[Aptitude-devel] Bug#710689: Bug#710689: aptitude: use unicode character in the trees

[Daniel Hartwig]

On 2 June 2013 10:42, Christoph Anton Mitterer <calestyo at scientia.net> wrote:
> On Sun, 2013-06-02 at 09:25 +0800, Daniel Hartwig wrote:
>> Do you have any complaint with the current drawing of tree nodes,
>> other than “its not the precise unicode graphing characters”?
> Well not "compliant"... it's just that time moved on,... unicode is
> standard now unless for the most bare minimum situations where you
> perhaps end up with C locale

> (probably not situations where aptitude is
> used?)

Systems where LANG or LC was set incorrectly, or where package locale
is missing, broken, or under configured, or with fonts deliberately
lacking unicode ranges to save on space.  In such cases the fallback
is C locale and it must be supported.

>
> Many "standard" tools, e.g. tree move on and at least
> provide/auto-detect unicode enabled output in addition to ASCII
> fallbacks.

If tree maintainers want to support that then that is up to them.  I
do not want to support using different glyphs depending on locale,
even if it is only two options depending on ascii or unicode.  I have
already spent far too much time fixing problems caused by this when
e.g. translations do not render properly, or did previously and do no
longer because of changes elsewhere.


>
>
> Anyway... I don't quite understand why this should be new to aptitude,
> as it already seems to use these box drawing characters (which are
> surely not ASCII), take e.g.
> ftp://ftp.debian.┌───────────────────────────────────────────┐iff [Downloaded]
> ftp://ftp.debian.│Downloaded 6906 kB in 1min 38s (70,1 kB/s).│iff [Downloaded]
> ftp://ftp.debian.│     [ Continue ]          [ Cancel ]      │ownloaded]
> ftp://ftp.debian.└───────────────────────────────────────────┘ownloaded]
>
> or any "scroll bars" chich use: ▒

A timely reminder to eliminate such faults.  There are outstanding bug
reports about e.g. misrenderings on non-utf8 locales due to precisely
use of those glyphs.

>
>
> WRT, GNU Coding... well... they're quite old fashioned in some ways ;)
>

These are conservative for precisely the goal of being portable across
systems and languages.


I am not really interested in tweaking glyphs where such already work
fine.  Now returning to finish up the wheezy branch so stable systems
can have less quirky multiarch support.

Regards