Re: Namespaces for Lintian Tags
Hi, On Wed, 20 Nov 2019, Felix Lechner wrote: > There are many motivations: Among those motivations, which one is the one that triggered this process and which one are there as "additional benefits" that you could identify to justify the change? > 1. Shortens tag names. I don't see that as a benefit, we copy/paste the tags into overrides or full lines into lintian-info. We rarely need to type them. > 2. Points to the code that issued the tag. "grep -r" on the codebase has been working well for me. This mapping is only really needed when you want to dig into the code anyway. > 3. Frees up name space (good tags are rare). Can you show examples of how this would help you concretely? I have a hard time seeing how difficult it can be to invent a new name for a new tag. > 4. Multiple checks can use the same tag in different contexts (i.e. > 'spelling'). spelling-error-in-binary spelling-error-in-description spelling-error-in-changelog etc is perfectly fine. > 5. Preempts name conflicts in case some check-writing is delegated to > expert teams. This is not a real problem, that's the kind of pseudo-benefit that you try to imagine to justify the change that you want (I have done that many times ;-)). > 6. Quicker to split large checks when components reuse tag names. I don't follow you. For me splitting checks means they get renamed and thus your tag names are renamed too. > 7. Brings consistency between Lintian and custom profile users, such > pkg-perl-tools and pkg-js-tools, who already have private namespaces. A very minor benefit IMO. Now can you do the same analysis with the disadvantages? Since the check is embedded in the tag name: * It's harder to move a tag from one check to another. * It's harder to rename a check. What else can you come up with? > The change is technically easy. (Lintian even has a way to track > renamed tags for overrides.) On an optical level, however, the change I was about to ask for overrides, but this would be a massive usage of this rename feature and it will confuse many persons. People will start to use the new name in overrides to avoid this confusion and it won't work with old lintians (not a big deal but still). > will affect a lot of people. It could even cause headaches for some > users, for example in derivatives. We would like to solicit your > input. What kind of headaches are you referring to? Cheers, -- Raphaël Hertzog ◈ Debian Developer Support Debian LTS: https://www.freexian.com/services/debian-lts.html Learn to master Debian: https://debian-handbook.info/get/
Re: Namespaces for Lintian Tags
On 21.11.19 10:55, Wouter Verhelst wrote: >> 1. Shortens tag names. > I do not see how that is a good thing. > > "debian-copyright-file-uses-obsolete-national-encoding" is a sentence, which > inherently makes it extremely human-readable. For a tool that is > primarily meant to validate the output of a human, this is an immensely > useful feature. to be honest, i already find myself resorting to lintian-info for an explanation of what these supposedly "extremely human-readable sentences" actually mean (and what these meanings imply) far too often. so i probably would be just as happy with new, more terse but namespaced tags. >> 3. Frees up name space (good tags are rare). i guess this is the problem, and everything that helps with it has my +1. just my 1¢. fgamsdr IOhannes
Re: Namespaces for Lintian Tags
On Wed, Nov 20, 2019 at 09:55:04AM -0800, Felix Lechner wrote: [...] > For example, the tag > 'debian-copyright-file-uses-obsolete-national-encoding' might become > 'national-encoding@debian/copyright'. > > There are many motivations: > > 1. Shortens tag names. I do not see how that is a good thing. "debian-copyright-file-uses-obsolete-national-encoding" is a sentence, which inherently makes it extremely human-readable. For a tool that is primarily meant to validate the output of a human, this is an immensely useful feature. "national-encoding@debian/copyright" is not very useful in that respect, and is a huge step backwards IMO. > 2. Points to the code that issued the tag. While I can see how that might be useful from a lintian maintainer's point of view, this is not really of benefit to the majority of users of lintian. > 3. Frees up name space (good tags are rare). This of course *does* make sense. > 4. Multiple checks can use the same tag in different contexts (i.e. > 'spelling'). Don't see that as a benefit, tbh. > 5. Preempts name conflicts in case some check-writing is delegated to > expert teams. > 6. Quicker to split large checks when components reuse tag names. > 7. Brings consistency between Lintian and custom profile users, such > pkg-perl-tools and pkg-js-tools, who already have private namespaces. These make sense, I guess. I guess namespacing things might have some limited benefit, I suppose. But I like that a plain "lintian" run provides enough context for most people to understand what's going on, without having to use lintian-info. Same for lintian overrides. Please don't break that. HTH, [...] -- To the thief who stole my anti-depressants: I hope you're happy -- seen somewhere on the Internet on a photo of a billboard
Re: Namespaces for Lintian Tags
On 11/20/19 3:01 PM, Sean Whitton wrote: > For example, I would request obsolete-national-encoding@debian/copyright > rather than national-encoding@debian/copyright. Or perhaps obsolete-encoding@? -- Richard signature.asc Description: OpenPGP digital signature
Re: Namespaces for Lintian Tags
Hello Felix, Thanks to you and others for working on this. On Wed 20 Nov 2019 at 09:55AM -08, Felix Lechner wrote: > For example, the tag > 'debian-copyright-file-uses-obsolete-national-encoding' might become > 'national-encoding@debian/copyright'. > > [...] > > The change is technically easy. (Lintian even has a way to track > renamed tags for overrides.) On an optical level, however, the change > will affect a lot of people. It could even cause headaches for some > users, for example in derivatives. We would like to solicit your > input. I hope that the change to the tag names could be done in such a way that the tag names do not become less descriptive. Right now it is nice to be able to read a tag name, and often know what it means, without having to look at the long text for the tag. I hope that can remain true. For example, I would request obsolete-national-encoding@debian/copyright rather than national-encoding@debian/copyright. -- Sean Whitton signature.asc Description: PGP signature
Re: Namespaces for Lintian Tags
Quoting Felix Lechner (2019-11-20 18:55:04) > I am about to introduce private namespaces for tags. Many tags already > point to the check that issued them. (For those who don't know, a > 'check' in Lintian is a module that issues a tag.) Namespaces would > formalize the relationship. > > For example, the tag > 'debian-copyright-file-uses-obsolete-national-encoding' might become > 'national-encoding@debian/copyright'. [...] > Please provide your thoughts, perspectives and suggestions by amending > the bug report at #943525. Thank you! Great! I very much welcome that change! - Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private signature.asc Description: signature