| -!- Netsplit *.net <-> *.split quits: heroux | 02:14 | |
| -!- Netsplit over, joins: heroux | 02:15 | |
| stekern | I'm starting to part more and more with derRichard, let's get rid of uClibc and start using glibc instead | 10:18 |
|---|---|---|
| jeremybennett | stekern: It's a *lot* bigger. But why not do both? | 11:04 |
| stekern | sure | 11:06 |
| jeremybennett | Anyone seen jonibo recently. I emailed him about the problems with the openrisc.net mailing lists last week and haven't had any reply. | 11:24 |
| stekern | this c++ library is starting to look like frankensteins monster... | 12:02 |
| stekern | a little libsupc++ mixed with hacked uclibc mixed with libc++... | 12:17 |
| stekern | jeremybennett: jonibo have been awfully quiet lately | 12:18 |
| jeremybennett | stekern: That's what I thought. Hope he's OK | 12:36 |
| derRichard | stekern: :-) | 13:05 |
| derRichard | yeah glibc is bigger. but it's the only way to provide a sane/complete linux userspace | 13:06 |
| stekern | so what's the magnitude of bigger? a quick comparison of just libc.so for glibc and uclibc shows about 3x (500k vs 1.5M), but that's comparing uclibc for or1k and glibc for x86 | 13:38 |
| derRichard | most embedded systems totay use glibc | 13:39 |
| derRichard | because they need sane pthread support, etc... | 13:39 |
| derRichard | AFAIM glibc is not that big | 13:39 |
| derRichard | *AFAIK | 13:39 |
| derRichard | btw: you know http://www.eglibc.org/? | 13:40 |
| derRichard | this is basically glibc plus patches to make it smaller... | 13:40 |
| stekern | I've heard about it, but have no hands-on experience with it | 13:41 |
| derRichard | debian is using it as default libc | 13:41 |
| stekern | aha, I wasn't aware of that | 13:41 |
| derRichard | but most eglibc features comes back to glibc now. because drepper is no longer maintaining glibc and some eglibc guys took over the glibc maintaince :) | 13:42 |
| derRichard | IOW drepper was kicked out of glibc | 13:43 |
| jeremybennett | derRichard: I think you need to define "embedded systems" a bit more clearly. I'm prepared to believe that embedded control processors in SmartPhones use glibc. Is it really the case for things like routers, which although they run Linux/BusyBox do not have vast memory resources. | 13:43 |
| jeremybennett | And it is certainly not the case for most embedded systems, because they don't run Linux. | 13:44 |
| derRichard | jeremybennett: the main point is, you need glibc support. if someone wants openrisc on a micro embedded system, he can still use uclibc, klibc, etc... | 13:45 |
| jeremybennett | I'm not arguing against glibc - just making the point that it isn't the answer for "most" embedded systems. | 13:45 |
| jeremybennett | jonibo: Good morning. Did you get my email about the openrisc.net mailing lists. | 13:45 |
| jeremybennett | I'm all for glibc on OpenRISC. | 13:46 |
| jonibo | yeah, got it... just checked this morning and your account looks fine | 13:46 |
| jonibo | no reason why you shouldn't be receiving emails aside from the fact that there's almost no traffic on the lists at the moment | 13:47 |
| jeremybennett | Hmmm. I wonder why I haven't been getting the emails. I'll check my filters to see they aren't throwing stuff away. | 13:47 |
| jonibo | like I said, there hasn't been much to receive anyway | 13:47 |
| jonibo | was someone against glibc on openrisc??? | 13:48 |
| jeremybennett | Agreed - but I noticed a number of cross-posted emails to both opencores.org and openrisc.net lists, and I only got the opencores.org ones. | 13:48 |
| jeremybennett | No - just me being picky about whether it is the library used on "most" embedded systems. | 13:48 |
| stekern | yeah, I haven't got any mail from openrisc.net since august 22 | 13:48 |
| jeremybennett | I wonder what happened on that date - same date I stopped getting emails | 13:49 |
| jonibo | is there anything in the archives since then??? | 13:49 |
| jeremybennett | Yes 20 messages in September alone | 13:49 |
| jeremybennett | http://lists.openrisc.net/pipermail/openrisc/2012-September/date.html | 13:50 |
| jeremybennett | derRichard was making the very valid point that if you want a fully functional pthread implementation you have to have glibc. | 13:51 |
| jonibo | ah ok... I haven't been receiving those emails either... will look into it | 13:52 |
| jonibo | ok, should be working now... let me know if mail issues continue | 14:07 |
| jeremybennett | thx | 14:28 |
| juliusb | jonibo: good to hear from you again! | 20:06 |
| juliusb | you going to be around Stockholm October 13? | 20:06 |
| kristianpaul | stekern: hi, sorry ask this that much,but ... did you finally got permission to release the openrisc port for the milkymist one board? | 21:36 |
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!