The cross between the awesome redbean.dev open source webserver in a zip executable that runs on six operating systems and feather.wiki for creating personal non-linear notebooks, databases, and wikis that is entirely self-contained and runs in your browser! Phew - it's smaller than a TiddlyWiki file.
Modern antivirus software is arguably a scam. To avoid having to determine if a program is actually malicious (which to do definitively would require solving the halting problem), they simply mark anything they haven't seen as suspicious. Granted, malware developers did use obtuse binary formats, packers, and strangely tweaked executable headers for years.
Plenty of demoscene productions are like that too. If anything, it's a good chance to educate others on whose interests "antimalware" really protects.
Around 2 decades ago, I remember one company's product detecting all binaries produced with GCC (on Windows) as "suspicious" and quarantining/deleting them, while those produced with MSVC were fine. Lots of "why can't I find the 'hello world' exe I compiled" questions on forums and newsgroups...
I think I can see the ransomware developers, criminals, scammers and malware writers celebrating in the background on the capabilities of this new species of αcτµαlly pδrταblε εxεcµταblεs for a new generation of a single all in one cross-platform portable binary payload that can run on six or more OSes.
Once again, the road to hell is paved with good intentions, and it is going to get much more chaotic with this thing aiding a whole new evolution of easy single binary payload deployments to the arsenal of the ransomware griftopia.
Malware author can’t send you an email with an attachment targeted at the OS(es) you happen to read the email on. So this definitely helps to some extent.
the difference with SQLite is that this is the same binary file that can be run on different oses, with SQLite you need the binary compiled for the target platform
I ported a C codebase (routino) to an redbean APE, it was a little bit of work but not alot. If fossil does not use many external dependencies it'd probably not be that much work, especially considering sqlite is already there.
this statement is not in line with the current reality for app downloads.
Yes, it could be smaller, but that does not mean it is "not small" I'm surprised when i download an app and it's less than 100MB these days.
that's not to say that 100MB is good, or that it's reasonable that downloads are as big as they are, but the fact is that they _are_ way bigger than 2.1MB. As designating something "small" or "large" is literally a relative measurement. So, unless you're comparing it against an atypical set of apps, or apps from a prior era of development, it is definitely "small".
No, I wrote "binary". Also, Featherwiki is not an app.
Most apps I download from f-droid are quite small and the bigger ones contain images, sounds and even animations.
2 MB are a lot for just CPU instructions assuming we are talking about compiled languages.
Of course some languages can create very large binaries anyways - but that's due to their inefficiency.
In fact, ls -Ssrh /usr/bin shows that the vast majority of binaries are small than 200KB, ignoring symlinks, and the outliers are databases, compiler and few other very large applications.
> In order to keep Feather Wiki small, many of its method names are extra short and can seem somewhat opaque. Hopefully this documentation will help make things a little more clear.
What kind of size difference would it make to have longer method names?
28 comments
[ 2.9 ms ] story [ 72.4 ms ] threadIt pains me to see that MS Defender marks it as suspicious, so I can't easily hand it to my non-tech friends. I understand why this happens though :(
Around 2 decades ago, I remember one company's product detecting all binaries produced with GCC (on Windows) as "suspicious" and quarantining/deleting them, while those produced with MSVC were fine. Lots of "why can't I find the 'hello world' exe I compiled" questions on forums and newsgroups...
Once again, the road to hell is paved with good intentions, and it is going to get much more chaotic with this thing aiding a whole new evolution of easy single binary payload deployments to the arsenal of the ransomware griftopia.
Yes, it could be smaller, but that does not mean it is "not small" I'm surprised when i download an app and it's less than 100MB these days.
that's not to say that 100MB is good, or that it's reasonable that downloads are as big as they are, but the fact is that they _are_ way bigger than 2.1MB. As designating something "small" or "large" is literally a relative measurement. So, unless you're comparing it against an atypical set of apps, or apps from a prior era of development, it is definitely "small".
Most apps I download from f-droid are quite small and the bigger ones contain images, sounds and even animations.
2 MB are a lot for just CPU instructions assuming we are talking about compiled languages.
Of course some languages can create very large binaries anyways - but that's due to their inefficiency.
In fact, ls -Ssrh /usr/bin shows that the vast majority of binaries are small than 200KB, ignoring symlinks, and the outliers are databases, compiler and few other very large applications.
Is there something like this which supports multiple users, permissions etc?
> In order to keep Feather Wiki small, many of its method names are extra short and can seem somewhat opaque. Hopefully this documentation will help make things a little more clear.
What kind of size difference would it make to have longer method names?
[1] https://feather.wiki/?page=documentation_intro