Re: Build servers going offline due to inclement weather
Ok, thanks. In the future then, I might do that if after 2-3 days, the outage was expected to continue for perhaps that much more, so that I didn't get too far behind; and I'd probably blow away the clone and switch back after things were back to normal (since it wouldn't be worth the slowdown the rest of the time). Basically, I like it if I can keep a "port sync;port upgrade outdated" down to under 10 minutes, not counting compiler recompiles or other really slow builds. That's because I have 21 different systems or VMs (or LDOMs or zones) (only 5 of which are Macs, the rest are Solaris, CentOS, Ubuntu, or a CentOS-similar SPARC Linux) that I do semi-scripted updates (each in its own Terminal tab) to all at once, and the slowest non-Macs (which, to be fair, aren't compiling anything) never take that long. That's not counting a couple of oddballs done separately because there's a limit to how many VMs at once I want to run on a host, or the three Windows instances (RDP access plus a script that runs 6 different GUI tools to do updates, since they also have WSL, Cygwin, and iCloud/iTunes on them; PatchMyPC does the none-of-the-above on those). Yeah, that's at home. I used to work in a place with a LOT of toys, and ran quite a few of 'em, and when I retired, so I wouldn't miss the toys too much, I have a few of my own (mostly virtual). Also mainframe emulators, PDP-11 emulator, old workstation emulators, etc - most of it something I used before one time or another. > On Feb 20, 2021, at 02:26, Ryan Schmidt wrote: > > On Feb 20, 2021, at 00:54, Richard L. Hamilton wrote: > >> Thank you for the information, and for your work! >> >> Question: if someone changed their configuration as you mentioned to get >> Portfiles from GitHub, would they EVER (in principle) see a version of a >> Portfile that would NEVER make it to the rsync server (if/when the >> buildmaster etc was up)? In other words, I might change my configuration >> (so as to be less affected by similar future outages) if I'm not risking >> getting a Portfile that's different from what I'd get otherwise (just >> sooner, sometimes), and assuming there's no other downside to me from change. > > What's on the rsync server is a time-delayed copy of what's on GitHub. > Usually the delay is only up to about an hour, when the server that does the > copying isn't offline. > > The downsides of using GitHub instead of rsync are that it will take more > space on your disk (because you have a clone of the entire history of the > repository, not just the current files as you would have with rsync) and it > will take more time when you sync (because you will be generating the > portindex locally instead of getting a copy from the rsync server). > > smime.p7s Description: S/MIME cryptographic signature
Re: Build servers going offline due to inclement weather
On Feb 20, 2021, at 00:54, Richard L. Hamilton wrote: > Thank you for the information, and for your work! > > Question: if someone changed their configuration as you mentioned to get > Portfiles from GitHub, would they EVER (in principle) see a version of a > Portfile that would NEVER make it to the rsync server (if/when the > buildmaster etc was up)? In other words, I might change my configuration (so > as to be less affected by similar future outages) if I'm not risking getting > a Portfile that's different from what I'd get otherwise (just sooner, > sometimes), and assuming there's no other downside to me from change. What's on the rsync server is a time-delayed copy of what's on GitHub. Usually the delay is only up to about an hour, when the server that does the copying isn't offline. The downsides of using GitHub instead of rsync are that it will take more space on your disk (because you have a clone of the entire history of the repository, not just the current files as you would have with rsync) and it will take more time when you sync (because you will be generating the portindex locally instead of getting a copy from the rsync server).
Re: Build servers going offline due to inclement weather
Thank you for the information, and for your work! Question: if someone changed their configuration as you mentioned to get Portfiles from GitHub, would they EVER (in principle) see a version of a Portfile that would NEVER make it to the rsync server (if/when the buildmaster etc was up)? In other words, I might change my configuration (so as to be less affected by similar future outages) if I'm not risking getting a Portfile that's different from what I'd get otherwise (just sooner, sometimes), and assuming there's no other downside to me from change. > On Feb 20, 2021, at 01:05, Ryan Schmidt wrote: > > On Feb 19, 2021, at 21:38, Richard L. Hamilton wrote: >> >> While in principle most updates that fetch source might still work, I've >> seen few if any since the outage began, and I'm curious about that. Haven't >> seen a fetch failure, just no updates showing up. > > The buildmaster is also responsible for sending the updated set of ports to > the rsync server along with newly generated portindexes. Since the > buildmaster is down, any port updates that have been committed over the past > days have not been sent to the rsync server so you don't see them in `port > outdated' (unless you have configured your MacPorts installation's > sources.conf to get the Portfiles from GitHub rather than from rsync -- > normally only MacPorts developers do that). > > As of Friday the power grid is stable and energy conservation is no longer > required so I'll be bringing the build servers back up and you should see > updates appearing normally within the next few hours. > > https://twitter.com/ERCOT_ISO/status/1362820173078687744 > > smime.p7s Description: S/MIME cryptographic signature
Re: Build servers going offline due to inclement weather
On Feb 19, 2021, at 21:38, Richard L. Hamilton wrote: > > While in principle most updates that fetch source might still work, I've seen > few if any since the outage began, and I'm curious about that. Haven't seen a > fetch failure, just no updates showing up. The buildmaster is also responsible for sending the updated set of ports to the rsync server along with newly generated portindexes. Since the buildmaster is down, any port updates that have been committed over the past days have not been sent to the rsync server so you don't see them in `port outdated' (unless you have configured your MacPorts installation's sources.conf to get the Portfiles from GitHub rather than from rsync -- normally only MacPorts developers do that). As of Friday the power grid is stable and energy conservation is no longer required so I'll be bringing the build servers back up and you should see updates appearing normally within the next few hours. https://twitter.com/ERCOT_ISO/status/1362820173078687744
