OpenWRT and LEDE agree on Linux-for-routers peace plan Competing Linux-for-routers distributions OpenWRT and LEDE will soon vote on a proposal to heal the schism between the two. OpenWRT is often used as firmware for small routers, largely SOHO WiFi kit. But in March 2016 a group of developers decided they didn't like the directions OpenWRT was taking and forked the project by …
Forking can be good, sometimes it's just ego.
I never exactly understood the reason for this particular split, so I don't know.
OpenOffice -> LibreOffice, MySQL -> MariaDB, Hudson -> Jenkins forks all have one common explanation. Oracle. Actually even Davik vs JVM and C# from J++ are result of Sun/Oracle pressure and licences (J++ was too like Java, but not Java. The "full" Desktop Java is "free", but you have to licence a cut down version for phones, Sun or Oracle wouldn't agree a price for the free desktop version for phones, hence Davik. Google bought in Android)
The Linux Desktop GUI forks made sense too, as some teams drank the the worst MS & Apple "koolaid" and went for either "eye candy" or a fantasy cross device (one GUI for phone, tablet, TV and Desktop inherently is always going to be a failure). Choice between Classic Xerox 1976 principles (Older Mac OS, Win9x/Win2000 as in Mate with Redmond Theme), Ultra-flat and daft 3D eye-candy as a user preference seems a good idea.
I can understand the Systemd vs Unix Init split, even though I don't understand the details, nor know which is "right", thus choice is good.
I was active in the OpenWRT community but sort of zoned out about a year before the split so am not completely on top of it. From what I gather from the dev mailinglist, however, is that is not necessarily an issue in the code itself that was a reason for the fork. I get the sense that it was more of a context / project management thing. A chunk of developers wanted to use more modern dev tools (I believe moving to GitHub for instance), faster dev cycles etc. but the lead maintainer of OpenWRT wasn't too keen on it.
Considering the unhappy coders were a fairly large chunk of all active coders, a fork was easily done and it looks as if LEDE very quickly overtook OpenWRT as the more active one. An OpenOffice/LibreOffice analogy comes to mind.
I just hope they take the best bits of both projects (I wouldn't have minded if they kept the LEDE name but I get the resistance to dropping the well-recognised OpenWRT name). The LEDE website is a lot better than OpenWRT's one if you ask me and I hope they use this opportunity to sort out all their documentation and hardware compatibility tables as they were always all over the place, lots of redundancies etc.
DD-WRT forked from OpenWRT too.
I've used OpenWRT and LEDE on a number of devices (a bunch of BT HomeHub5's) and it seemed a bit fragmented as to which site / code base one was best for any one topic. LEDE seems to be better in terms of functionality and the overlay file system is great for config and upgrades (run sysupgrade and you are done), but the documentation on how to tweak a particular router seemed more up-to-date on the OpenWRT site. I just hope that they merge the "good" from both projects - like dropping the pointless drinks recipe from the login process
I kind of expected the target project to be called something like LibreWRT, but I guess that OpenWRT has a certain heritage.
Do you also have scales (or feathers according to some paleontologists) and limber around on two legs while emitting thundering noises and the pesky mammals scurrying under your feet?
Tomato is beyond prehistoric and has not fixed a long list of issues for a nearly a decade. It a paleontological exhibit worthy of a place in the museum of technology, not actual production use.
Do not understand me wrong - it was great in its day. It was the first proper open source embedded build. That was about a decade ago.
"I use Tomato on all my routers."
Oh, so yours are too old to fit anything newer in the NAND too... ;)
"Tomato is beyond prehistoric"
I found it thoroughly amusing that at the time of the Great Heartbleed Extinction Event, Tomato turned out to be immune due to its sheer, uh, age...
I hope merging stabilises things. I'm a great fan of OpenWrt and LEDE, but the LEDE update to then current OpenWrt code broke the BearDropper app (which is a kind of Fail2Ban, written as an OpenWrt shell script), and I don't have the time to develop the skills to fix it myself.
It pushed me to move SSH to a non-standard port number* on the routers controlled by me, which was good, because all the automated access attempts on the standard SSH port were irritating me (they weren't going to get in, but they kept filling up the log).
I have a mixture of OpenWrt, LEDE and DrayOS (Draytek) routers, and I far prefer the OpenWrt/LEDE approach. SSH access to the Draytek devices from Linux appears to be 'somewhat' buggy, and getting some bog-standard VPNs up seems to be harder than it should be - but that could be my ignorance/idiocy at work.
*Top tip: add a firewall rule allowing access for the new port number BEFORE changing the Dropbear config. That way you don't have a long journey to re-enable access to remote routers.
Does this mean they'll start properly supporting the WRT54GS chipset now? I walked away from OpenWRT when the response to my problem (nothing serious, just the router hanging and dropping all packets if you tried to get more than 1Mbit/s throughput) was (let me think... oh, yes...)
"Can't be arsed, they should have designed the chipset better".
The conversation is right there in the mailing list.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
