[wdiff-bugs] Re: wdiff 0.6.0 released

[Martin von Gagern]

On 30.03.2010 13:19, Santiago Vila wrote:
> BTW: I wonder if you could take a look at this:
> 
> http://patch-tracker.debian.org/package/wdiff/0.5-21/
> 
> My task now is to determine which ones among those patches are already
> applied in 0.6.0 and which ones are not.
> 
> The ones which are not will be applied to the Debian package and
> resubmitted, but if you take a look at them and consider them as
> feature requests for 0.6.1, 0.6.2 or 0.7, it would help.

I thought I had looked at the patches applied to some debian wdiff
package in the past, but iirc that list was shorter, and definitely less
colorful. Thanks for the link!

I'll comment on them below, distinguishing those I consider done (+)
from those you might still want to apply (-) and those requiring
additional work (*). The (-) cases at least I'll consider for 0.6.1.

+ http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/01
The fclose(input_file) mentioned in the patch has been dropped in GNU
wdiff. So I assume this patch shouldn't be needed any more.

+ http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/02
Applied.

- http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/03
Still an issue. The strstr declaration is now enclosed in #ifndef
strstr, but that doesn't prevent redeclaration. Will look for a proper
solution to this using autotools and perhaps gnulib.

- http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/04
Unfortunately the Bug-Debian link given here seems unavailable, so I
have to guess about the motivation. Is this patch aiming for the case
where users have a pager which is a symlink to less, but doesn't contain
less in its file name? Considering it for inclusion, although I hate the
concept of PATH_MAX.

+ http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/05
The order of options is somewhat changed, but those three addressed by
the patch are all included these days.

- http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/06
I'll need a closer look to decide whether I want to replace popen with
writepipe, so I don't overlook possible implications. It would be easy
to make PAGER point to a short wrapper script providing extra options.
That solution would be perfectly portable, so I'd suggest it to users in
general. Nevertheless, if I'm convinced I won't loose anything, I'll
consider the patch.

* http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/07
* http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/10
* http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/12
wdiff has -v for --version. I believe it uses the lower case version
consistently at least throughout the wdiff application itself. I might
consider changing it to -V for consistency with other apps, and adding
-v as a backwards-compatible option. But that will require more work,
additional translations in particular. This will have to wait for 0.7.0
if at all. In general I believe --version to be good enough. I
personally seldom use that command line switch at all, on any app.

+ http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/08
Applied.

+ http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/09
Current wdiff looks somewhat different, but the core improvement of
using mkstemp over tmpnam has been implemented.

* http://patch-tracker.debian.org/patch/series/view/wdiff/0.5-21/11
In current wdiff, the category is named as "Individual utilities"
instead of "Utilities". Do you know whether this makes a difference in
the generated index in any setup? Else I'll have to experiment.

Greetings,
 Martin

signature.asc
Description: OpenPGP digital signature