L8ter
RC Linux bloat does not one’s ire provoke, says sometimes arsey Linux bloke.
Progress on Linux 4.10 has to date been steady. So steady, in fact, that it's not been worth reporting. But that changed on Sunday when Linus Torvalds declared that release candidate six, a previously “small and calm release candidate somehow blew up to not be all that small after all.” Torvalds isn't panicking because he's …
GPU bloating the RC doesn't really surprise me, given some of the work being done in wine and in HPC, and the HPC GPU code will need quite a number of fixes to networking path code.
That said, its January. Everyone is back at work. So Linus' quiet time through the winter solstice is over.
>How cute, they still use a mailing list!
What do you suggest they use, Skype for Business ? ROFL! Teams ? [Pukes guts out and dies]!
It so happens that a mailing list fulfills the task perfectly, has been for DECADES, no need to "improve" what doesn't need fixin' ... next, you'll say they need a nice new tiled touch-only gui to go with it ?
It so happens that a mailing list fulfills the task perfectly...
A newsgroup (perhaps on a private NNTP server) might be better, but not everyone likes to let NNTP traffic through their firewall so a mailing list is a pragmatic choice.
... next, you'll say they need a nice new tiled touch-only gui to go with it ?
Hah! Kids these days ...
A newsgroup is not better. Mailing lists can be filtered and sorted into folders to the receiver's liking, can be accessed on multiple clients syncing the read/unread/tagged status. A newsgroup requires the same name everywhere without any server-side support for flags and tags.
NNTP is mostly dead. Mailing lists rule.
The initial "fixes" ARE a bloat to the size of the codebase, but it can be made smaller, given more time. That is what I gathered from reading the article. If you ever wrote code that is the point of good coding practice; not merely making your thing work, but making it work in fewer lines and less processing/resource usage overall. No? I value "tight code" rather than just "make it work, however large and shitty, and release it like the dump that it is" code. There is a certain level of elegance and clever thinking that goes into really well designed code, even if the code is small like the tiny Atari 2600 games of old. In the larger picture, those bulk of code fixes look huge and need trimming down, according to Linus. And I would tend to agree. Especially where security is involved. For example, fixes for security are going to be top priority and get released right away, yet the code can be cleaned up after the initial vector is closed down saving space or calls or any number of things to improve it.