SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux
SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux
SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux
belgacem1958 has joined #crux
<belgacem1958>
while booting from usb media with the latest Iso, i have message like "id" c1" respawning too faste: disabled for 5 minutes ". I can't install
<SiFuh>
belgacem1958: how much RAM you got?
<belgacem1958>
have 8 GB of ram
<SiFuh>
That's plenty. Are you sure the transfer to USB is error free?
<belgacem1958>
used rufus, fedora media writer under win10
<SiFuh>
I don't even know if it works with rufus.
<SiFuh>
belgacem1958: there are a few reasons for this error, but the most common is missing files.
<belgacem1958>
I downloaded it from ninja
<SiFuh>
It can be many reasons for missing files. Bad memory, not enough memory, ISO to USB didn't work correctly, USB loaded to memory didn't work. A USB port is faulty, a faulty USB drive
<belgacem1958>
okay, thank you very much
<SiFuh>
A corrupt downlaod, or the actual ISO you downloaded is corrupt on the server side.
belgacem1958 has quit [Remote host closed the connection]
ppetrov^ has quit [Quit: Leaving]
<farkuhar>
The 'prt-get sysup' sorting algorithm might require further tweaking. Just now it tried to update gcc-fortran *before* gcc, and thus /usr/lib/gcc/$arch/$version was populated with header files owned by the wrong port.
<farkuhar>
Running 'pkgadd -u -f gcc#12.3.0-1.tar.xz' was enough to reset the pkg database with the correct file ownership, but a smarter sorting algorithm would prevent the "conflicting files" error in the first place.
<farkuhar>
If gcc-fortran had declared gcc as a dependency, then the existing sorting algorithm would work, but it's frowned upon to list core ports among the dependencies unless they're dynamically linked in. Making 'prt-get sysup' aware of this convention could require sorting two different sets---the out-of-date core ports, and the out-of-date non-core ports---rather than the one set that is currently sorted for a
<farkuhar>
sysup operation.
<chrcav>
farkuhar: seems like a core only or core first option for sysup would be nice for this situation but then it would be nice to see what core ports need updates. A note could be added to the handbook under prt-get sysup to indicate a way to update core ports ahead of a sysup operation.
<SiFuh>
prt-get sysup --core --opt --contrib
<SiFuh>
:-P
<SiFuh>
Although --opt should include --core automatically
<chrcav>
I guess one pitfall with sorting core separate from the other repos is that some people may not have core. It's possible to rename the folder.
<SiFuh>
What a weird thing to do
<chrcav>
I agree. It would be a weird thing to do.
<chrcav>
I would settle for running two commands one for limiting to a single repo such as core and then sysup a second time to update all other repos. It would be helpful if prt-get diff supported the path option.
<chrcav>
It's not that hard to create a core only prt-get config and use --config=
<SiFuh>
ports -u core && prt-get sysup
<chrcav>
That's another good idea.
ppetrov^ has joined #crux
DaViruz has quit [Ping timeout: 240 seconds]
DaViruz has joined #crux
joacim has quit [Ping timeout: 240 seconds]
joacim has joined #crux
<farkuhar>
chrcav: Thanks for the ideas. If you've already synced all your repos (too late for SiFuh's solution), you can still do a core-only sysup as follows: prt-get printf '%i:%p:%n\n' | grep '^diff:/usr/ports/core' | cut -d: -f3 | xargs prt-get update
<SiFuh>
farkuhar: yeah too late for me. Can test another time
<SiFuh>
Oops. Speed reading again and skipped the word solution :-D