From: "SZEDER Gábor" <szeder.dev@gmail.com>
To: Jeff King <peff@peff.net>
Cc: "Junio C Hamano" <gitster@pobox.com>,
"Nguyễn Thái Ngọc Duy" <pclouds@gmail.com>,
"Leho Kraav" <leho@conversionready.com>,
git@vger.kernel.org
Subject: Re: [PATCHv2 0/7] Fix and generalize version sort reordering
Date: Tue, 20 Dec 2016 09:50:42 +0100 [thread overview]
Message-ID: <CAM0VKjmDDKgYCvtbwpx=GcwRENzvSDLW_Xhia3btdeMjtAjAvg@mail.gmail.com> (raw)
In-Reply-To: <20161214170852.bzh5pyl4bov6rwbt@sigill.intra.peff.net>
On Wed, Dec 14, 2016 at 6:08 PM, Jeff King <peff@peff.net> wrote:
> On Thu, Dec 08, 2016 at 03:23:54PM +0100, SZEDER Gábor wrote:
>
>> > With my patches it looks like this:
>> >
>> > $ git -c versionsort.prereleasesuffix=-pre \
>> > -c versionsort.prereleasesuffix=-prerelease \
>> > tag -l --sort=version:refname
>> > v1.0.0-prerelease1
>> > v1.0.0-pre1
>> > v1.0.0-pre2
>> > v1.0.0
>>
>> And when there happen to be more than one matching suffixes starting
>> at the same earliest position, then we should pick the longest of
>> them. The new patch 6/7 implements this behavior.
>
> The whole approach taken by the suffix code (before your patches) leaves
> me with the nagging feeling that the comparison is not always going to
> be transitive (i.e., that "a < b && b < c" does not always imply "a <
> c"), which is going to cause nonsensical sorting results.
>
> And that may be part of the issue your 6/7 fixes.
>
> I spent some time playing with this the other day, though, and couldn't
> come up with a specific example that fails the condition above.
>
> It just seems like the whole thing would conceptually easier if we
> pre-parsed the versions into a sequence of elements, then the comparison
> between any two elements would just walk that sequence. The benefit
> there is that you can implement whatever rules you like for the parsing
> (like "prefer longer suffixes to shorter"), but you know the comparison
> will always be consistent.
I considered parsing tagnames into prefix, version number and suffix,
and then work from that, but decided against it.
versioncmp() is taken from glibc, so I assume that it's thoroughly
tested, even in corner cases (e.g. multiple leading zeros).
Furthermore, I think it's a good thing that by default (i.e. without
suffix reordering) our version sort orders the same way as glibc's
version sort does. Introducing a different algorithm would risk bugs
in the more subtle cases.
Then there are all the weird release suffixes out there, and I didn't
want to decide on a policy for splitting them sanely; don't know
whether there exist any universal rules for this splitting at
all. E.g. one of the packages here has the following version (let's
ignore the fact that because of the '~' this is an invalid refname in
git):
1.1.0~rc1-2ubuntu7-1linuxmint1
Now, it's clear that the version number is "1.1.0", and the user
should configure the suffix "~rc" for prerelease reordering. But what
about the rest? How should we split it "into a sequence of elements",
is it { "1.1.0", "~rc1", "-2ubuntu7", "-1linuxmint1" } or { "1.1.0",
"~rc1-2", "ubuntu7-1", "linuxmint1" }?
What if there is a hard-working developer who is involved in a lot of
Debian derivatives (and derivatives of derivatives...), and, for
whatever reason, wants to put derivative-specific versions in a
particular order? With my series, or conceptually even with master if
it weren't buggy, it's possible to specify the order of suffixes of
suffixes, and that dev could do this:
$ git -c versionsort.suffix=-rc
-c versionsort.suffix=linuxmint
-c versionsort.suffix=YADoaDD
tag -l --sort=version:refname
'1.1.0*'
1.1.0-rc1-2ubuntu7-1linuxmint1
1.1.0-rc1-2ubuntu7-1YADoaDD2
1.1.0
1.1.0-2ubuntu7-1linuxmint1
1.1.0-2ubuntu7-1YADoaDD2
and would get Linux Mint-specific tags before "Yet Another Derivative
of a Debian Derivative"-specific ones. Not sure whether this is
relevant in practice, but I think it's a nice property nonetheless.
(Btw, just for fun, I also found a package version
2.0.0~beta2+isreally1.8.6-0ubuntu1. "isreally". Oh yeah :)
> It would also be more efficient, I think (it seems like the sort is
> O(nr_tags * lg(nr_tags) * nr_suffixes) due to parsing suffixes in the
> comparator). Though that probably doesn't matter much in practice.
