff_changes(CURRENT, $Rev1) +
> diff_changes($Rev100, $Rev1).
> Note, that in case that your diffs are none  diff_changes(CURRENT, $Rev1) =
> 0, then you can simply get new file.
> Same thing in case that only $FreeBSD$ changed.
> 
> I have some PoC-es for this, but not in shell, maybe I'll come up someday
> with full tool.
> 
> 
> 
> 
> 2013/10/1 Ivan Voras 
> 
>> This is the first time I've used freebsd-update in years, and I'm
>> immediately flagging it as something I won't use in the future. For
>> the last half hour it has been forcing me to manually resolve, one by
>> one, in an editor, hundreds of "merge conflicts" such as these:
>> 
>>  1 <<< current version
>>  2 # $FreeBSD: release/9.0.0/etc/gettytab 209954 2010-07-12 19:09:18Z bcr
>> $
>>  3 ===
>>  4 # $FreeBSD: release/9.2.0/etc/gettytab 243623 2012-11-27 19:23:54Z
>> peterj $
>>  5 >>> 9.2-RELEASE
>>  6 #   from: @(#)gettytab  5.14 (Berkeley) 3/27/91
>>  7 #
>> 
>> This is fairly ridiculous. Then, at a random file it has proposed its
>> own merge and asked the prompt "Does this look reasonable?". It was
>> not, but when I answered it with "n" it stopped the whole process
>> (instead of maybe opening the file in the editor for me to merge
>> again).
>> 
>> I've since retried the process and it behaves the same, and then tried
>> it on another system and again - the same type of manual merges and
>> the same exit from the process when answering "n" to a botched merge.
>> 
>> In both cases, I'm upgrading from either 9.0-RELEASE or 9.1-RELEASE to
>> 9.2-RELEASE and the command line was "freebsd-update upgrade -r
>> 9.2-RELEASE".
>> 
>> Am I doing something wrong, or is freebsd-update simply quirky and not
>> that useful?
>> ___
>> freebsd-hackers@freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
>> To unsubscribe, send any mail to "freebsd-hackers-unsubscr...@freebsd.org"
>> 
> 
> 
> 
> -- 
> Regards,
> Alexander Yerenkow
> ___
> freebsd-hackers@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
> To unsubscribe, send any mail to "freebsd-hackers-unsubscr...@freebsd.org"

___
freebsd-hackers@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
To unsubscribe, send any mail to "freebsd-hackers-unsubscr...@freebsd.org"


Re: Is freebsd-update braindamaged, or I'm using it wrong?

2013-10-01 Thread Alexander Yerenkow
To make better tool (than current behaviour of mergemaster regarding
configs/source files) which could make merge an easy task it *must* have
such things:

a) way to get Original configs/files from revision from you are upgrading
($Rev1)
b) way to get Original configs/files from revision to you are upgrading
($Rev100)
c) have ability to ignore differencies in comments
d) have ability to treat special cases (as $FreeBSD$ - just took newer line)

Then, your each new file will be  $Rev100 + diff_changes(CURRENT, $Rev1) +
diff_changes($Rev100, $Rev1).
Note, that in case that your diffs are none  diff_changes(CURRENT, $Rev1) =
0, then you can simply get new file.
Same thing in case that only $FreeBSD$ changed.

I have some PoC-es for this, but not in shell, maybe I'll come up someday
with full tool.




2013/10/1 Ivan Voras 

> This is the first time I've used freebsd-update in years, and I'm
> immediately flagging it as something I won't use in the future. For
> the last half hour it has been forcing me to manually resolve, one by
> one, in an editor, hundreds of "merge conflicts" such as these:
>
>   1 <<< current version
>   2 # $FreeBSD: release/9.0.0/etc/gettytab 209954 2010-07-12 19:09:18Z bcr
> $
>   3 ===
>   4 # $FreeBSD: release/9.2.0/etc/gettytab 243623 2012-11-27 19:23:54Z
> peterj $
>   5 >>> 9.2-RELEASE
>   6 #   from: @(#)gettytab  5.14 (Berkeley) 3/27/91
>   7 #
>
> This is fairly ridiculous. Then, at a random file it has proposed its
> own merge and asked the prompt "Does this look reasonable?". It was
> not, but when I answered it with "n" it stopped the whole process
> (instead of maybe opening the file in the editor for me to merge
> again).
>
> I've since retried the process and it behaves the same, and then tried
> it on another system and again - the same type of manual merges and
> the same exit from the process when answering "n" to a botched merge.
>
> In both cases, I'm upgrading from either 9.0-RELEASE or 9.1-RELEASE to
> 9.2-RELEASE and the command line was "freebsd-update upgrade -r
> 9.2-RELEASE".
>
> Am I doing something wrong, or is freebsd-update simply quirky and not
> that useful?
> ___
> freebsd-hackers@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
> To unsubscribe, send any mail to "freebsd-hackers-unsubscr...@freebsd.org"
>



-- 
Regards,
Alexander Yerenkow
___
freebsd-hackers@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
To unsubscribe, send any mail to "freebsd-hackers-unsubscr...@freebsd.org"


Re: Is freebsd-update braindamaged, or I'm using it wrong?

