Created attachment 218821 [details] Patch to 3.1.0 Maintainer update, 3.1.0 release. Important note on Feedparser: The Feedparser option is currently broken (requires feedparser >= 6.0.0). I have submitted a patch against bug 250380 to bring feedparser up to date (6.0.1). The port otherwise works fine.
Created attachment 218822 [details] Poudriere build log
^Triage: Bugfix release, merge to quarterly branch. ^Triage: Maintainer-feedback flag (+) not required unless requested (?) first. Thanks!
Created attachment 219010 [details] Patch to 3.1.0 - Update cheroot Upstream were holding with a dependency of <= 8.4.3 on www/py-cheroot due to a bug in 8.4.4. 8.4.5 has hit the ports tree, and has a fix but may have an additional regression. Updating the patch to be without the dependency issue. Previous versions of sabnzbd would be similarly affected, so we'll see if anyone has issues and go from there.
A commit references this bug: Author: adamw Date: Mon Oct 26 19:45:48 UTC 2020 New revision: 553378 URL: https://svnweb.freebsd.org/changeset/ports/553378 Log: news/sabnzbdplus: Update to 3.1.0 PR: 250411 Submitted by: maintainer Changes: head/news/sabnzbdplus/Makefile head/news/sabnzbdplus/distinfo head/news/sabnzbdplus/pkg-plist
I've committed the patch here, but I'm holding off on the MFH until we know there's no fallout from the regressions you mentioned. James, when you feel confident that it's safe to MFH, please let me know (or if you think this update doesn't need to be merged to quarterly).
(In reply to Adam Weinberger from comment #5) Ok, I've had a deeper dive through the issues, and I think what's left is log spam only and should be safe for this port. If anyone stumbles across this, the relevant upstream issues are: SabNZB: https://github.com/sabnzbd/sabnzbd/issues/1572 Cheroot: https://github.com/cherrypy/cheroot/issues/317 Cheroot 8.4.4 had a crash caused by a threading race condition. 8.4.5 the crash is resolved, but the race condition is still there which may cause log spam. No fix available yet but upstream SabNZB have given the ok on 8.4.5. In either case, the quarterly branch is on 8.4.3 which is safe from both issues. This version resolves a number of issues with SabNZB so merging quarterly is probably a good idea. Cheers
A commit references this bug: Author: adamw Date: Tue Oct 27 13:41:25 UTC 2020 New revision: 553440 URL: https://svnweb.freebsd.org/changeset/ports/553440 Log: MFH: r553378 Approved by: portmgr (with hat) news/sabnzbdplus: Update to 3.1.0 PR: 250411 Submitted by: maintainer Changes: _U branches/2020Q4/ branches/2020Q4/news/sabnzbdplus/Makefile branches/2020Q4/news/sabnzbdplus/distinfo branches/2020Q4/news/sabnzbdplus/pkg-plist
Created attachment 219156 [details] 3.1.0: Fix feedparser hard dependency An oversight on my part, feedparser is still in the default options list, but the patch to 6.0.1 on bug 250380 is still pending maintainer approval with still another 3 days before timeout. Builds are failing with the default option set. Further to that, it looks like this version has made feedparser a hard requirement too so just disabling the option introduces a runtime problem. Attached makes feed parser a hard dependency, but the port will be broken until the other bug is resolved. I am wondering if rolling back sabnzbd makes more sense at this time.
Ugh. I don't know why my poudriere was happy with this before commit. What is the state of bug 250380? Your last comment in there talks about marking the port broken. Is the patch ready for commit (maintainer approval aside)?
Ok well, sbz was never added to that PR so really the timeout period should probably start from today. I'll back this out of head and quarterly.
A commit references this bug: Author: adamw Date: Tue Oct 27 23:37:38 UTC 2020 New revision: 553512 URL: https://svnweb.freebsd.org/changeset/ports/553512 Log: news/sabnzbdplus: Back out r553378 Feedburner needs to be a hard dependency, and the version currently in ports is incompatible. PR: 250411, 250380 Changes: head/news/sabnzbdplus/Makefile head/news/sabnzbdplus/distinfo head/news/sabnzbdplus/pkg-plist
A commit references this bug: Author: adamw Date: Tue Oct 27 23:38:36 UTC 2020 New revision: 553513 URL: https://svnweb.freebsd.org/changeset/ports/553513 Log: MFH: r553512 Approved by: portmgr (with hat) news/sabnzbdplus: Back out r553378 Feedburner needs to be a hard dependency, and the version currently in ports is incompatible. PR: 250411, 250380 Changes: _U branches/2020Q4/ branches/2020Q4/news/sabnzbdplus/Makefile branches/2020Q4/news/sabnzbdplus/distinfo branches/2020Q4/news/sabnzbdplus/pkg-plist
(In reply to Adam Weinberger from comment #9) > Ugh. I don't know why my poudriere was happy with this before commit. Mine was too, but I might have had the 6.0.1 feedparser patch in my overlay ports tree which was a mistake. I did set this bug as depending on bug 250380, but it broke considerably more spectacularly than I was expecting. > Your last comment in there talks about marking the port broken. Is the patch ready for commit (maintainer approval aside)? Ready to go aside from the clean-up you've identified which I'll fix now. The "broken" aspect is that the latest version of feedparser requires shipping a Python 2 standard library that has had 2to3 run on it. If we don't do that, the port is broken, if we do (which is a bit messy) it works fine. There's no getting around that unfortunately as upstream have yet to port the code properly to Python 3.
