[tex-live] tlmgr --help

[Victor Ivrii]

On Sun, Sep 21, 2008 at 9:36 AM, Manuel Pégourié-Gonnard <mpg at elzevir.fr> wrote:
> Victor Ivrii a écrit :
>> (and strangly nowhere mentioned)
>>
>> %texdoc  tlmgr
>So the pdf you
> see is generated from tlmgr.1 and I'd bet this one is generated from
> tlmgr --help (or both are generate form the same source in pod format).

Yes, it looks this way

>
>> "tlmgr update --all
>>
>> Update all packages."
>>
>> missing: and installs new.
>>
> ... in already installed collections.

So, if collection is new it is not installed? Say, mathabx was
installed by tlmg update --all because it was part of fontsextra (sp?)
collection but intro-scientific was not because it was not part

Still: it should be

tlmgr update --all

Update all collections (including new packages which are parts of
these collections)

So, my suggestion to check new packages
%tlmgr list | egrep -v "^i|binary|00|win32|wintools|darwin|scheme-"

makes sense (without egrep -v list is too huge: there are 1264 alien binaries!)

>
>> 3) Also --dry-run is more verbose than --list it is still less verbose
>> than an actual update --all
>> Also only --list returns version numbers (so some information about
>> forthcoming update could be provided only by it)
>>
> Actually, I don't feel comfortable either with --dry-run output being
> different from the actual output. That's just not the way I'd expect it
> to work. tlmgr gurus, any comment on this?

Not a guru, but definitely it should mock-remove before mock-install

>
>> b) if package is no more part of TeXLive tlmgr does not remove
>
> Still wondering if this is actually "smart" btw, since it implies there
> will be a difference between a fresh install and a old installation kept
> up-to-date with tlmgr. Anyway, if you like it...

Well, the alternative of disappearing packages is worse. Ideally
(IMHO) it should ask what should be done with this package: left,
removed or "removed and reinstalled into local TDS but I know that you
do not want to touch it (and have a good reason)


>
>> Actually tlmgr is smarter than documentation (which is also a part of
>> tlmgr.pl); so in a sense tlmgr is smarter than it tells you :-)
>>
> Isn't it good news? ;-)

Not sure, but an opposite is much worse and happens much more often with s/w

>
> Manuel.
>

Victor

-- 
========================
Victor Ivrii, Professor, Department of Mathematics, University of Toronto
http://www.math.toronto.edu/ivrii