Use of xz instead of gzip

Chris Rees utisoft at gmail.com
Thu Aug 11 21:12:47 UTC 2011


On 10 August 2011 16:03, Julian H. Stacey <jhs at berklix.com> wrote:
> Hi,
> Reference:
>> From:         Chris Rees <utisoft at gmail.com>
>> Date:         Wed, 10 Aug 2011 14:17:01 +0100
>> Message-id:   <CADLo839=-Wc6cohQ24otOjFN-_jw9CfAjfJt5eJGeS85RU=V0A at mail.gmail.com>
>
> Chris Rees wrote:
>> --20cf307f3aaee51c1704aa267e44
>> Content-Type: text/plain; charset=ISO-8859-1
>>
>> On 10 Aug 2011 13:53, "Julian H. Stacey" <jhs at berklix.com> wrote:
>> >
>> > > I can see this presenting a problem on 7.x (and earlier, but they are
>> > > no longer supported) because the base system shouldn't have dependencies
>> > > on ports.  The options would seem to be:
>> > > 1) Stick with gzip for 7-src and earlier (as well as ports until 7.x
>> > >    goes EOL)
>> >
>> > (Not that I'm suggesting for or against Option 1, but it needs modifying,
>> > Applying the same logic, append to Option 1: "as well as cvs-cur"
>> >        (as cvs can be used to extract eg 4 & ports & doc as well as head).
>> > As the ctm collection comprises:
>> >         cvs-cur ports-cur src-4 src-5 src-6 src-7 src-8 src-cur
>> > Option 1 thus becomes "Stick with gzip for all except src-8 & src-cur"
>> >
>> > Whether a CTM src-stable stream is EOL is irrelevant.
>> > All that matters is whether extractor hosts have xz or not.
>> >  One does not need to run ctm on same `uname -r` as one extracts for,
>> > eg my CTM extractor runs on 8.3, but I extract for all of
>> >        cvs-cur ports-cur src-4 src-5 src-6 src-7 src-8 src-cur
>> > Some (4, maybe 6) have long been EOL, but I use them for reasons beyond
>> > remit of this thread.
>> >
>> >
>> > > 2) Make ctm a port (so it can have a dependency on xz)
>> >
>> > I cant imagine many CTM users saying "please kick CTM out of src/" ;-)
>> >
>> >
>> > > 3) If dispensation is granted,
>> >
>> > Dispensation for what ?
>> >
>> >
>> > >    ensure the dependency is documented in
>> > >    UPDATING and ctm.1 when it is MFC'd to 7.x and before.
>> >
>> > Man ctm says: "The ctm command appeared in FreeBSD 2.1"
>> > In theory someone could be extracting on a really old host ?
>> > We might be lucky & find all people are extracting on 8.x ?
>> > We won't know unless we ask,
>> > Stephen could request on ctm-announce@ eg:
>> >  "People extracting CTM on old systems are requested to run
>> >        uname -r
>> >  & if not 8.something-RELEASE or current, report to .... at ....
>> >
>> > Personaly, I'd think an acceptable middle course could be to switch to
>> > whatever reduces the CTM bandwidth the most & put a note in man ctm:
>> >        "If running FreeBSD-x or earlier you need to
>> >         cd /usr/ports/archivers/xz ; make install"
>> >
>>
>> How does a manpage entry help? It'll only appear on later versions, where
>> It's irrelevant, and not appear on earlier versions, where it would help!
>
> I see what you mean, but often I read a manual on 1 host while
> running a prog on another. Notleast cos newer manuals sometimes
> list more limitations, eg this. Adding a note to manual would be useful,
>
> As above would Stephen's run time err. message.
>
>> Perhaps a note could be put into the Handbook entry?
>
> Sure, good idea. though not all have handbook, not all on line,
> (though personaly I mirror handbook to laptop for ref. before traveling).
>
>> Content-Type: text/html; charset=ISO-8859-1
>> Content-Transfer-Encoding: quoted-printable
>
> Duplicate ;-).
>
> Cheers,
> Julian

Suitably chastised, I'm sorry.

Blame the gmail Android app :(

Chris

-- 
Chris Rees          | FreeBSD Developer
crees at FreeBSD.org   | http://people.freebsd.org/~crees


More information about the ctm-users mailing list