Alexander E. Patrakov wrote these words on 12/27/05 07:33 CST:
> I suggest that a warning should be put on each affected package's page,
> with approximately the following text:
>
> =========
> This package doesn't work correctly in UTF-8 based locales. If you use a
> UTF-8 based locale, don't install this package. If you want to install
> this package, don't use UTF-8 locales.
> =========
That warning is simply too stiff, in my opinion. In fact, a warning
is not called for at all. A note is sufficient. Additionally, the
text needs some revising:
"This package doesn't work correctly in UTF-8 based locales".
This can never be placed on a BLFS page. Something needs to be
placed on the page that accurately describes the breakage, or
limited functionality you *may* see. "Doesn't work correctly"
is simply too vague.
"If you use a UTF-8 based locale, don't install this package".
This is also inappropriate for some of the packages, i.e., ed.
So a custom note will be required on some pages, instead of a
generic note.
"If you want to install this package, don't use UTF-8 locales"
This text is totally unnecessary. It is not our job to say "don't
use something". All we need to do is identify what will happen
if someone installs/uses (whichever it may be) the package on
a system set up in a UTF-8 locale. The choice is then up to the
reader to do what she wishes.
Some random thoughts about the individual package notes:
> Ed-0.2 (use only as a build dependency for teTeX)
Would/Does anyone use it for anything else? Unlikely. So, the
note on this page may be something to that affect.
> ASH-0.4.0 (The ${#VARIABLE} construction returns the number of bytes in
> $VARIABLE, while it should return the number of characters).
Is this the *only* thing broken? Doesn't really seem like a show
stopper that warrants a "warning" that says "Don't use this ..."
:-)
> Tcsh-6.14.00 (use only if it is a build dependency of something)
We will need some technical explanation of what is broken. Something
similar to ASH would be nice.
> ZSH-4.2.5 (upgrading to a development snapshot is reported to help)
Can this not be tested by someone? What is broken? Why should we
say "is reported to help"? Surely we can determine if something
does help, or does not help.
> UnZip-5.52,
Can't a note be placed on the page that says you *may* need to
do ("whatever it is") for UTF-8. Won't it work for what the
BLFS book uses it for? (mozilla, docbook, etc)
> Tcl-8.4.11, Tk-8.4.11: untested
Please, more information. :-)
> Links-2.1pre17: no way to properly configure for UTF-8 based character
> cell display, there is no UTF-8 terminal encoding in the menu. If run on
> the framebuffer console, doesn't properly accept keystrokes. If run from
> X as a graphical browser, seems to work.
Can't these things be put on the page?
> Nail-11.25: untested
See Tcl, Tk.
> [snip all others]
Overall, very good reports, Alex. However, I'm not sure we need to
dramatize this so much in the book. For packages that have some issues,
we identify the issues and let the readers decide what is best for
them.
--
Randy
rmlscsi: [GNU ld version 2.15.94.0.2 20041220] [gcc (GCC) 3.4.3]
[GNU C Library stable release version 2.3.4] [Linux 2.6.10 i686]
10:30:00 up 93 days, 19:54, 3 users, load average: 0.17, 0.30, 0.49
--
http://linuxfromscratch.org/mailman/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