2013-10-01 Thread Vincent Hoffman
On 01/10/2013 10:16, Ivan Voras wrote:
> This is the first time I've used freebsd-update in years, and I'm
> immediately flagging it as something I won't use in the future. For
> the last half hour it has been forcing me to manually resolve, one by
> one, in an editor, hundreds of "merge conflicts" such as these:
>
>   1 <<< current version
>   2 # $FreeBSD: release/9.0.0/etc/gettytab 209954 2010-07-12 19:09:18Z bcr $
>   3 ===
>   4 # $FreeBSD: release/9.2.0/etc/gettytab 243623 2012-11-27 19:23:54Z peterj 
> $
>   5 >>> 9.2-RELEASE
>   6 #   from: @(#)gettytab  5.14 (Berkeley) 3/27/91
>   7 #
>
> This is fairly ridiculous. Then, at a random file it has proposed its
> own merge and asked the prompt "Does this look reasonable?". It was
> not, but when I answered it with "n" it stopped the whole process
> (instead of maybe opening the file in the editor for me to merge
> again).
>
> I've since retried the process and it behaves the same, and then tried
> it on another system and again - the same type of manual merges and
> the same exit from the process when answering "n" to a botched merge.
>
> In both cases, I'm upgrading from either 9.0-RELEASE or 9.1-RELEASE to
> 9.2-RELEASE and the command line was "freebsd-update upgrade -r
> 9.2-RELEASE".
>
> Am I doing something wrong, or is freebsd-update simply quirky and not
> that useful?
It is a little quirky, you arent the first to note this. For example
http://freebsd.1045724.n5.nabble.com/Should-I-be-able-to-use-mergemaster-with-freebsd-update-td5822457.html

"not that useful?" is subjective, some machines I've used it on happily,
I'm more likely to build and use mergemaster though.

Vince
> ___
> freebsd-hackers@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
> To unsubscribe, send any mail to "freebsd-hackers-unsubscr...@freebsd.org"

___
freebsd-hackers@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
To unsubscribe, send any mail to "freebsd-hackers-unsubscr...@freebsd.org"


Re: Is freebsd-update braindamaged, or I'm using it wrong?

2013-10-01 Thread Uffe Jakobsen



On 2013-10-01 11:16, Ivan Voras wrote:

This is the first time I've used freebsd-update in years, and I'm
immediately flagging it as something I won't use in the future. For
the last half hour it has been forcing me to manually resolve, one by
one, in an editor, hundreds of "merge conflicts" such as these:

   1 <<< current version
   2 # $FreeBSD: release/9.0.0/etc/gettytab 209954 2010-07-12 19:09:18Z bcr $
   3 ===
   4 # $FreeBSD: release/9.2.0/etc/gettytab 243623 2012-11-27 19:23:54Z peterj $
   5 >>> 9.2-RELEASE
   6 #   from: @(#)gettytab  5.14 (Berkeley) 3/27/91
   7 #



I've had the same expierience during some upgrades.
I've never been able to find the exact reason.
But my theory is that the mergemaster reacts on some specific settings 
on the affected systems - it be env or config files.


Unfortunately I've reinstalled the affected system (in rage) so further 
bug hunting is not possible. On newer installs the problem has not shown 
up...


/Uffe



___
freebsd-hackers@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
To unsubscribe, send any mail to "freebsd-hackers-unsubscr...@freebsd.org"


Re: Is freebsd-update braindamaged, or I'm using it wrong?

2013-10-01 Thread Florent Peterschmitt
Le 01/10/2013 11:16, Ivan Voras a écrit :
> This is fairly ridiculous. Then, at a random file it has proposed its
> own merge and asked the prompt "Does this look reasonable?". It was
> not, but when I answered it with "n" it stopped the whole process
> (instead of maybe opening the file in the editor for me to merge
> again).

You can merge files from freebsd-update, using (m) if I remember well.
Then you choose right or left for each step.

But there is something strange since you have hundreds of conflicts… Did
you built each release from sources?

-- 
Florent Peterschmitt   | Please:
flor...@peterschmitt.fr|  * Avoid HTML/RTF in E-mail.
+33 (0)6 64 33 97 92   |  * Send PDF for documents.
http://florent.peterschmitt.fr |  * Trim your quotations. Really.
Proudly powered by Open Source | Thank you :)



signature.asc
Description: OpenPGP digital signature