Closed core-code closed 5 years ago
2018-12-07 is online ๐จ
@core-code Done.
phocus - I can't find this anywhere and the download requires a user account.
@ran-dall thanks! you're the best :) I will have a look on phocus
2018-12-08 is online
hey @ran-dall - phocus is up-to-date :)
@core-code Done. No problems today. ๐
@ran-dall thanks for your good work on the repairs! i'll have little time for HBC throughout the rest of the year but @suschizu is handling this nicely now ;)
@core-code NP. I'm also running a little light on time for the rest of the year. BUT I've actually automated a lot of the process so it's pretty easy to do. So it doesn't really matter...as long as I get a heads up, everything is peachy...๐
2018-12-09 is online :)
Done. ๐
electron - beta darktable - Release Candidiate
2018-12-10 is online - just a few updates today :)
@ran-dall thanks for the info! the electron and darktable - update will be ignored now sorry for the iridium-issue (see #55941) - it would be good for us to have the "real" version number in the version-stanza
@suschizu Thanks for keeping me in the loop! ๐โโ๏ธ
The updates have been sent. โ
Problems:
2018-12-11 is online
@ran-dall thanks for your contribution :) sorry for musescore - I thought I've dropped it from the list โฆ that seem to be a beta-release yes.
18-12-11 we've added a new check-list ( crgithub-2018-12-11.txt ) - the new script looks only for GitHub-releases ๐
i've updated the original post title and text with explanation about two new checks that should be uploaded to the server daily now
@suschizu Thanks for the constant communication! Way to go! โ
Done.
aseprite - https://github.com/Homebrew/homebrew-cask/pull/55996 nanowallet - https://github.com/Homebrew/homebrew-cask/pull/55989
@core-code These updates make sense. ๐ I updated my automation to support both outputs now. We'll see how they do in the next couple of days. ๐
awesome thanks! also, i've just added something to let you know if a PR is already open for the app in question - hope that helps ;)
2018-12-12 two new sheets online ๐ including the brand new crsparklefeed-check :D
@suschizu done. ๐
glyphs - https://github.com/Homebrew/homebrew-cask/pull/56040 - This one was acting strange. rightfont - 5.2.3 > 4.12 - It's already at a whole major release higher.
2018-12-13 all 3 lists online for today - much todo this time ๐ฃ
@ran-dall - you're awesome! Thanks for all your efforts ๐
@suschizu done. โ
boostnote - Pre-release - 0.11.12-pre brave-browser - Developer Channel v0.59.5|v0.59.5 => Beta Channel v0.59.6|v0.59.6 - ?? Never seen this bytecoin - Beta - 3.4.0-beta-20181212 jbrowse - RC - 1.16.0-rc1 nextiva - No newer version available
@ran-dall - wow โฆ that was fast ๐ฎ - thank you! ๐
we probably need some form of postprocessing for the github output as well if there are so many non-stable "latest releases" in there.
i suspect some developers don't get that the "latest release" in github should point to the latest STABLE one ;p
2018-12-14 cr-processed,crgithub and crsparklefeed are online - yay ๐ป
Done
datadog - https://github.com/DataDog/datadog-agent/releases / Not released yet milanote - No newer release taskexplorer - https://objective-see.com/products/changelogs/TaskExplorer.txt / No newer release jcryptool - https://www.cryptool.org/en/jcryptool/143-news/407-release-candidate-8-en / RC gog-galaxy- No newer release media-center - No newer release cloudtv - No newer release teamviewer - @core-code I think this should be removed from our list electron - v3.0.11 => v2.0.15 ??
sorry at the moment i have no way to postprocess or filter the sparkle or github files. will be looking to add that
@core-code Sounds good. Let me know if you need any help with that.
Just to note, I had a lot more returned casks today than normal. The ones with 'No Newer Release' I've manually verified to have no newer release on the appcast
. There might be a newer version, but as you know, sometimes I just don't have time to look hard enough... ๐ณ
any help with that.
i think we could just reuse the "postprocessCaskRepairOutput.py" and adapt it slightly. i will do that when i have time, but if you beat me to it .. ;)
I had a lot more returned casks today than normal
well, i think we kind of need to treat the 3 checks differently. the 'database' check has been going for about 2 months now, so it its working quite well i believe.
the appcast check should be quite good since in theory we don't need to verify that those are real releases. anything that occurs in a valid appcast is a real release by definition. i had no time yet to veriy that this check really works as well as expected though.
the github check is also new and i think a bit problematic because many projects seem to tag beta releases as the "latest release". this is the one where we should apply some postprocessing to filter releases that are not official. from a quick look this should be quite easy with regular expressions. but we need to monitor it a bit.
just reuse the "postprocessCaskRepairOutput.py" and adapt it slightly.
This seems like a really decent idea. I'll look into this weekend. โ
the appcast check should be quite good
I can confirm that there haven't been any real issues with these unless the URL changes. ๐
github check is also new and i think a bit problematic
I'd also add that some developers have the strangest habits of numbering things. I especially love it ๐ when a developer has a bunch of pre-releases and/or betas but labels them all as releases.
Also, this hasn't had as many problems over the last few days. So at least I'd say we're progressing in the right direction.
2018-12-14 cr-processed,crgithub and crsparklefeed are online - sorry @ran-dall for all the duplicates and already updated casks!
@suschizu It really wasn't that big of a deal today, fortunately. ๐
nextiva - https://www.nextiva.com/support/articles/nextiva-app-mac-setup.html - No newer version on appcast
skype - No newer version on appcast
wondershare-video-converter-ultimate - No newer version on website
sip - curl: (60) SSL certificate problem: self signed certificate
This issue is already too long for another person to come in as a good first issue
, so Iโm removing the label and closing. But please do use it to continue your conversation at will. Thatโs absolutely fine (and encouraged).
