●● IRC: #boycottnovell @ FreeNode: Monday, February 22, 2021 ●● ● Feb 22 [03:55] Techrights-sec2 recent files are now hardlinked: [03:55] Techrights-sec2 $ stat --printf "%i\t%n\n" /home/gemini/gemini/tr_text_version/irc-log-techrights-210221.txt ~glr/tr_text_version/irc-log-techri [03:55] Techrights-sec2 ghts-210221.txt [03:55] Techrights-sec2 309272 /home/gemini/gemini/tr_text_version/irc-log-techrights-210221.txt [03:55] Techrights-sec2 309272 /home/glr/tr_text_version/irc-log-techrights-210221.txt [03:55] Techrights-sec2 but the old ones are not. perhaps the directory should be wiped and the [03:55] Techrights-sec2 full 'copy' be re-run manually? [03:55] schestowitz__ I noticed the same when I checked after the cron job had run [03:55] schestowitz__ On a positive note, the gemini links were all fine, got it right the first time around [03:56] schestowitz__ Shall we wipe tr-archives with care not to also delete the originals? I dread losing something in case hard links exist somewhere? [03:58] Techrights-sec2 ~gemini/gemini/tr_text_version/ [03:58] Techrights-sec2 has the redundant files so it if is wiped and then the copy script run, [03:58] Techrights-sec2 it will get populated by hardlinks. [03:58] schestowitz__ as long as that being done won't have any effect on the original files (depending on what was done prior to this, maybe tests included) [03:59] schestowitz__ the script is run with sudo by "pi", but I can run it manually ● Feb 22 [04:00] Techrights-sec2 ok [04:00] Techrights-sec2 IF run while files exist in the target directory, no changes are made [04:00] Techrights-sec2 the target has to be empty [04:01] schestowitz__ Yeah, I just want to be sure the emptying, however done, won't drain out anything ipfs uses as its files pool, about 1,000 files [04:03] Techrights-sec2 If ~gemini/gemini/tr_text_version/ is a copy, then nothing is lost. [04:03] Techrights-sec2 IF ~gemini/gemini/tr_text_version/ contains hardlinks, then only [04:03] Techrights-sec2 the directory entry gets removed, any other hardlinks to the original [04:03] Techrights-sec2 including the original, remain. [04:04] schestowitz__ do you have the privs to remove or to empty it safely? BTW, I think gemini is in sudoers already [04:10] Techrights-sec2 yes [04:10] Techrights-sec2 all clear [04:10] schestowitz__ I will rerun now.... then count files [04:11] schestowitz__ oh you beat me to it [04:11] Techrights-sec2 all copied [04:12] schestowitz__ Maybe I will try to better partition the page http://techrights.org/ipfs/ to avoid it getting so gigantic [04:12] -TechrightsBN/#boycottnovell-techrights.org | Techrights Full IPFS Index [04:13] schestowitz__ gemini@raspberrypi:~/gemini/not.tr_text_version $ ls -la | wc -l [04:13] schestowitz__ 997 [04:13] schestowitz__ gemini@raspberrypi:~/gemini/not.tr_text_version $ ls -la ../tr_text_version/ | wc -l [04:13] schestowitz__ 1006 [04:14] schestowitz__ regenrating the index to test linking to irc and bulletins [04:15] schestowitz__ the linking will save us hundreds of MBs over time, and seeing we have almost 4GB spare and almost all the capsule is now in place (I think 2020 misses some bits) we should be OK for at least a year [04:16] schestowitz__ ipfs does not scale well when the number of files grows and the operations (e.g. "add") grow linearly in duration of run [04:17] Techrights-sec2 yes there are months before there is an issue again regarding space [04:19] Techrights-sec2 is there a way to subdivide the IPFS collection so that the [04:19] Techrights-sec2 time and other resources it requires remain manageable? [04:19] schestowitz__ I asked some months ago and do not recall the exact answer, but it's not a dead end [04:22] schestowitz__ The linkage to the new linked objects seems to work correctly, I've tested about 20 semi-randomly [04:23] Techrights-sec2 I've put the 'copy' script in /usr/local/sbin/copy-tr-to-gemini.sh [04:23] Techrights-sec2 it can be called from cron from there [04:24] schestowitz__ diff says a comment line is the only difference to the original in the homedir for ipfs [04:26] schestowitz__ to avoid confusion of conflicting changes I've set pi crontab to point to the shared location and imported that file for backup purposes [04:26] Techrights-sec2 yes the clarification seems useful. [04:29] schestowitz__ Maybe in http://techrights.org/ipfs/ I should split into 5 subpages? [04:30] Techrights-sec2 At least two pages, perhaps make a new one each quarter? [04:35] schestowitz__ working on it now, should not take long, testing will take longer and over the long run... [04:36] Techrights-sec2 Is it the number of files that bother IPFS or the combined size? [04:36] schestowitz__ mostly size because it goes about scanning them for any changes, I believe (hashing) ● Feb 22 [05:24] schestowitz__ OK, http://techrights.org/ipfs/ seems to be properly split up now [05:25] schestowitz__ do we still need gemini/not.tr_text_version for anything or should I move it to /tmp as tentative for deletion (if that partition... maybe not even its own... is large enough?) [05:28] schestowitz__ OK, you did that already [05:41] schestowitz__
  • [05:41] schestowitz__
    Please do not put IP addresses into DNS MX records
    [05:41] schestowitz__
    [05:41] -TechrightsBN/#boycottnovell-blog.hboeck.de | Please do not put IP addresses into DNS MX records - Hanno's blog [05:41] schestowitz__

