Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 28 Oct 2009 21:48:10 +0300
From: Solar Designer <>
Subject: Re: My patches (Performance, options, mingw, vc, generic md5, phpass, etc)


On Wed, Oct 28, 2009 at 12:55:02PM -0500, JFoug wrote:
> Eric provided me with his original file (the COMINOSEC_fmt.c) , and I
> compared that with my original. By original, I mean ver- with
> the jumbo-1 patch installed.
> The file Eric provided, and mine were 100% identical, except for EOL
> markers, and that seems to cause problems on the Mac port of patch.
> The file Eric sent, had DOS end of line (\r\n), while the file I had
> was standard unix (\n) line endings.

How did it happen that the files used different linefeed conventions?
Doesn't this imply that one of them was in fact not the original file?
If that one was yours, then consequently your patch was wrong/buggy -
and that's what should be addressed (now and on any future occasions).

> I have no 'solution' to this, other than we need to make sure that
> all patches, and files about to be patched are in \n format.

Correct.  I think the burden should be on those of us making the
patches, not on those using the patches.

> Are there ways to make patch smart enough to ignore end of
> line issues?
> I do see in patch the -l  (or --ignore-whitespace)  I am pretty sure
> this flag would do the job, and from the patch I have, it appears to
> be the only available option for ignoring end of line differences.
> Not sure if the -l is the 'right' option to use or not,

I think it would work, but I'd rather not complicate the instructions
unnecessarily.  (We'd have to explain that it's letter "ell" and not
digit "one".)

> or if it is better to work to keep all source/patch in unix EOL format.


> If the
> -l is a good option, then someone should (I could do it, I guess)
> change the wiki page on how to properly install a patch.
> I would not change the wiki, unless there was an 'official' ok for
> using that swich in the patch program.

I don't have strong objections to this, but it feels like a last resort
approach.  I think we should figure out where the discrepancy came from
first.  Maybe we'll end up adjusting the instructions for creating
patches instead.



Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.