If in the future we need to create another of these issues, donโt hesitate to contact me.
Iโm going to unsubscribe from the issue, so @mention me if you need me to see something.
On second thought, Iโll leave it open a while longer, since the top post is still quite useful. See if anyone else is interested.
@vitorgalvao i agree this issue has gotten quite lengthy and probably not the easiest to jump into as a newcomer. also, @ran-dall is taking great care of all the required cask-repairs at the moment, so it doesn't look like we need additional help currently. i think we are making good progress with keeping everything up-to-date, especially with the new checks introduced a few days ago that check appcasts of sparkle-rss and github-atom kind although there have been a few setbacks with submitted beta versions.
initially i thought our output would be a good starting point for the community to jump in an keep the project up to date, which would suggest it would be best to keep this issue open. however, it seems like @ran-dall is fully on top of things especially with the help of automation, so keeping the issue open may not be necessary or even helpful. so, i defer to both of you on this question. if things change, we can always re-evaluate.
Then lets keep with the plan of closing, and if things change, we can always make a new one in the future. Thank you all!
2018-12-16 cr-processed,crgithub & crsparklefeed ready to check ๐จ
@suschizu Thank you! I'll process these in a few hours when I get back.โ
@core-code @vitorgalvao Sorry gentlemen for not being around yesterday. ๐คช My wife is making me go to the middle of nowhere for X-mas ๐ and I've been running around like a crazy person. I'll spend some time later looking over things in detail. I'll also start adding some specs for the Machine Learning / "AI" update bot (nothing yet is a 'real' AI, but for lack of a better word).
no problem - if you want us to do today's cask-repairs, just us know, we can always jump in ;)
@core-code Thanks for offering but I think I should be good today. However, on the 18th, I'll be travelling. โ๏ธ If you or @suschizu could take care of it that day. I would be very grateful. ๐ I loaded up the system into my MacBook Pro this morning with a friend of mine, so I could be able to pick it back up the 19th. I'm off to run today's work. โ
no problem, we will pick up tuesday
@suschizu @core-code Crazy day today! Anyway's here's the works.
appcast
appcast
Doesn't this count as the appcast
? https://www.cryptool.org/en/jct-downloads/stableThe details of my automation will come sometime in the next few days when I have time to write it out. Sorry for the delay. ๐ช
i'll have a look at datadog-agent.
taskexplorer this seems a beta, should be ignored jcryptool this seems to be a rc, should be ignored
i think we prefer 'real' appcasts, i.e. either sparkle RSS or github ATOM feeds, or whatever configuration file the app uses for its own a built-in update check. but if those aren't available, we can use whatever URL contains the latest version in some way, e.g. the download or release-notes website. so, you are right, the download site of jcryptool should be added as an appcast, i've done it now, thanks.
2018-12-17 cr-processed,crgithub & crsparklefeed are online ๐ผ
@ran-dall - no problem I'll check the lists in the next few days :) Have a nice holiday and a merry Christmas! โ๏ธ ๐ ๐ yes - the datadog-agent is special (https://github.com/Homebrew/homebrew-cask/pull/54739)
@suschizu Awesome! ๐ It's only tomorrow. But I'll keep the offer in mind because I might need your help on the 28th when I'm supposed to be travelling back. โ๏ธ
๐ I'll start to process the casks for today...
hello HBC community,
one of the 'problems' of Homebrew Cask is that it requires constant maintainance to keep working. any Cask with a version number needs to be updated every time a new version of this app is released.
for apps that have an sparkle/rss or github/atom feed from the vendor, keeping them up-to-date is relatively straightforward. @brianmorton has a great system implemented here. (his system incidentally also works for non-sparkle appcasts, but currently most of the Casks that have no sparkle, also have no appcast-stanza)
therefore i've been focusing my efforts on automatically finding outdated Casks on those without "sparkle appcast". every day i let a script run that compares the version of the Casks without appcast, to the database of app-versions that our company maintains. every day the script finds around 5 - 10 outdated Casks that should be updated e.g. using 'cask-repair'.
the design and implementation of the system is discussed here: https://github.com/Homebrew/homebrew-cask/issues/52837 and there was also a bit discussion about this here: https://github.com/Homebrew/homebrew-cask/issues/54811
so, long story short, basically it would be great if the community could have a look at the "cr-processed-DATE.txt" files i place here: http://macupdater.net/cask-repair/ each day and look through them to see which of those entries are real, official new releases of the apps and submit cask-repairs accordingly. i think this would also be a good task to get started with contributing to HBC.
for the time being, it would be best to note false positives (i.e. entries that are not real official releases of this app) here so i'll add them to the different filter-lists in the postprocessing script so that they won't appear in the next days output anymore.
UPDATE 11.dec.2018: ok, the cask-repair system for apps without appcast has been working quite well over the past month, i guess around 80% success rate (i.e. percentage of entries in the txt files that are actually valid new releases). but, it seems like there are quite a lot of casks WITH appcast that are outdated despite the appcast. so i've implemented two new checks to try to improve the situation there:
crsparklefeed-DATE.txt => this checks each cask with an valid sparkle appcast each day and generates an output if the appcast has a new 'newest' entry since yesterday. since sparkle feeds should only contain valid releases, 100% of those should be valid updates to be done via 'cask-repair' (if someone else hasn't beaten you to it).
crgithub-DATE.txt => this checks each cask that has its releases on github (as denoted via the URL or APPCAST stanza) and checks github if a new release has been posted since yesterday. if available, it tries to query the 'Latest Release' as denoted on github by the developer, so that beta releases are not picked up. if no 'Latest Release' is marked, it falls back to just using the newest release and downgrades from error to warning.