Re: ARM gdb compilation fails with implicit function declaration error ioctl
On Feb 19, 2021, at 15:14, Dave Horsfall wrote: > On Fri, 19 Feb 2021, Ryan Schmidt wrote: > > [...] > >> As I said in my previous message from December: >> >> * Don't forget to Reply All so the conversation stays on the list. > > Better still, set the reply to the list only? I have a beef with "Reply > All" because I don't need my own personal copy. > > It's a simple setting in Mailman. We've answered this question many times. Doing so would contravene internet standards so we won't do that. https://woozle.org/papers/reply-to-still-harmful.html
Re: Build servers going offline due to inclement weather
While in principle most updates that fetch source might still work, I've seen few if any since the outage began, and I'm curious about that. Haven't seen a fetch failure, just no updates showing up. > On Feb 19, 2021, at 15:26, Bill Cole > wrote: > > On 19 Feb 2021, at 15:12, Todd Doucet wrote: > >> Many new users, like me, might not know what the consequences of the build >> servers being down are, exactly. I am speculating that this does not mean >> that MacPorts is not available generally, but instead that if I ask for a >> port it might build it on my machine instead of pulling a pre-built binary >> for me. > > Correct. In some cases there could also be problems with getting source > distributions of some ports automatically. > > From Ryan's initial message: > >> Binary archives of newly updated ports will not be produced during this time >> so MacPorts will compile from source on your computer when needed. Source >> distfiles will not be mirrored during this time either which may impact your >> ability to install some ports on older OS versions, particularly OS X 10.8 >> and earlier. If you get an error during the fetch phase, consult: >> >> https://trac.macports.org/wiki/ProblemHotlist#fetch-failures > > That link explains that if you can find and manually fetch the right > distfile, you can work around a 'port fetch' failure. > > -- > Bill Cole > b...@scconsult.com or billc...@apache.org > (AKA @grumpybozo and many *@billmail.scconsult.com addresses) > Not Currently Available For Hire > smime.p7s Description: S/MIME cryptographic signature
Re: ARM gdb compilation fails with implicit function declaration error ioctl
On Fri, 19 Feb 2021, Ryan Schmidt wrote: [...] As I said in my previous message from December: * Don't forget to Reply All so the conversation stays on the list. Better still, set the reply to the list only? I have a beef with "Reply All" because I don't need my own personal copy. It's a simple setting in Mailman. -- Dave
Re: Build servers going offline due to inclement weather
On 19 Feb 2021, at 15:12, Todd Doucet wrote: Many new users, like me, might not know what the consequences of the build servers being down are, exactly. I am speculating that this does not mean that MacPorts is not available generally, but instead that if I ask for a port it might build it on my machine instead of pulling a pre-built binary for me. Correct. In some cases there could also be problems with getting source distributions of some ports automatically. From Ryan's initial message: Binary archives of newly updated ports will not be produced during this time so MacPorts will compile from source on your computer when needed. Source distfiles will not be mirrored during this time either which may impact your ability to install some ports on older OS versions, particularly OS X 10.8 and earlier. If you get an error during the fetch phase, consult: https://trac.macports.org/wiki/ProblemHotlist#fetch-failures That link explains that if you can find and manually fetch the right distfile, you can work around a 'port fetch' failure. -- Bill Cole b...@scconsult.com or billc...@apache.org (AKA @grumpybozo and many *@billmail.scconsult.com addresses) Not Currently Available For Hire
Re: Build servers going offline due to inclement weather
Many new users, like me, might not know what the consequences of the build servers being down are, exactly. I am speculating that this does not mean that MacPorts is not available generally, but instead that if I ask for a port it might build it on my machine instead of pulling a pre-built binary for me. > Hello Ryan, > > On 19/02/2021 16.38, Ryan Schmidt wrote: > > Build servers remain offline. Since my last update, power was out here for > > 32 hours but is back on now. Power is restored in most of Austin but the > > Texas electric grid operator still asks us to conserve electricity. > > Thank you for the update. I hope you are well and stay warm! > > Rainer >
Re: Build servers going offline due to inclement weather
Hello Ryan, On 19/02/2021 16.38, Ryan Schmidt wrote: > Build servers remain offline. Since my last update, power was out here for 32 > hours but is back on now. Power is restored in most of Austin but the Texas > electric grid operator still asks us to conserve electricity. Thank you for the update. I hope you are well and stay warm! Rainer
Re: Build servers going offline due to inclement weather
Build servers remain offline. Since my last update, power was out here for 32 hours but is back on now. Power is restored in most of Austin but the Texas electric grid operator still asks us to conserve electricity. https://twitter.com/ERCOT_ISO/status/1362780731001499650
Re: ARM gdb compilation fails with implicit function declaration error ioctl
On Feb 17, 2021, at 15:49, Alex Alexandrov wrote: > Hello. > I have same problem with building gdb for macbook pro m1. > Do you have any suggestions how to fix it? > — > Yours, Alex As I said in my previous message from December: * Don't forget to Reply All so the conversation stays on the list. * Implicit function declarations are not allowed. Report the problem to the developers of the software in question so that they can fix it.