Roland Mainz wrote:
> What about the idea to create a tool called "pathtomanpath" which takes
> care about the conversion in /etc/profile and may be used by plain users
> in their ~/.profile, too ?
>
Simply awful (IMHO). Of course, "awfulness" is for the bunch of us to
decide.
> 1. /usr/bin/more is _not_ CSI-enabled and barely works with multibyte
> locales (I am ignoring /usr/xpg4/bin/more since most end-users never
> used that unless some admin was really clever and/or (more likely) set
> PATH to include "/usr/xpg4/bin:/usr/bin"). The newer "less" versions are
> better than that.
> 2. In theory we could add a plain check for "zh_CN.GB18030" and set
> PAGER=/usr/xpg4/bin/more to make sure it works in this case... but we
> don't have any manual pages in "zh_CN.GB18030" which reduces the issue
> to "mailx".
>
> An alternative may be that a Sun manager steps up and declares that some
> resources get allocated to "fix" /usr/xpg4/bin/more (e.g. add cursor
> keys for document navigation like in "less", add seach result
> highlighting (=reverse video), a way to edit he regular expression
> lines, fix any remaining multibyte locale issues etc.) and move the
> resulting application to /usr/bin/more (e.g. EOL/EOF /usr/bin/more and
> replace it with /usr/xpg4/bin/more).
>
1) 1) works for some locales that less doesn't, correct? Why it works
is irrelevant.
(I note below, you assert that it really doesn't work, which is
an interesting
avenue to pursue.)
2) 2) is a gross hack. IMHO, the cure is far worse than the disease. It
also is a
clear violation of the relevant "big rule". Summarizing the
"big rule": There
are a set of about 10 locales which are all treated as "most
important".
The alternate I saw was for this project team to fix less so that it
offered no regressions
relative to more. If that fix is accomplished directly by the project
team or by
getting somebody else to do it is irrelevant. This is business as usual.
>> I see the smiley, but the important point is "no regression". I see
>> that your view is
>> that the "good" outweighs the "bad". We officially don't think that
>> way (and
>> in practical terms, acceptable regressions need to be tiny and I don't
>> think being
>> "in your face" to a couple of billion folk is tiny).
>>
>
> Right... but did you use /usr/bin/more in ja_JP.PCK recently ? I doubt
> "less" would be much worse for those users...
>
I guess declaring more broken anyway in ja*, zh* and whatever else is a
way to
avoid the regression concern. However, I think we need somebody with a good
understanding of these locales and the behavior within them, to declare
it so.
I don't know who that would be. Maybe somebody on this alias can suggest
someone (or better yet, a group).
- jek3
