Bug#915816: RM: uptimed -- ROM; long time bugs non-fixed, better alternatives
Hi, gustavo panizzo wrote: > Alex, any update on this? I just looked through the open bug reports. The _only_ non-wishlist bug report open is the one about duplicated bug reports by myself. So the plural in "long-time bugs non-fixed" is IMHO exaggeration. So I'm slightly fail to see the reason to remove it. So far I also doubt that tuptime is superior in every way. > Is there anything I can help with to move this forward? I'm not keen at all to take over yet another package, but I fear there's no other way. Regards, Axel -- ,''`. | Axel Beckert , https://people.debian.org/~abe/ : :' : | Debian Developer, ftp.ch.debian.org Admin `. `' | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5 `-| 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
Bug#915816: RM: uptimed -- ROM; long time bugs non-fixed, better alternatives
Hello Alex, any update on this? Is there anything I can help with to move this forward? -- IRC: gfa GPG: 0X44BB1BA79F6C6333
Bug#915816: RM: uptimed -- ROM; long time bugs non-fixed, better alternatives
On December 8, 2018 9:27:22 AM UTC, gustavo panizzo wrote: > >On Sat, Dec 08, 2018 at 04:03:05AM +, Scott Kitterman wrote: > >> >>It seems like it's not for sure this should be removed right now. >Axel: are you willing to take over as maintainer? I see there is a new >upstream release. >> > >This is upstream changelog since the last snapshot of the repo I >packaged for Debian until today. > >$ git log v0.4.0+git20150923.6b22106..v0.4.1 --no-merges >--pretty=oneline > >f895c8fef774952af8d435b10b0dded5e648aec9 (HEAD, tag: v0.4.1, >upstream/master) version bump to 0.4.1 > >7beceef840d8ba8836c04fa410fb91af63972d3f readme markup improvements > >777d6fc65251734d6b63706bb60d0885f8a62430 Use utmpx instead of utmp for >Solaris >80a09909c73a7ca2f3e6507d77c09a7cd5e1db5f markdown formatting > >2fd3dcdf6916e0454315b90daff8e41317481f3f typo fix > >5ecddd498484b4a837c97005d48d121057f9ac02 readme formatting updates > >4a5b397cf0a97965f531701b576864a69072bbf1 readme formatting updates > >d143b069ecd16b1a89bdb880ef016e9521de50d8 got rid of hgignore > >5a3c60f4132fa1568460da43f55b66a19e5ef213 typo fix > >a9d8b40b944bffaba2b6e089aa765faa0bed9588 whitespace formatting > > >This is why I opened an ROM bug instead of a RFH bug. I think our users >are better served with tuptime than keep bumping the d/control and >Standards-Version in uptimed. I'm not arguing with you about what's best. I just want to avoid removing something and then having it show up again shortly after in New. Scott K
Bug#915816: RM: uptimed -- ROM; long time bugs non-fixed, better alternatives
Control: tag -1 + moreinfo Hi Scott, Scott Kitterman wrote: > It seems like it's not for sure this should be removed right now. > Axel: are you willing to take over as maintainer? I'll have a look. In general, the things gustavo wrote, are correct. > I see there is a new upstream release. Indeed, from Jun 2018, but sole changelog entry is "updates for modern solaris", so not approaching any of the mentioned issues. > Perhaps this bug should be put on hold for now? For the moment, yes. > If you agree, please add a moreinfo tag. Done. Thanks! Regards, Axel -- ,''`. | Axel Beckert , https://people.debian.org/~abe/ : :' : | Debian Developer, ftp.ch.debian.org Admin `. `' | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5 `-| 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
Bug#915816: RM: uptimed -- ROM; long time bugs non-fixed, better alternatives
On Sat, Dec 08, 2018 at 04:03:05AM +, Scott Kitterman wrote: It seems like it's not for sure this should be removed right now. Axel: are you willing to take over as maintainer? I see there is a new upstream release. This is upstream changelog since the last snapshot of the repo I packaged for Debian until today. $ git log v0.4.0+git20150923.6b22106..v0.4.1 --no-merges --pretty=oneline f895c8fef774952af8d435b10b0dded5e648aec9 (HEAD, tag: v0.4.1, upstream/master) version bump to 0.4.1 7beceef840d8ba8836c04fa410fb91af63972d3f readme markup improvements 777d6fc65251734d6b63706bb60d0885f8a62430 Use utmpx instead of utmp for Solaris 80a09909c73a7ca2f3e6507d77c09a7cd5e1db5f markdown formatting 2fd3dcdf6916e0454315b90daff8e41317481f3f typo fix 5ecddd498484b4a837c97005d48d121057f9ac02 readme formatting updates 4a5b397cf0a97965f531701b576864a69072bbf1 readme formatting updates d143b069ecd16b1a89bdb880ef016e9521de50d8 got rid of hgignore 5a3c60f4132fa1568460da43f55b66a19e5ef213 typo fix a9d8b40b944bffaba2b6e089aa765faa0bed9588 whitespace formatting This is why I opened an ROM bug instead of a RFH bug. I think our users are better served with tuptime than keep bumping the d/control and Standards-Version in uptimed. -- IRC: gfa GPG: 0X44BB1BA79F6C6333
