[07:33] good morning everyone [07:34] RikMills: not a big drama, bu there's an extra test changelog entry in you last kglobalaccel (fw5) upload [07:35] santa_: will look in a few mins [07:40] santa_: fix uploaded and git pushed [07:45] RikMills: nice! thank you [09:12] santa_: I found an issue with ka-update-metadata. If I #comment #out a build dep in debian/control, ka-update-metadata appears to stop parsing or recognising build deps from then on in the file [09:13] RikMills: I know, but that's a well known limitation of the python-debian library, in fact I had to remove your comments from purpose when packaging 6.4 [09:14] RikMills: given that it's a limitation from python-debian, comments in control files also trigger wrap-and-sort malfunctioning [09:14] santa_: ok, know I know then. thanks [09:14] I'm really sorry about that [09:14] *now [09:14] I whish I had a solution [09:14] don't be. it happens and I learned something [09:15] furthermore, there's a snitize-debian-control program in KA, precisely to handle that [09:16] had not spotted that [09:16] there's also another tricky thing: consider a control file with a couple of paragraphs separated by a blank line [09:17] imagine you put one one various spaces in that blank line [09:18] that, at least back in the days would make wrap-and-sort and some KA programs go crazy [09:18] urgh [09:18] e.g. like wrap-and-sort removing one of the paragraphs [09:19] I remember we hit both issues when the Plasma 4 -> Plasmna 5 upgrade happened here [09:20] I came in at the tail end of a lot of that, so likely missed it [09:21] I guess so, that sanitize-control-file programs must be one of the most ancient of the current programs of KA [09:22] * RikMills goes to specsavers [10:38] morning [11:37] sgmoore: hi [11:42] sgmoore: hi (x2) [11:42] hi! [11:44] RikMills: the thing you "workarounded" in purpose was missing from git, last time I checked :) [11:49] Hi all [12:30] santa_: have pushed the workarounded thing in git [12:31] may change it at some point so the workaround is in the KF5 package instead [12:31] which long term may be better [12:34] RikMills: yes, perhabps we can add a separate package for the icons in kf6-purpose, then add a depend in the fw 5 counterpart [12:38] perhaps. putting icons in a -bin package seems odd to start with [13:07] I don't understand how to do falkon. gbp-archive fetches the tar, but doesn't remove the files-excluded and so of course the patch fails. I have tried putting my own tar brought in by uscan —download-current-version that properly removes the files, but it is overwritten bad tar. @RikMills how did you successfully upload this darn thing? [13:08] Is that in ocular or noble? (re @sgmoore: I don't understand how to do falkon. gbp-archive fetches the tar, but doesn't remove the files-excluded and so of course the patch fails. I have tried putting my own tar brought in by uscan —download-current-version that properly removes the files, but it is overwritten bad tar. @RikMills how did you successfully upload this darn thing?) [13:08] ocular [13:08] oracular too [13:08] what they did with falkon is something terrible [13:08] I agree [13:08] both in debian and ubuntu [13:09] and worst of it, I bthered to fix it and the someone who tought he knows better synced the thing from debian [13:09] * bothered [13:09] santa_: yeah, I saw that :? [13:10] sgmoore: if you have other things to do in the meantime please go ahead with them, I will have another look into it soon ;) [13:11] I think the upstream tarball repacking is not actually needed, but I need to check better [13:11] furthermore, they did the repacking in debian without adding a repack suffix [13:11] what a disaster [13:12] I did pull-lp-source and made changes in that and uploaded. than added those changes to git and tagged (re @sgmoore: I don't understand how to do falkon. gbp-archive fetches the tar, but doesn't remove the files-excluded and so of course the patch fails. I have tried putting my own tar brought in by uscan —download-current-version that properly removes the files, but it is overwritten bad tar. @RikMills how did you successfully uploa [13:13] I did pull-lp-source and made changes in that and uploaded. then added those changes to git and tagged (re @sgmoore: I don't understand how to do falkon. gbp-archive fetches the tar, but doesn't remove the files-excluded and so of course the patch fails. I have tried putting my own tar brought in by uscan —download-current-version that properly removes the files, but it is overwritten bad tar. @RikMills how did you successfully uploa [13:13] Ok, I will try this method, thank you (re @RikMills: I did pull-lp-source and made changes in that and uploaded. then added those changes to git and tagged) [13:13] and even worse: we don't have repacking support in KA yet because I have been busy with the fw6 + Attic dir support :) [13:15] santa_: was it a repack? at one point they were using the tarball that kde or github (I forget which) automatically makes for a git tag [13:15] *kde invents gitlab [13:15] RikMills: yes, and also a repack if I'm not mistaken [13:16] urgh^2 [13:18] I was doing this before that person synced: https://launchpad.net/ubuntu/+source/falkon/23.08.0-1ubuntu1 [13:23] santa_: there is a new upstream 24.05.2 falkon in gear, so we could do that with the correct tarball ;) [13:23] oh yes, please [13:24] in any case the falkon case it's