FEATURES="candy"

[Paapaa]

100% useless feature, I wouldn't mind removing it. I say: aim for simplicity.

[Bob P]

candy has always been an undocumented feature. the only people who are even aware of it are the people who have actually examined the portage source code. it should come as no surprise, then, to observe that the vast majority of people responding to this poll aren't aware that the feature exists and aren't familiar with what it does. moreover, it should come as no surprise that people who have been living in blissful ignorance won't miss anything that they never knew existed. in this context it might be most meaningful to sort the responses to this poll, group them into pools of people who knew that the feature existed and those that didn't, and analyze the pools of responses separately.

on the one hand there's the "efficient code" argument. if the developer is basing his opinion on the efficient code argument, this thread is really quite meaningless, as the answer to the question was known before the question was asked.

from a user's CPU utilization perspective, "candy" doesn't cost any significant amount of CPU overhead for the people who aren't using it. candy amounts to a block of code that is executed if and only if the user triggers its execution by implementing the features statement. realistically speaking, candy only costs you in CPU utilization if you intentionally turn the feature on.

in the big scheme of things, i think that its important to remember that undocumented features like candy amount to pearls that have been buried in the source code to reward people who bother to read it. somebody who had worked on portage before spent alot of time working on that code. leaving it in for posterity is an honor to them in acknowledgement of what they've done for Gentoo, and it doesn't cost anything. OTOH, removing it is the honorary equivalent of a sharp stick.

my recommendation would be to give users choice, and to leave it in. but then my recommendation would also have been not to document a feature that had been intentionally undocumented in the first place.

[loki99]

candy has always been an undocumented feature. the only people who are even aware of it are the people who have actually examined the portage source code.

Not quite true!

I never read the portage sources and know about it since it has been implemented.

[Cinder6]

I can't even read it when it's enabled (I only know it says "Thank you for using Gentoo" by looking for it), but I think it's kinda fun to have Easter eggs like this, so I say keep it (unless, of course, as others have said, it makes the code unmanageable).

[ali3nx]

definitely a keeper in my opinion. I've been using candy in features for awile now and try to have it enabled whenever possible. Thanks carpaski!

[syouth]

Replace "candy" with "offensive" to show ASCII porn and things like that...

[loki99]

I can't even read it when it's enabled (I only know it says "Thank you for using Gentoo" by looking for it), but I think it's kinda fun to have Easter eggs like this, so I say keep it (unless, of course, as others have said, it makes the code unmanageable).

One of the lines I remember seeing one time said something like this:

If you can read this, your computer is too slow!

[at240]

One of the lines I remember seeing one time said something like this:

If you can read this, your computer is too slow!

Enjoyed reading this thread---had no idea that this feature even existed...

[Stolz]

I use it but if it makes code easier to mantain, get rid of it.

[Omega21]

I wanna try it now.... How do you activate it? (If its not too late. Ive seen some GNOME-like ideals in the last few months with Gentoo... "If its a feature, remove it")

Although if it makes the developer's lifes that much easier, who am I to stand in your way?

[jedsen]

candy has always been an undocumented feature. the only people who are even aware of it are the people who have actually examined the portage source code. it should come as no surprise, then, to observe that the vast majority of people responding to this poll aren't aware that the feature exists and aren't familiar with what it does. moreover, it should come as no surprise that people who have been living in blissful ignorance won't miss anything that they never knew existed. in this context it might be most meaningful to sort the responses to this poll, group them into pools of people who knew that the feature existed and those that didn't, and analyze the pools of responses separately.

on the one hand there's the "efficient code" argument. if the developer is basing his opinion on the efficient code argument, this thread is really quite meaningless, as the answer to the question was known before the question was asked.

from a user's CPU utilization perspective, "candy" doesn't cost any significant amount of CPU overhead for the people who aren't using it. candy amounts to a block of code that is executed if and only if the user triggers its execution by implementing the features statement. realistically speaking, candy only costs you in CPU utilization if you intentionally turn the feature on.

in the big scheme of things, i think that its important to remember that undocumented features like candy amount to pearls that have been buried in the source code to reward people who bother to read it. somebody who had worked on portage before spent alot of time working on that code. leaving it in for posterity is an honor to them in acknowledgement of what they've done for Gentoo, and it doesn't cost anything. OTOH, removing it is the honorary equivalent of a sharp stick.

my recommendation would be to give users choice, and to leave it in. but then my recommendation would also have been not to document a feature that had been intentionally undocumented in the first place.

Again, Bob P, thank you for that steaming load of useless crap.

[goom]

Imho it is completely unneccesary. Got it running on a server here, but that was only some testing and nothing serial. As you said, a silly feature

[John R. Graham]

I found both of those undocumented features by studying Portage's Python source code. I have enjoyed each of these little discoveries over the years. Sometimes even the comments are hilarious. Once while working late into the night to find the root cause of a customer-reported anomaly, I found a quote that one of my employees had left behind years before for someone to find:

Code: Select all

/*
**  "Wanting every habitable planet to be inhabited is like wanting
**   everybody to get athlete's foot."
**  -- Kurt Vonnegut
*/

I burst out laughing, thinking that it must've been a similarly late night that spawned this little missive. It still makes me smile, just thinking about it.

So, by all means, leave them in. I've seen the code: there're no credible performance or maintainability argument to be made here. And they're fun!

