Resolving issue 2431 as invalid? (blame ignore eol-style)

[Erik Huelsmann <eh...@gmail.com>]

In the past, I have tried to resolve issue 2431 (blame ignores
eol-style). However I wasn't completely successful yet.

Giving this issue much thought, I came to the conclusion the issue
should be resolved as INVALID and all associated commits should be
reverted: blame output is command line output.  This means it should
use the platform default EOL style, not what happens to be in the
eol-style property.

This differs from diff, since diff - as input for patch - should
mirror the actual file as much as possible.


Opinions?


bye,


Erik.

Re: Resolving issue 2431 as invalid? (blame ignore eol-style)

["Peter N. Lundblad" <pe...@famlundblad.se>]

Erik Huelsmann writes:
 > That's right. The current diff code - as I understand it - has a
 > helper to estimate the current file's eol style. The blame algorithm
 > generating the final output doesn't use this helper yet. It surely
 > could be updated to do so though.
 > 

No, the "detect file's eol style" is used for merge to use the "best" eol style
for conflict markers.  The diff code itself always looks for all three eol styles
we support.  Blame could do the same.

//Peter

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: Resolving issue 2431 as invalid? (blame ignore eol-style)

[Erik Huelsmann <eh...@gmail.com>]

On 4/22/06, Peter N. Lundblad <pe...@famlundblad.se> wrote:
> Erik Huelsmann writes:
>  > Giving this issue much thought, I came to the conclusion the issue
>  > should be resolved as INVALID and all associated commits should be
>  > reverted: blame output is command line output.  This means it should
>  > use the platform default EOL style, not what happens to be in the
>  > eol-style property.
>
> +1.  Mainly because this output isn't fed to a program that will use it as the
> raw contents.  OTOH, the blame code should look for all eol styles,
> which I'm not sure if it does currently.

That's right. The current diff code - as I understand it - has a
helper to estimate the current file's eol style. The blame algorithm
generating the final output doesn't use this helper yet. It surely
could be updated to do so though.

Since blame uses diff internally, I presume it does use the helper to
find the deltas.

bye,


Erik.

Re: Resolving issue 2431 as invalid? (blame ignore eol-style)

["Peter N. Lundblad" <pe...@famlundblad.se>]

Erik Huelsmann writes:
 > Giving this issue much thought, I came to the conclusion the issue
 > should be resolved as INVALID and all associated commits should be
 > reverted: blame output is command line output.  This means it should
 > use the platform default EOL style, not what happens to be in the
 > eol-style property.

+1.  Mainly because this output isn't fed to a program that will use it as the
raw contents.  OTOH, the blame code should look for all eol styles,
which I'm not sure if it does currently.

//Peter

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org