Bug#915816: RM: uptimed -- ROM; long time bugs non-fixed, better alternatives
On Sat, 8 Dec 2018 08:12:54 +0800 gustavo panizzo wrote: > On Fri, Dec 07, 2018 at 11:06:01AM +0100, Axel Beckert wrote: > >Hi, > > >> leap second [2], > > > >Never noticed that despite having uptimed installed on more or less > >all my Debian systems for more like 15 years or so. You're breaking my > >tradition!!1!eleventy! ;-) > > > > I understand how you feel, I've used uptimed on all my systems for the > past 10 years. > > In the usecase were uptimes shines(servers) there is almost no reason to > no use tuptime [1] > > On boards, were usually run from SD, you want to minimize disk > writes, uptimed won't let you remount the filesystem as RO. > With tuptime you'll need to hack a bit to remount RW every now and > then, or run it just before shutting down, but is more manageable. > > > On laptops uptimed is buggy, tuptime is not. > > If you wanna build something on top of the uptime data, tuptime gives > you a sqlite db which is better that parse /var/spool/uptime/records [2] > > If you wanna "extend" tuptime you just need yo hack the Python script. > > I really don't see a situation where uptimed is better than tuptime. > > > >> and general bugginess that hasn't been solved in years [3] > > > >That one I noticed. But as the severity states, it's minor (but > >admittedly annoying). > > > >> Ricardo (tuptime upstream and Debian maintainer) has agreed to add a > >> note to tuptime's long description about uptimed (for people doing > >> `apt-cache search uptimed`) and a migration script [4] > > > >Actually for some reason, I thought the import of the uptimed database > >is done by default if present. I probably misread that addition to the > >package description. > > > >> I'll submit a patch to Buster release notes regarding the removal. > > > >Maybe there should be a transitional package from uptimed to tuptime > >which actually calls that migration script. (For that it needs to be > >packaged uncompressed and not under /usr/share/doc/. Will file bug > >reports for these two things later today.) > > > > No, I don't think tuptime should replace uptimed, their CLI is > different, tuptime pulls Python. > > tuptime could run the import script on postinst (I'd provide a patch if > Ricardo agrees with the idea and implementation), but it should not > Breaks/Conflicts/Replaces uptimed in case > someone wants to keep using uptimed after upgrading to Buster. [3] It seems like it's not for sure this should be removed right now. Axel: are you willing to take over as maintainer? I see there is a new upstream release. Perhaps this bug should be put on hold for now? If you agree, please add a moreinfo tag. That takes it out of the regular processing que. Scott K
Bug#915816: RM: uptimed -- ROM; long time bugs non-fixed, better alternatives
On Fri, Dec 07, 2018 at 11:06:01AM +0100, Axel Beckert wrote: Hi, leap second [2], Never noticed that despite having uptimed installed on more or less all my Debian systems for more like 15 years or so. You're breaking my tradition!!1!eleventy! ;-) I understand how you feel, I've used uptimed on all my systems for the past 10 years. In the usecase were uptimes shines(servers) there is almost no reason to no use tuptime [1] On boards, were usually run from SD, you want to minimize disk writes, uptimed won't let you remount the filesystem as RO. With tuptime you'll need to hack a bit to remount RW every now and then, or run it just before shutting down, but is more manageable. On laptops uptimed is buggy, tuptime is not. If you wanna build something on top of the uptime data, tuptime gives you a sqlite db which is better that parse /var/spool/uptime/records [2] If you wanna "extend" tuptime you just need yo hack the Python script. I really don't see a situation where uptimed is better than tuptime. and