    I want to highlight a common misconfiguration in the DNS records for e-mail servers.

    [05:41] schestowitz__

    When a domain is configured to receive mails usually a DNS record of the type MX is configured pointing to the host name of the mail server.

    [05:41] schestowitz__

    Notably, according to the respective RFC 1035 the MX record must contain a domain name and may not directly point to an IP address. However some mail servers do configure an IP address. Many mail servers are lenient when it comes to this misconfiguration and will deliver mails nevertheless, so this may stay undetected.

    • [05:41] schestowitz__
  • [05:41] schestowitz__
    Update your computer!
    [05:41] -TechrightsBN/#boycottnovell-blog.linuxmint.com | Update your computer! The Linux Mint Blog [05:41] schestowitz__
    [05:41] schestowitz__

    Linux Mint comes with one of the best update managers available. Its very easy to use, its configurable and it shows a lot of information.

    [05:41] schestowitz__

    It handles security updates for all your software. All you need to do is use it.

    • ● Feb 22 [08:08] *rianne has quit (Quit: Konversation terminated!) ● Feb 22 [09:21] schestowitz__ http://ipkitten.blogspot.com/2021/02/highlights-from-new-epo-guidelines-for.html?showComment=1613940219904#c8653901889325739911 [09:21] -TechrightsBN/#boycottnovell-ipkitten.blogspot.com | Highlights from the new EPO guidelines for examination 2021: ViCo oral proceedings, description amendments and antibodies - The IPKat [09:21] schestowitz__ "I recall filing a divisional (with the UKIPO) directed to a clearly-described but unillustrated modification of embodiments of the parent. In the divisional as filed, all references to "embodiment in accordance with the invention" were replaced by "arrangement useful for understanding the invention", and after the description of the "arrangements", said something like "in accordance with the invention, the such-and-such is [09:21] schestowitz__ replaced by a so-and-so". That modification had been explicitly disclosed in the description of the parent, and this procedure seemed to be the only way to avoid the possible added matter objection that would have arisen if a new drawing illustrating the described but unillustrated modification, had been filed with the divisional. No objection was raised, and I could see no reason why it should have been." [09:25] schestowitz__ x https://www.bloomberg.com/news/articles/2021-02-21/bill-gates-says-social-media-bans-aren-t-answer-to-divisiveness?srnd=technology-vp [09:25] -TechrightsBN/#boycottnovell-www.bloomberg.com | Bloomberg - Are you a robot? [09:25] schestowitz__ # bill sez ● Feb 22 [10:09] Techrights-sec2 2007 is in place [10:09] Techrights-sec2 2006 is also in place [10:09] Techrights-sec2 I will redo 2021 later in the day, but for now 2006 through 2015 all have [10:09] Techrights-sec2 the latest conversion process. Let me know if you spot any major room [10:09] Techrights-sec2 for improvement. [10:09] Techrights-sec2 2021 is now improved [10:09] Techrights-sec2 there is a local tarball of the articles: gemini-pages.2006-2021.tar.gz [10:09] Techrights-sec2 2016 - 2020 still have the old style block quotes however. [10:10] schestowitz__ Excellent, maybe we can announce this later today. I keep struggling in recommending a client/browser as many are jailed in **ithub [10:11] Techrights-sec2 I have not explored the clients, amfora was the first one that worked. [10:11] Techrights-sec2 None are in any repositories for convenient (and safe) download and [10:11] Techrights-sec2 automated maintenance. [10:15] Techrights-sec2 I sent the start URL for 2006 to GUS today. Hopefully the old pages will [10:15] Techrights-sec2 