7 ms·

> > file history > You can have that in Linux You can use lvm or zfs snapshots, but that's not what I'm talking about – I'm talking about in-app browsable hist

by patrec 8y ago

> > file history > You can have that in Linux You can use lvm or zfs snapshots, but that's not what I'm talking about – I'm talking about in-app browsable history of things like documents or presentations. > Krunner, for example, has all the same features as Spotlight Last I checked it used Baloo to do the actual indexing, the list of high priority features/bugs on the project site https://community.kde.org/Baloo https://community.kde.org/Baloo (Baloo crashes a lot in various places etc.) and a quick google make it look like it remains alpha software at best, I'm also pretty sure it doesn't have an equally reliable index update mechanism. The most important thing about spotlight for me is that it can search file names and content (filtered by type if necessary) fast and reliably and up-to-date. But you can also do types of searches that as far I'm aware of none of the linux utilities can do. E.g. show me all the items I downloaded from a google.com domain: mdfind "kMDItemWhereFroms == '*google.com*'" If I want to see the filetypes that where downloaded sorted by frequency I can add mdfind -0 "kMDItemWhereFroms == '*google.com*'" | xargs -0 -n1 mdls -name "kMDItemWhereFroms" | sort | uniq -c | sort -n > For example I can run math calculations in it - which I haven't yet worked out how to do in Spotlight. You literally just type what you want calculated, e.g. `sin(pi/4)`. > As a counterargument I could list a dozen things that are in Linux that aren't in OSX. Like up to date core utils and proper package management. nix. By my lights the only proper package management for any OS. Works fine under both linux and macOS (and will also trivially supply you with up to date coreutils). > But do you actually use desktop Linux on modern hardware? I have been using (well-specced) linux desktops for most of my work for a long time. > The problem there is the application. However Linux will just kill the last process that over allocates memory. I don't think that's how it works. The whole point of having a proper OS (rather than say DOS) is that misbehaving app won't just bring down everything else. Also if you have a process that wildly allocates memory, by default linux will start off swapping like mad, making your computer effectively unusable (and yeah, in fact my linux desktop does have an SSD and many times as much RAM as my macbook, so if I'm not comparing like to like my linux station is the one with much more powerful hardware). And even if it runs out of swap it doesn't just kill the last process, it uses a more complex scoring algorithm which has a good chance of killing something you didn't want to be killed.

laumars 8y ago
> You can use lvm or zfs snapshots, but that's not what I'm talking about – I'm talking about in-app browsable history of things like documents or presentations. I got that. It's still just some application UI wrapped around a CoW file system. Maybe a better way of saying your point is "doing the same on Linux lacks a lot of polish" - which is true. But that's when happens when Linux has to support a multitude of file systems but Apple can control every aspect of their ecosystem. > Last I checked it used Baloo to do the actual indexing, Possibly? Krunner has always "just worked" for me so I've never bothered to look under it's hood. Regarding the bug you found, well I'd argue that you should expect to read bugs on a bug tracker given that's the point of bug trackers. It does feel like what you're basically now doing is the equivalent of reading a 1 star review of a product (eg on Amazon) and claiming it doesn't work by the proxy of others while also ignoring all the 5 star reviews from people who haven't had any issues. It's a heavily biased way to hold a discussion and if we're both honest, Mac's haven't been without their fair share of bad publicity either. So is it really worth our time cherry picking all the negative things when you and I both know that they're the exception rather than the norm? > You literally just type what you want calculated, e.g. `sin(pi/4)`. Handy to know. I suspected it would have been possible but I kept prefixing the formula with `=` which Spotlight didn't like. > nix. By my lights the only proper package management for any OS. Works fine under both linux and macOS (and will also trivially supply you with up to date coreutils). My point is you shouldn't have to install a 3rd party package manager. That's the bare minimum a modern OS should provide out of the box. > I have been using (well-specced) linux desktops for most of my work for a long time. I struggle to believe that given the descriptions of faults that you've been discussing. Though you have also said you've tinkered with the "swappiness" parameters (plus more) so I guess it's possible that you are running current hardware but have inadvertently tweaked Linux into performing terribly? Or maybe you're just exaggerating all these problems to make a point (much like your "look, I've found a bug on a bug tracker" comment above). Either way, if the problems were as prevalent and severe as you keep describing then you and I - and millions of other techies for that matter - wouldn't be running Linux. > The whole point of having a proper OS (rather than say DOS) is that misbehaving app won't just bring down everything else. "Proper OS" is such a flakey term and what you described isn't even the "whole point" of running an OS. But that's a whole other tangent. More importantly Linux doesn't do what you're accusing it of doing. Thus your statement is simply untrue in a multitude of ways. > Also if you have a process that wildly allocates memory, by default linux will start off swapping like mad It's actually a great deal more complicated than that. It depends on the size of your swap file, what applications you have open and their current running state (ie can they be paged). It depends on whether your cache is non-zero and it also depends on the kernel parameters you define. > And even if it runs out of swap it doesn't just kill the last process, it uses a more complex scoring algorithm which has a good chance of killing something you didn't want to be killed. Depends on the version of Linux (the kernel) you're running. Older kernels will just kill the last requester. Newer kernels do have a scoring algorithm but it's really not that complex at all (if memory serves, it's ostensibly a just a percentage*10 figure of used memory)

