| 1 |
On Wednesday, April 22, 2020 7:35 PM, Michael Orlitzky <mjo@g.o> wrote: |
| 2 |
|
| 3 |
> On 4/22/20 11:22 AM, Caveman Al Toraboran wrote: |
| 4 |
> |
| 5 |
> > On Tuesday, April 21, 2020 11:01 PM, Consus consus@××××.net wrote: |
| 6 |
> > |
| 7 |
> > > Yeah, mgorny likes to do some provocative stuff like forking Portage. |
| 8 |
> > |
| 9 |
> > patching P*****E is heretic, and forking it is |
| 10 |
> > outright blasphemous. |
| 11 |
> |
| 12 |
> For everyone complaining about how long emerge @world takes, and about |
| 13 |
> the incomprehensible error messages -- this fork was a step towards |
| 14 |
> fixing that. Portage does some slow, unpredictable, undocumented magic |
| 15 |
> when resolving dependencies that it never should have done in the first |
| 16 |
> place. Developers using portage then make commits that appear to work |
| 17 |
> with portage, but won't work in any other PMS-compliant package manager, |
| 18 |
> and often don't work in portage itself when given slightly different |
| 19 |
> command-line options. |
| 20 |
> |
| 21 |
> Portage was forked because the current maintainers insist on leaving it |
| 22 |
> broken to "avoid the phone calls." There are still problems, but this |
| 23 |
> way people don't realize they're portage's fault. |
| 24 |
|
| 25 |
i was joking. i agree with you + mgorny. |
| 26 |
|
| 27 |
in fact, i think portage sucks so much it must be |
| 28 |
rewritten from scratch, in such a way that it has |
| 29 |
least run-time dependencies, so we stop worrying |
| 30 |
about upgrading other packages, such as python. |
| 31 |
|
| 32 |
e.g. perhaps gne (gne is not emerge) should better |
| 33 |
be statically linked (no stupid python run-time |
| 34 |
that freaks us every time we upgrade python). |
| 35 |
|
| 36 |
just my thought. but mgorny knows much better |
| 37 |
than me most likely. i like his work. and i hope |
| 38 |
politics around emerge/portage gets dropped. |
Archives