- John

[brims]

So, by all means, leave them in. I've seen the code: there're no credible performance or maintainability argument to be made here. And they're fun!

- John

Things like that can be amusing like in one kernel when there was a printer problem it would say "lp0 on fire" or something. But the original poster, who is a dev, states that the code is fairly long. Though antarus has few posts here on the forums, he must know how to code and have some idea of what he is doing otherwise the other developers wouldn't let him be a developer.

If the code is long and no one uses it, well then there is no need to keep it unless a better solution could be devised. One poster suggested modularizing it which would work. But then you get back to the topic of, is anyone even using it and is it worth the time to actually modularize it.

To me it looks like everyone that's bothering to post on this thread recommends getting rid of the feature, and those who don't post either don't care or see that their point of view is well covered so they don't post.

[sponge]

But the original poster, who is a dev, states that the code is fairly long.

Err... how long? Searching /usr/bin/emerge only finds one instance of candy, at line 75, where it's checked for in the feature list. Under that conditional, there's an assignment that points to a list at line 24. Then, there's an assignment that points to a function at line 54, which is 8 lines long. The conditional is 3 lines long, and while the list is 18 lines long, it's hardly difficult code. All in all, 29 lines. That doesn't seem like much, even in python.

[andrewd18]

Didn't knew it existed, but now I do I'm not really interested either.
Kick it.

++

[Ferdinando]

I think it should be kept, maybe in a separate module only installed with a particular use flag. I've been using it for a long time now, I just like it even though I can't read the sentence (I went in the source code to read them - the spinner_msgs var, at the beginning); it's not funny anymore, but it's original.

Bye

[brims]

Err... how long? Searching /usr/bin/emerge only finds one instance of candy, at line 75, where it's checked for in the feature list. Under that conditional, there's an assignment that points to a list at line 24. Then, there's an assignment that points to a function at line 54, which is 8 lines long. The conditional is 3 lines long, and while the list is 18 lines long, it's hardly difficult code. All in all, 29 lines. That doesn't seem like much, even in python.

Ok, then maybe antarus can clarify that point. But again, if it isn't used, is it really needed?

[no idea]

Again, Bob P, thank you for that steaming load of useless crap.

Hey, he is 100% right.

It is an easter egg, a reward for those who actually read the emerge code.

And as he said, it should stay in honor of those people who really did something for gentoo.

What do you gain by throwing it out?

[no idea]

Ok, then maybe antarus can clarify that point. But again, if it isn't used, is it really needed?

I don't use emerge --deep, so let's throw that out as well, shall we?

[brims]

Ok, then maybe antarus can clarify that point. But again, if it isn't used, is it really needed?

I don't use emerge --deep, so let's throw that out as well, shall we?

Good job trolling. You don't use --deep, other people do. As the thread shows, very few use FEATURES="candy".

[Ferdinando]

As the thread shows, very few use FEATURES="candy".

I don't think so; yes, there were some who didn't even know of that, but believe me, I've read people on these forums who didn't even know of cflags. It wasn't documented in the installing manual nor in any man page I know of, but it was mentioned in a GWN, so a lot of people read about it, and I think that a lot of people likes some candy here and there: ok, it's not very pretty, but I know many people who use it.

Bye

[brims]

As the thread shows, very few use FEATURES="candy".

I don't think so; yes, there were some who didn't even know of that, but believe me, I've read people on these forums who didn't even know of cflags. It wasn't documented in the installing manual nor in any man page I know of, but it was mentioned in a GWN, so a lot of people read about it, and I think that a lot of people likes some candy here and there: ok, it's not very pretty, but I know many people who use it.

Bye

Ok, then you explain all the posts "interesting, but don't use it" or "not very useful" or any post similar to these examples. A couple of posts say "never knew it existed", for the most part the tone is to get rid of it.

EDIT: As someone brought up, on a semi fast machine, you can't even read the sentences. On my 356MHz machine I'd be able to read it, but I don't use it. On my 2 desktops, I'd never be able to read it, the normal spinner gets spinning so fast it looks like a propeller spinning, can't see the different characters.

EDIT: Also if you know so many people that use the feature, why aren't they on this thread telling not to axe it. If enough people say to axe it, well, it may get axed. If some enough people say "keep it" the feature will be left alone. But they clearly don't want it.

[Ferdinando]

Ok, then you explain all the posts "interesting, but don't use it" or "not very useful" or any post similar to these examples. A couple of posts say "never knew it existed", for the most part the tone is to get rid of it.

Wait: most posts were to get rid of it if it was a problem in terms of performance and/or code simplicity, and I agree with that. But you wrote that very few use it, and I disagree: if you search for "features candy" you'll get 5 pages of topics.

as someone brought up, on a semi fast machine, you can't even read the sentences

True; and that happens to me too, as I wrote a few posts above. But I still like the idea.
I wouldn't mind if it were to change in a better spinner, cute, coloured and funny, but I feel that losing this feature completely would be sad.
I've chosen this distribution a long time ago because it gave me the opportunity to find my own compromise between performance and beauty, and I still want to be free to decide.

My 2 ecu cents.
Bye

[amne]

Again, Bob P, thank you for that steaming load of useless crap.

Feel free to disagree with your fellow community members, but please do it in a more professional way next time.