How to avoid upgrading an installed Python 2.7 package?

[Leonardo.b]

"Good morning," said the little prince.

"Good morning," said the merchant.

This was a merchant who sold pills that had been invented to quench thirst. You need only swallow one pill a week, and you would feel no need of anything to drink.

"Why are you selling those?" asked the little prince.

"Because they save a tremendous amount of time," said the merchant. "Computations have been made by experts. With these pills, you save fifty-three minutes in every week."

"And what do I do with those fifty-three minutes?"

"Anything you like..."

"As for me," said the little prince to himself, "if I had fifty-three minutes to spend as I liked, I should walk at my leisure toward a spring of fresh water."

[Hu]

OP: you have made very clear your position that you want Portage to do more to help you follow the upgrade path you chose. I hope that we have made clear that you are a severe outlier, that your problems are self-inflicted by insisting to retain without any modification packages that can no longer fully work without modification, and that there seems to be little community interest in developing a patch to Portage to improve its ability to handle your chosen upgrade style. That is not to say that, if someone were to submit such an improvement, it could not be accepted. However, I see no indication that anyone active in this thread is interested in working on that improvement for you. Therefore, you need either to improve your messaging to better convince someone to make the proposed improvements, or you need to demonstrate to us that it is feasible by submitting a prototype implementation. As I wrote above, I would be pleased to see Portage better able to report all the conflicts, because I think some forum threads would be more readily solved if Portage could provide an exhaustive list of all the things the user will need to fix before a full upgrade is possible. However, this would not improve my uses enough that I am willing to develop the required proficiency to write this. In part, this is based on the belief that if it were easy, the Portage developers would already have provided it as an option. Therefore, I am guessing, with no specific evidence, that this is not easy. The code for it might be very complicated or very slow. I would be pleased to see that assumption proven wrong.

[pjp]

Don't be fooled by the "update often" mania. Whether you use your time productively for 1 year and 250 days (as I chose to do), or spend 15 minutes every day on your daily update hassle, you are going to spend more or less the same order of magnitude in time.

I disagree with your opinion about time consumption. At a minimum, 15 minutes per day (I don't work on updates every day) is less disruptive than 13+ days and counting of not working at all on things I value more than performing updates. I'd take a net loss of time at 15 minutes per day if it meant rarely having multi-hour or multi-day efforts to perform an update.

As for the "mania," I'd say it is more manic to spend a concentration of time spanning weeks than i is minutes or hours. By common use of the term mania, "excessive" seems to fit an update that takes multiple weeks.

More to the point about "update often," Gentoo is a rolling release. Since my early use of Gentoo, I have wished there could be a way to "pin" versions of software to reduce the speed at which updates occur. A proposal had been put forth, but if I recall correctly, it was abandoned due to the human time that would have been necessary to provide such a feature. Disappointing, yes, but that doesn't change the nature of a rolling release.

I generally try to update once per month, or not more than 3 months between updates. More than that requires too much time in my experience, for the value I place on performing updates. Most of the time I spend on updates has more to do with understanding the impact of "important" changes and improving my process rather than the updates themselves.

As I mentioned, I'm glad your choice of how to perform updates works for you. It seems horrifying to me, which is why I offered my input on a way to avoid multi-week updates. But you've made it quite clear that only your process is one that you are willing to choose. That's perfectly fine, but choices have consequences, and it is you who decides whether or not those consequences are a good thing or a bad thing, or at least worth making that choice.

[segmentation-fault]

Using my world-updatable set (an idea inspired by Hu's proposal of a set technique to overcome the original problem of this thread), I have been able to make significant progress: more than 1000 packages of the ~5000 installed have already been updated (including their dependencies). This significantly eases things, as they were the most important, most "connected" components of the dependency graph. I don't expect any other serious "blockers" on the way. Now, the majority of computer's time is spent on compiling, so I am confident all will go well.

That I am a "severe outlier" is due to the self-amplifying dynamics of a system that favors small, simple-to-resolve dependency graphs, being increasingly helpless with increasing graph complexity. This encourages frequent updaters and punishes infrequent ones. At the end, only very few infrequent updaters remain, as the majority of them turns to alternatives. This does not mean that the current system is ripe, advanced, or in good shape - it only means that the system has "found" its users through a darwinian natural selection process.

Nevertheless, I am not discouraged and I will do think about ways to make the system better - and these "issues" less critical.

Thanks to all

S-F

[pjp]

Split off [topic=1147180][split] upgrade philosophies[/topic]

[segmentation-fault]

Thank you pjp for not diluting the thread...

Finally, on Day 20 of my update hell, after countless partial updates that paved the way (compiling 1000-1500 packages...manually...) and freed it from slot conflicts and other annoyances, I was able to type my "voodoo" (as I call it, looking at its options ) command

Code: Select all

emerge -vUDua --backtrack=2000 @world-updatable

and, after more than one hour of thinking, get the offer to upgrade:

Code: Select all

Total: 1306 packages (1067 upgrades, 5 downgrades, 88 new, 23 in new slots, 123 reinstalls, 2 uninstalls), Size of downloads: 6,329,150 KiB
Conflict: 7 blocks (all satisfied)

Now, the 1 million $ question for you is...shall I say Yes, or shall I deny?

...

Forget it, just kidding - of course I said Yes!

Yes, Yes, Yes!!!