get indexed. [10:15] Techrights-sec2 $ find /home/gemini/gemini/2* -mindepth 3 -type f -name '*.gmi' -print | wc -l [10:15] Techrights-sec2 32458 [10:15] Techrights-sec2 just over 32k articles [10:15] schestowitz__ Biggest capsule in a matter of less than a fortnight :-) [10:16] Techrights-sec2 yes, it took a week to write the conversion code (about 1 week at 1 FTE) [10:16] Techrights-sec2 there after it took a week of waiting for the downloads (about 2 hours at 1 FTE) [10:17] schestowitz__ Once it's done it's done as I very rarely change anything old (by rarely I mean almost never( [10:18] Techrights-sec2 the scripts are in our Git repository [10:18] Techrights-sec2 and mirrored on the RPi [10:18] Techrights-sec2 I figure the old articles will remain static, but the scripts are there [10:18] Techrights-sec2 in case anything needs updating. [10:18] Techrights-sec2 If the layout / structure changes, then the parser will need adjustment. [10:19] schestowitz__ Maybe they can be generalise to make a toolset of wordpress->gmi conversions. Can help 'recruit' many more sites for the space... [10:22] Techrights-sec2 They're too specific to be of much use outside of TR. However they [10:22] Techrights-sec2 can serve as exanples and the approximate workflow might be of use to many [10:22] Techrights-sec2 others. [10:22] Techrights-sec2 The wordpress part can contain too much variation, it is all custom HTML there. [10:22] Techrights-sec2 Fortunately you have been very consistent in use of HTML within the articles [10:22] Techrights-sec2 so it was possible to parse. The daily links needed their own subroutine, but [10:22] Techrights-sec2 everything else seems to fit into the same set of rules. [10:24] schestowitz__ Maybe I can do some blog posts explaining various aspects of the conversion of the code, if I can comprehend Perl well enough (I have not tried). This way people can search and find useful code files or at least code samples they can reuse. [10:24] Techrights-sec2 gemini-scripts-README.txt has the internal write up [10:24] Techrights-sec2 wget-tr-pages.sh can get a year at time of the back articles [10:25] schestowitz__ Most people's wordpress sites are vastly smaller, so we sort of stress-tested it 'at scale', I suppose... ● Feb 22 [11:12] Techrights-sec2 The scale worked in our favor and made it more worth it to script. [11:12] Techrights-sec2 The initial scripts took only maybe one day (at 1 FTE) but then tweaks [11:12] Techrights-sec2 etc and working with Git for the first time added to that. [11:12] Techrights-sec2 The scripts can be shown but our Git repository has not been put into [11:12] Techrights-sec2 the HTTP server yet. I have not read up on that yet, and wonder about [11:12] Techrights-sec2 a lot of the features. I was talking with my xxxxxxxxxxxxxxxxx about Git recently [11:12] Techrights-sec2 and got a lot of tips but now have to learn. [11:12] Techrights-sec2 As for the scripts, if people have been consistent in their document structure, [11:12] Techrights-sec2 then the XPath approach will work for them too. [11:14] schestowitz__ Thanks for all the hard work. I think we need to do what we can to give back to gemini and help it grow. Later on I'll examine the code to see if I personally can make sense of it, though I suspect publishing anything about it must be done after git goes public. vis a vis git, mind the ongoing TR series about github etc. [11:15] Techrights-sec2 No problem. Gemini is a worthy project. So contributing a decent capsule [11:15] Techrights-sec2 helps. I am writing up some notes for the Capsule. [11:16] schestowitz__ Maybe at a later point I will convert my personal blog (>2000 posts) to gemini as it is also wordpress and the domain