by patrec 8y ago

laumars 8y ago

> You can use lvm or zfs snapshots, but that's not what I'm talking about – I'm talking about in-app browsable history of things like documents or presentations. I got that. It's still just some application UI wrapped around a CoW file system. Maybe a better way of saying your point is "doing the same on Linux lacks a lot of polish" - which is true. But that's when happens when Linux has to support a multitude of file systems but Apple can control every aspect of their ecosystem. > Last I checked it used Baloo to do the actual indexing, Possibly? Krunner has always "just worked" for me so I've never bothered to look under it's hood. Regarding the bug you found, well I'd argue that you should expect to read bugs on a bug tracker given that's the point of bug trackers. It does feel like what you're basically now doing is the equivalent of reading a 1 star review of a product (eg on Amazon) and claiming it doesn't work by the proxy of others while also ignoring all the 5 star reviews from people who haven't had any issues. It's a heavily biased way to hold a discussion and if we're both honest, Mac's haven't been without their fair share of bad publicity either. So is it really worth our time cherry picking all the negative things when you and I both know that they're the exception rather than the norm? > You literally just type what you want calculated, e.g. `sin(pi/4)`. Handy to know. I suspected it would have been possible but I kept prefixing the formula with `=` which Spotlight didn't like. > nix. By my lights the only proper package management for any OS. Works fine under both linux and macOS (and will also trivially supply you with up to date coreutils). My point is you shouldn't have to install a 3rd party package manager. That's the bare minimum a modern OS should provide out of the box. > I have been using (well-specced) linux desktops for most of my work for a long time. I struggle to believe that given the descriptions of faults that you've been discussing. Though you have also said you've tinkered with the "swappiness" parameters (plus more) so I guess it's possible that you are running current hardware but have inadvertently tweaked Linux into performing terribly? Or maybe you're just exaggerating all these problems to make a point (much like your "look, I've found a bug on a bug tracker" comment above). Either way, if the problems were as prevalent and severe as you keep describing then you and I - and millions of other techies for that matter - wouldn't be running Linux. > The whole point of having a proper OS (rather than say DOS) is that misbehaving app won't just bring down everything else. "Proper OS" is such a flakey term and what you described isn't even the "whole point" of running an OS. But that's a whole other tangent. More importantly Linux doesn't do what you're accusing it of doing. Thus your statement is simply untrue in a multitude of ways. > Also if you have a process that wildly allocates memory, by default linux will start off swapping like mad It's actually a great deal more complicated than that. It depends on the size of your swap file, what applications you have open and their current running state (ie can they be paged). It depends on whether your cache is non-zero and it also depends on the kernel parameters you define. > And even if it runs out of swap it doesn't just kill the last process, it uses a more complex scoring algorithm which has a good chance of killing something you didn't want to be killed. Depends on the version of Linux (the kernel) you're running. Older kernels will just kill the last requester. Newer kernels do have a scoring algorithm but it's really not that complex at all (if memory serves, it's ostensibly a just a percentage*10 figure of used memory)