@CarlLange. That is good news and thank you very much for making the effort. I'm still on OS X 10.13.4, so I am using Safari 11.1, but I expect to upgrade to 10.13.5 shortly, which I presume will install Safari 12, and I will need the new extension because I use halirutan's button quite a lot.

@m_goldberg I already explained in detail here that I would like to ensure that you specifically have access to the script.

@m_goldberg I think you should be ok upgrading to 10.13.5/10.13.6, but the upgrade to 10.14 will be the problematic one.

@CarlLange. My coming upgrade will be to 10.13.6. I see that I have procra

:46609814, thanks Henrik, something like ToSystemMatrix is probably what I'll go for.

Is there anyone here using some Linux that's old enough for the default gcc to be 4.8 or earlier? Please let me know.

@m_goldberg sorry, I'm not aware of any way to use @halirutan 's javascript without Ninjakit in Safari, but I'm not a good person to ask

@ChrisK @m_goldberg The critical point is that although some of our people can write javascript, I wouldn't know anyone who is a real Apple developer who writes Safari extensions. So the first hurdle is to convert the current editor buttons script into a native Safari App Extension. To publish it as we do in the Chrome Web Store, you need to be an Apple Developer which

costs $99/year

@Szabolcs I have a VM with gcc (GCC) 4.4.7 20120313, what do you need?

@JasonB. Do you have Mathematica in it?

@JasonB. In short, I am trying to understand what determines binary compatibility with Linux. I got some complaints about IGraph/M not working on very old systems. I am trying to figure out what I can do (and what is reasonable to do) to maximize compatibility.

@JasonB. If you could just try to run some binary executables I send you that would already be a very useful test. If you could try to use a LibraryLink library, that would be even better. I have a few test executables prepared based on my guesses about what is important for compatibility.

@halirutan Yeah, the real annoyance is the money/developer account - I can manage the actual creation of the app extension fine.

@Szabolcs I happen to have many Mathematica's on that machine, do you need me to install IGraphM the usual way, or something more? You can email me if it's complicated

@JasonB. I'll email you!

I was reading this and it seems that Mathematica has many of the same quirks that python does in terms of deciding what's numeric and not.

@JasonB. Could you check your libc version please? Apparently one can find that out by directly running the libc.so.6 binary. The location might differ between systems. On mine it's /lib/x86_64-linux-gnu/libc.so.6

I didn't know one could run .so files

Is it 2.14 or later? Or earlier than that?

"GNU C Library stable release version 2.12, by Roland McGrath et al."

@Szabolcs - did not tag you with the above reply.

Sometimes I wonder about people who accept an answer, but apparently don't give a positive vote for it.

I mean, I just wonder the logic behind it. That vote is free. It doesn't cost them any points.

My perspective on such an approach is that "this answer solved my problem but has no value."

Then again, world is full of different perspectives, many of which I probably don't comprehend at all the way their holders do... :I

@kirma Remember that low-rep users cannot upvote...

@HenrikSchumacher Really? Even on their own questions?

I think that's unfortunate.

@kirma I am pretty sure about that. This was my initial motivation for gaining some rep...

@HenrikSchumacher Ok, good to know!

Sure, there can be logic behind that, but... hmm. All effects of it are not the best.

@kirma Well, it's only 15 points you need for the priviledge (see stackoverflow.com/help/privileges). But for people who visit the site only once...

That would be 3 upvotes for the first question ever posted. And be honest. Most questions of newcomers become quickly classified as "mistake" and only few get three or more upvotes.

Hum. :/

Another reason to upvote newcomers, even if their question is not among the most orginal ones...

Hmm hmm. I always thought one could upvote answers of their own question, but then again, that would be an attack surface for multiple accounts trying to play the system.

Which results trouble on the other side...

Has anyone run into a notebook where any block of cell they try to cut and paste hangs the FE then crashes?

Actually also copy and paste does it

I think it might be related to the fact that if I copy literally any textual data the FE inserts these obnoxious continuation characters and a newline after every character. So if there're any cells that use text-form data it hangs and crashes as it does that and runs out of available memory.

I wonder if it'd be possible to use something like this to interoperate Julia and Mathematica. I've never used Julia, but I often hear the zealots rave about how wonderful it is.