general bugginess that hasn't been solved in years [3] That one I noticed. But as the severity states, it's minor (but admittedly annoying). Ricardo (tuptime upstream and Debian maintainer) has agreed to add a note to tuptime's long description about uptimed (for people doing `apt-cache search uptimed`) and a migration script [4] Actually for some reason, I thought the import of the uptimed database is done by default if present. I probably misread that addition to the package description. I'll submit a patch to Buster release notes regarding the removal. Maybe there should be a transitional package from uptimed to tuptime which actually calls that migration script. (For that it needs to be packaged uncompressed and not under /usr/share/doc/. Will file bug reports for these two things later today.) No, I don't think tuptime should replace uptimed, their CLI is different, tuptime pulls Python. tuptime could run the import script on postinst (I'd provide a patch if Ricardo agrees with the idea and implementation), but it should not Breaks/Conflicts/Replaces uptimed in case someone wants to keep using uptimed after upgrading to Buster. [3] [1] - tuptime pulls Python as a dep, I don't remeber any server without python installed in the last years (python apps, or fail2ban, or ansible, or saltstack or monitoring scripts, and the list goes on) some people may not want to have python installed [2] - which may have duplicate data anyway [3] - uptimed is composed by a binary, a daemon and a library, it only links against glibc so an old deb should keep working for a few releases -- IRC: gfa GPG: 0X44BB1BA79F6C6333
Bug#915816: RM: uptimed -- ROM; long time bugs non-fixed, better alternatives
Hi, thanks for taking me in Cc. gustavo panizzo wrote: > I want to request the removal of the uptimed package *sigh* > package works on long running servers, but breaks with non-rtc systems Which are common (Raspberry Pi, Alix, etc.), but far from being the majority. > leap second [2], Never noticed that despite having uptimed installed on more or less all my Debian systems for more like 15 years or so. You're breaking my tradition!!1!eleventy! ;-) > and general bugginess that hasn't been solved in years [3] That one I noticed. But as the severity states, it's minor (but admittedly annoying). > Ricardo (tuptime upstream and Debian maintainer) has agreed to add a > note to tuptime's long description about uptimed (for people doing > `apt-cache search uptimed`) and a migration script [4] Actually for some reason, I thought the import of the uptimed database is done by default if present. I probably misread that addition to the package description. > I'll submit a patch to Buster release notes regarding the removal. Maybe there should be a transitional package from uptimed to tuptime which actually calls that migration script. (For that it needs to be packaged uncompressed and not under /usr/share/doc/. Will file bug reports for these two things later today.) P.S. to Ricardo: I saw your reply on #914954, will reply soonish. Regards, Axel -- ,''`. | Axel Beckert , https://people.debian.org/~abe/ : :' : | Debian Developer, ftp.ch.debian.org Admin `. `' | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5 `-| 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
Bug#915816: RM: uptimed -- ROM; long time bugs non-fixed, better alternatives
Package: ftp.debian.org Severity: normal Hello I want to request the removal of the uptimed package package works on long running servers, but breaks with non-rtc systems [1], leap second [2], and general bugginess that hasn't been solved in years [3] Some time ago I discovered tuptime which has none of this bugs, a responsive upstream and a modern codebase. Ricardo (tuptime upstream and Debian maintainer) has agreed to add a note to tuptime's long description about uptimed (for people doing `apt-cache search uptimed`) and a migration script [4] I'll submit a patch to Buster release notes regarding the removal. thanks [1] https://github.com/rpodgorny/uptimed/issues/1 [2] https://github.com/rpodgorny/uptimed/issues/11 [3] #654830 6 years old https://github.com/rpodgorny/uptimed/issues/2 4 y.o https://bugs.launchpad.net/ubuntu/+source/uptimed/+bug/498439 9 y.o [4] https://metadata.ftp-master.debian.org/changelogs/main/t/tuptime/tuptime_3.4.1_changelog