I don't think there will be more than only a few configured suffixes
in any repository. However, if you consider O as "number of
starts_with() invocations", then there is an additional suffix_length
factor. But then again, these suffixes tend to be short.
> I dunno. I think maybe your 6/7 has converged on an equivalent behavior.
> And I am certainly not volunteering to re-write it, so if what you have
> works, I'm not opposed to it.
>
> -Peff
next prev parent reply other threads:[~2016-12-20 8:50 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-05 22:42 2.10.0: multiple versionsort.prereleasesuffix buggy? Leho Kraav (Conversion Ready)
2016-09-05 23:21 ` Jeff King
2016-09-06 1:07 ` SZEDER Gábor
2016-09-06 4:07 ` Jeff King
2016-09-06 19:45 ` SZEDER Gábor
2016-09-07 15:12 ` [PATCH 0/5] Fix version sort prerelease reordering bug SZEDER Gábor
2016-09-07 15:12 ` [PATCH 1/5] t7004-tag: delete unnecessary tags with test_when_finished SZEDER Gábor
2016-09-07 15:12 ` [PATCH 2/5] t7004-tag: use test_config helper SZEDER Gábor
2016-09-07 15:12 ` [PATCH 3/5] t7004-tag: add version sort tests to show prerelease reordering issues SZEDER Gábor
2016-09-07 15:12 ` [PATCH 4/5] versioncmp: pass full tagnames to swap_prereleases() SZEDER Gábor
2016-09-08 17:49 ` Junio C Hamano
2016-09-08 20:37 ` SZEDER Gábor
2016-09-08 21:31 ` Junio C Hamano
2016-09-07 15:12 ` [PATCH 5/5] versioncmp: cope with common leading parts in versionsort.prereleaseSuffix SZEDER Gábor
2016-09-07 15:48 ` SZEDER Gábor
2016-09-09 10:43 ` Duy Nguyen
2016-10-05 1:33 ` SZEDER Gábor
2016-10-05 17:01 ` Junio C Hamano
2016-10-05 21:26 ` SZEDER Gábor
2016-10-05 22:15 ` Junio C Hamano
2016-10-06 0:40 ` Jacob Keller
2016-10-06 5:48 ` Duy Nguyen
2016-12-08 14:23 ` [PATCHv2 0/7] Fix and generalize version sort reordering SZEDER Gábor
2016-12-08 14:23 ` [PATCHv2 1/7] t7004-tag: delete unnecessary tags with test_when_finished SZEDER Gábor
2016-12-08 14:23 ` [PATCHv2 2/7] t7004-tag: use test_config helper SZEDER Gábor
2016-12-08 14:23 ` [PATCHv2 3/7] t7004-tag: add version sort tests to show prerelease reordering issues SZEDER Gábor
2016-12-08 14:23 ` [PATCHv2 4/7] versioncmp: pass full tagnames to swap_prereleases() SZEDER Gábor
2016-12-08 14:23 ` [PATCHv2 5/7] versioncmp: cope with common part overlapping with prerelease suffix SZEDER Gábor
2016-12-12 21:27 ` Junio C Hamano
2016-12-13 0:27 ` SZEDER Gábor
2016-12-13 6:39 ` Junio C Hamano
2016-12-08 14:24 ` [PATCHv2 6/7] versioncmp: use earliest-longest contained suffix to determine sorting order SZEDER Gábor
2016-12-08 14:48 ` [PATCHv2 6.5/7] squash! " SZEDER Gábor
2016-12-08 14:24 ` [PATCHv2 7/7] versioncmp: generalize version sort suffix reordering SZEDER Gábor
2016-12-08 19:36 ` Junio C Hamano
2016-12-14 17:08 ` [PATCHv2 0/7] Fix and generalize version sort reordering Jeff King
2016-12-14 17:36 ` Junio C Hamano
2016-12-20 8:50 ` SZEDER Gábor [this message]
2016-12-20 16:49 ` Jeff King
2016-09-06 7:12 ` 2.10.0: multiple versionsort.prereleasesuffix buggy? Leho Kraav (Conversion Ready)
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CAM0VKjmDDKgYCvtbwpx=GcwRENzvSDLW_Xhia3btdeMjtAjAvg@mail.gmail.com' \
--to=szeder.dev@gmail.com \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=leho@conversionready.com \
--cc=pclouds@gmail.com \
--cc=peff@peff.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).