[pacman-dev] Blockers for releasing pacman-6.0.0?
Hi all, I think the 6.0.0 release is roughly ready to go. Here are my known blockers: 1) Remove "total download" callback There is a patch submitted which I briefly looked at. Need to do a proper review, but looked fine apart from copy/paste issue with comment. 2) prevent spamming at the end of downloads when the top server in the mirrorlist is bad. I sent an RFC for this. It is not the best of patches, but serves as a start... The only other things I have in the roadmap on Flyspray are: FS#61497 - add a timestamp file to repos FS#65677 - [makepkg] not all source files get included in debug packages Both are nice to have, but not regressions so can be delayed a release (unless someone else propose patches in the meantime). Is there anything else that people think is essential to get in pacman-6.0.0? I know there are a few other downloader issues listed in the meta-bug, but I think these are not critical and can be future improvements. Allan
Hi Allan On Sun, Mar 7, 2021 at 8:38 PM Allan McRae <allan@archlinux.org> wrote:
Hi all,
I think the 6.0.0 release is roughly ready to go.
Here are my known blockers:
1) Remove "total download" callback There is a patch submitted which I briefly looked at. Need to do a proper review, but looked fine apart from copy/paste issue with comment.
I just sent an update with a new comment. Still waiting for a feedback from Andrew as it is the item we've discussed with him at IRC channel. cc-ing him to this thread.
2) prevent spamming at the end of downloads when the top server in the mirrorlist is bad. I sent an RFC for this. It is not the best of patches, but serves as a start...
The only other things I have in the roadmap on Flyspray are:
FS#61497 - add a timestamp file to repos FS#65677 - [makepkg] not all source files get included in debug packages
Both are nice to have, but not regressions so can be delayed a release (unless someone else propose patches in the meantime).
Is there anything else that people think is essential to get in pacman-6.0.0? I know there are a few other downloader issues listed in the meta-bug, but I think these are not critical and can be future improvements.
A few people reported about corrupted UI for parallel downloads. It seems to be related to "slow downloads" or "failed downloads" situations. I would really love to see this bug fixed. Though I do not know how to reproduce the problem.
Em março 8, 2021 1:55 Anatol Pomozov escreveu:
A few people reported about corrupted UI for parallel downloads. It seems to be related to "slow downloads" or "failed downloads" situations. I would really love to see this bug fixed. Though I do not know how to reproduce the problem.
I get this occasionally when using our tier 0 mirror and pacman 6.0.0 alpha. I think you could reproduce on a VM/container by capping the download speed to less than 100Kb/s. Regards, Giancarlo Razzolini
On 8/3/21 9:43 pm, Giancarlo Razzolini wrote:
Em março 8, 2021 1:55 Anatol Pomozov escreveu:
A few people reported about corrupted UI for parallel downloads. It seems to be related to "slow downloads" or "failed downloads" situations. I would really love to see this bug fixed. Though I do not know how to reproduce the problem.
I get this occasionally when using our tier 0 mirror and pacman 6.0.0 alpha. I think you could reproduce on a VM/container by capping the download speed to less than 100Kb/s.
I am regularly on networks with less than 100Kb/s and do not see that issue. So perhaps more than just speed issues. Allan
Em março 8, 2021 8:50 Allan McRae escreveu:
On 8/3/21 9:43 pm, Giancarlo Razzolini wrote:
Em março 8, 2021 1:55 Anatol Pomozov escreveu:
A few people reported about corrupted UI for parallel downloads. It seems to be related to "slow downloads" or "failed downloads" situations. I would really love to see this bug fixed. Though I do not know how to reproduce the problem.
I get this occasionally when using our tier 0 mirror and pacman 6.0.0 alpha. I think you could reproduce on a VM/container by capping the download speed to less than 100Kb/s.
I am regularly on networks with less than 100Kb/s and do not see that issue. So perhaps more than just speed issues.
Sure, could be a situation of both slow network and packet loss. Also, I didn't tried with pacman-git in about a month, so things might have changed. But the issue described is one I can confirm seeing when you asked for testing with the alpha. Regards, Giancarlo Razzolini
On 03/08/21 at 08:55am, Giancarlo Razzolini wrote:
Em março 8, 2021 8:50 Allan McRae escreveu:
On 8/3/21 9:43 pm, Giancarlo Razzolini wrote:
Em março 8, 2021 1:55 Anatol Pomozov escreveu:
A few people reported about corrupted UI for parallel downloads. It seems to be related to "slow downloads" or "failed downloads" situations. I would really love to see this bug fixed. Though I do not know how to reproduce the problem.
I get this occasionally when using our tier 0 mirror and pacman 6.0.0 alpha. I think you could reproduce on a VM/container by capping the download speed to less than 100Kb/s.
I am regularly on networks with less than 100Kb/s and do not see that issue. So perhaps more than just speed issues.
Sure, could be a situation of both slow network and packet loss. Also, I didn't tried with pacman-git in about a month, so things might have changed. But the issue described is one I can confirm seeing when you asked for testing with the alpha.
Regards, Giancarlo Razzolini
My sync-server-failover branch has a patch to simulate the network failure as well as test cases that can do the same.
participants (4)
-
Allan McRae
-
Anatol Pomozov
-
Andrew Gregory
-
Giancarlo Razzolini