is managed similarly. I can try it on the pi and use another server software to test searching. [11:26] schestowitz__ thought: if we contract the right people and the code is well documented, gemini core people and pages will point to our repo and increase use among wordpress users (there's one for hugo that I saw) [11:28] Techrights-sec2 We should have some internal review of the scripts first, to ensure that [11:28] Techrights-sec2 nothing obvious is wrong. [11:28] schestowitz__ MS-PL licence, definitely! [11:28] Techrights-sec2 Should the scripts be AGPL? [11:51] Techrights-sec2 gemini://gemini.techrights.org/2021-html-to-gemini-notes/ [11:51] Techrights-sec2 I suppose it is ready for internal review. Comments on where comments are needed in the code are important now, as would be com [11:51] Techrights-sec2 ments on what to change or redact. [11:51] Techrights-sec2 back in a bit ● Feb 22 [12:01] schestowitz__ I have just read the whole page and did not spot typos. It is written concisely and clearly. [12:05] schestowitz__ about 1000 requests so far this morning [12:06] schestowitz__ (though I cannot distinguish what types, I just check for communications in/out, over port 1965) ● Feb 22 [16:58] Techrights-sec2 https://git.sr.ht/~hsanjuan/gemini-ipfs-gateway [16:58] -TechrightsBN/#boycottnovell-git.sr.ht | ~hsanjuan/gemini-ipfs-gateway - sourcehut git [16:58] schestowitz__ interesting! [16:59] schestowitz__ so we could serve over ipfs what we already have in the "tr" directory anyway ● Feb 22 [17:00] schestowitz__ so we could serve over ipfs and gemini what we already have in the "tr" directory anyway (for both ipfs and gemini) [17:00] Techrights-sec2 yes, but would it be of use for TR? [17:01] Techrights-sec2 I think so. [17:01] Techrights-sec2 I'm not up on IPFS though [17:01] schestowitz__ They solve very different problems [17:03] schestowitz__ I struggle to think of a practical real-world scenario where you want to combine both [17:06] Techrights-sec2 The bulletins are already mirrored via the filesystem too. [17:07] schestowitz__ yes, and if both the pi dies and the server get seized or something, ipfs will still be able to serve a copy (not that such a scenario ought to ever arise) [17:08] schestowitz__ I think of it as deterrent (against SLAPP or takedown demands with deadline) [17:08] schestowitz__ It's easier for them when there's a third party like Google that buckles for its own business reasons, or even Twitter without incentive to fight for you [17:09] *schestowitz__ has quit (Quit: Konversation term) [17:09] -NickServ-schestowitz__!~schestowi@host81-154-169-167.range81-154.btcentralplus.com has just authenticated as you (schestowitz) [17:09] *schestowitz__ (~schestowi@unaffiliated/schestowitz) has joined #boycottnovell [17:09] *ChanServ gives channel operator status to schestowitz__ ● Feb 22 [19:27] Techrights-sec2 https://adambrown.info/p/wp_hooks/hook/%7B$new_status%7D_%7B$post-%3Epost_type%7D [19:27] Techrights-sec2 can that be used to have WordPress trigger updates in Gemini and IPFS? [19:27] -TechrightsBN/#boycottnovell-adambrown.info | {$new_status}_{$post->post_type} Wordpress hook details -- Adam Brown, BYU Political Science [19:28] schestowitz__ I am not sure; wasting lots of time fighting off a major ddos attack on tm at the moment :( [19:33] Techrights-sec2 I notice that there have been TM outages [19:33] Techrights-sec2 Is there anything that can be done upstream to mitigate the attacks? [19:33] schestowitz__ not serve CSS files, but then the attack pattern would just shift [19:34] Techrights-sec2 Can a Vanish cache be placed in front? [19:35] schestowitz__ that might help only to some degree, depending on patterns. 20k reqs per 30 sec is still a lot