interesting for whever I get into KA repack support [13:25] previously I was checking if the tarball was repacked or not by checking the watch file [13:25] * RikMills hands santa_ a ghostwriter [13:25] now I know that's not good enough :) [13:43] hmm. falkon is qt6. guess that won't work on my jammy laptop. putting aside for now. [13:46] Your change is in git? If so, I could add that to a new uptream release upload that uses the correct tarbal? [13:46] yeah, let me set to unreleaed and delete git tag [14:57] sgmoore: santa_: https://launchpad.net/ubuntu/+source/falkon/24.05.2-0ubuntu1 [15:02] ah, thanks a lot [15:03] reading the changelog, now I understand what I did [15:04] still, someone could argue that the tarball repacking is needed [15:05] yeah... skipping ghostwrite for now as is a repack [15:05] ghostwriter too [15:05] but personally, once those js files come from source IMHO we could skip the repacking [15:06] that's what I don't understand. They are in source, we delete them, then patch them back in? It makes zero sense. [15:11] it used to make sense [15:12] @sgmoore I was also naughty and git a force push to git to clean things up, so you should delete you falkon clone and reclone if you want to have one still locally [15:12] let me have a look so I can explain [15:12] @sgmoore I was also naughty and did a force push to git to clean things up, so you should delete you falkon clone and reclone if you want to have one still locally [15:12] ok (re @RikMills: @sgmoore I was also naughty and did a force push to git to clean things up, so you should delete you falkon clone and reclone if you want to have one still locally) [15:17] alright, so here it goes the explanation about the js files in falkon: [15:18] these days falkon orig source tarball provide 4 js files in src/lib/data/html/ [15:19] - jquery-ui.js and jquery-ui-1.12.1.js [15:20] - jquery.js and jquery-1.12.4.js [15:21] in both pairs the second is the source code of the first [15:21] however, some time ago the falkon orig tarball provided only: [15:21] - jquery-ui.js [15:22] - jquery.js [15:22] i.e. only the minified js without the source code [15:23] so what they did is removing the minified files with Files-Excluded and add the non-minified files with the patch [15:24] so that thing fits better with the DFSG [15:25] but now, we have the source js files in the code, therefore, in my opinion, it's no longer needed to do that [15:25] @sgmoore makes sense now? [15:25] yeah [15:25] furthermore: [15:26] now that I had a look at the debian/rules file, there's a few lines that we also should probably remove [15:28] lets the apparmor fix migrate, then we can deal with any other changes [15:28] *let [15:28] * RikMills is very tempted to add an epoch to prevent any more syncing ;) [15:28] but I resist [15:29] lol [15:33] hahah [15:34] in any case, when they did that they should have used a +dfsg prefix in the tarball [15:34] and put the thing in the watch file [15:34] yup. also, why do they use github?! [15:35] but what do you expect if they put a github link? [15:35] falkon is in release service.. [15:35] sgmoore: here is my wild guess: [15:36] the guy looked in download.kde.org, the he found that tarballs were separate in stable and unstable [15:36] Why they have an upstream signing key in the packaging when they don't use the signed tars is bizarre [15:37] so to have the pre-release versions he put the github thing, which covers both stable and unstable [15:37] santa_: but you can have multiple watch line IIRC? [15:37] but if you want to do that you can use a regex [15:38] or put 2 lines [15:38] grumbles [15:38] maybe the guy doesn't know [15:38] maybe he doesn't know about download.kde.org [15:38] who knows? [15:38] I know he ignored the email I sent [15:39] in any case, what do you expect from a linux distribution where the following happens to me: [15:39] 1. I type man gcc [15:40] 2. I find out theres no manpage [15:40] 3. I try to install gcc-doc [15:40] 4. wtf I can't find gcc-doc in the repos, what is going on? [15:41] 5. I find out that gcc-doc is in the non-free repos. WTF THE GCC DOCS IN NON-FREE REALLY!? [15:42] hahaha [15:43] 6. then I learn about this https://www.debian.org/vote/2006/vote_001 [15:44] and that GR was vote by a majority of the people with the DD nobiliary title [15:44] :facepalm [15:44] go figure [15:44] * was voted [15:45] omg [16:08] be back in an hour or so [16:09] ack, good luck in whatever you are going to do [18:08] ok, so... [18:08] we have to post-pone the brefing about the fw 6.4 changes [18:09] today, a new version of system arrived from -proposed to the normal release breaking all my autopkgtests containers [18:09] so I had to do a workarounding session [18:10] hopefully everything is fine now, and I'm rebuilding packages to test [18:10] * RikMills adds workarounding to local dictionary [18:10] but that's a derivative of the original one: workarounded [18:11] spellcheckers don't know that [18:11] I still have some brain fog today, even after having a night of normal sleep [18:49] That moment when I get an email that something broke, I have no idea what it is, and finally trace it to here. 😆 [18:57] Eickmeyer: heh, you got a reprepro build failure email too didn't you? :P [18:57] Yuuuuuuup! 😂 [18:59] I was like "Tritemio Maintainers... reprepro... BIONIC? What on earth did Launchpad eat this morning?" [18:59] "oh wait... ok nvm" [19:07] lol