Created attachment 219582 [details] Update to 3.1.1 3.1.1 has been released, updating this patch. Still pending Feedparser update to >= 6.0.0 before this can be committed, see bug 250380
Created attachment 219583 [details] Poudriere build log
Created attachment 222897 [details] Update to 3.2.0 Still blocked by bug 250380 on textproc/py-feedparser
Created attachment 222898 [details] Poudriere build log
Created attachment 224134 [details] Update to 3.2.1 Feedparser 6.0.2 has been committed to the ports tree, this upgrade is no longer blocked. Bumping the patch to 3.2.1. Thanks for the patience all.
Created attachment 224135 [details] Poudriere build log
The SAB version that is currently in the tree (3.0.2,1) is now broken because of the Feedparser update. Is there anything I can do to help get SAB updated to 3.2.1? Traceback (most recent call last): File "/usr/local/bin/SABnzbd.py", line 1721, in <module> main() File "/usr/local/bin/SABnzbd.py", line 1224, in main sabnzbd.initialize(pause, clean_up, evalSched=True, repair=repair) File "/usr/local/lib/python3.7/site-packages/sabnzbd/decorators.py", line 36, in call_func return f(*args, **kw) File "/usr/local/lib/python3.7/site-packages/sabnzbd/__init__.py", line 323, in initialize rss.init() File "/usr/local/lib/python3.7/site-packages/sabnzbd/rss.py", line 48, in init __RSS = RSSQueue() File "/usr/local/lib/python3.7/site-packages/sabnzbd/rss.py", line 209, in __init__ patch_feedparser() File "/usr/local/lib/python3.7/site-packages/sabnzbd/rss.py", line 648, in patch_feedparser feedparser._FeedParserMixin.namespaces["http://www.newznab.com/DTD/2010/feeds/attributes/"] = "newznab" AttributeError: module 'feedparser' has no attribute '_FeedParserMixin'
The patch does need to be committed. I'm returning the PR to the pool so someone can make this happen.
I see the 3.2.1 patch attached, what exactly is the block? Can we please merge the patch? 3.0.2,1 is broken due to py38-feedparser-6 update.
No block from me (port maintainer), just needs someone to commit the patch to the tree. I've been dogfooding 3.2.1 for some time without issue. It was blocked by feed-parser being out of date, but is now broken because that issue has been fixed. 3.3.0 is about to drop upstream shortly, I was toying with raising a new bug (and closing this one) to give it a bit more visibility. This bug is now 7 months old and is a ways down the list.
Do we want to bump the importance? This affects me and clearly others as well - is the "Affects Only Me" possibly keeping this lower on the TODO list?
@James French: who could commit the patch?
A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/ports/commit/?id=903174f474abeb09977c1ec9ddce5957ac386344 commit 903174f474abeb09977c1ec9ddce5957ac386344 Author: James French <james@french.id.au> AuthorDate: 2021-05-05 09:09:43 +0000 Commit: Tim Bishop <tdb@FreeBSD.org> CommitDate: 2021-05-05 09:09:43 +0000 news/sabnzbdplus: Update to 3.2.1 PR: 250411 news/sabnzbdplus/Makefile | 13 ++++++------- news/sabnzbdplus/distinfo | 6 +++--- news/sabnzbdplus/files/sabnzbd.in | 2 +- news/sabnzbdplus/pkg-plist | 12 +++++++++++- 4 files changed, 21 insertions(+), 12 deletions(-)
Committed. Sorry that it took so long for somebody to do it.
@James French: Now that the patch was merged, can we release the port?
(In reply to Floren Munteanu from comment #28) The updated port is in the ports tree, that's the patch that was committed. If you're using git for your ports directory, git pull to get the updates and it's there right now ready to build. If you're using portsnap or nightly packages, it should come through in the next day or two when the automated tools catch up. If that's the case and you need it sooner, you can apply the patch to your own ports tree. Portsnap will replace the patched files when they come through, but it will replace them with the same file so no harm done. The quarterly branch (and packages) are still on 3.0.2, but feed parser 5.0.x is also still in the tree for the quarterly branch.
(In reply to James French from comment #29) Any idea when the FreeBSD:12:amd64 build will be processed? The check was done on 2010-05-10 but the repo build and processed dates are behind. All repos have been updated, except the one I need. How lovely. :) https://www.freshports.org/news/sabnzbdplus/
(In reply to Floren Munteanu from comment #30) FYI for posterities sake - It showed up on ports about 10 hours after the last comment looking for FreeBSD:12:amd64. This was about 6 days after the fix was checked in. I think the fix was likely checked in shortly after a build started and it seems to take about 3 days for the full build process to complete, so the fix got in the next build after the one it just missed.