Steffen Nurpmeso <stef...@sdaoden.eu> wrote: > Hello. > > u...@net9.ga wrote in > <e1kfa4f-0005p9...@mx1.net9.ga>: > > If you wanted to subscribe only for this thread, just say, i will > remove you again. You are whitelisted, and can simply post > without subscribing. sorry for the delay. > > |Steffen Nurpmeso <stef...@sdaoden.eu> wrote: > |> I Cc: the list, i hope this is ok? > | > |Yes, it is ok. > > Great, thanks. > > |> I just got it totally wrong, i apologise for that. > |> > |> u...@net9.ga wrote in > |> <e1kfkno-0004rv...@mx1.net9.ga>: > |>|The truncation is with s-nail 14.9.11-2 on debian buster. > |>|In my post I mentioned s-mailx becuase I was confusing s-nail with > |>|the list name. > |>|I think the cause of truncation are unicode characters that the poster \ > |>|included in his message. > |> > |> Ah! Now i understand -- you tried to read the message of the URL > |> you posted, and _that_ was impossible? > | > |Right. That message, which is archived at > |https://lists.archlinux.org/pipermail/arch-general/2020-September/048110\ > |.html , > |was truncated at the beginning of the line where the author > |had a right arrow symbol. The line in the body of the message. > |Rethinking, it could be the subject was shown without the right > |arrow character. Just skipping that symbol, and otherwise fine. > |But i am not sure. > > It just makes we wonder why the list archive shows the correct > Unicode symbol in the message body, whereas GNU iconv(3) failed > with an illegal sequence error. But anyhow, it is very likely > that the reverse commit i had shown fixes it, as that we now > simply continue again after such errors. > > I implemented a clear violation of "be conservative in what you do > and liberal in what you accept", and i should have known better > since faulty content-transfer encodings etc are quite common > indeed. > > |> Apologies, this was a complete misunderstanding on my side then!! > |> Sorry! > |> > |>|At this time I can not test it on archlinux, which defenitly has > |>|a newer s-nail version. Do you think it will not happen with a > |>|newer version? > |> > |> On Debian you could upgrade to v14.9.19 from the "testing" > |> repository, but i do not know how to accomplish that, i only used > |> on release of Debian, that was 3.0 (Woody), a long time ago. > | > |Unfortunately, in order to go from buster to bullseye (testing), > |I will have to upgrade some of s-nail dependencies. Which will > |probably requires upgrading more packages, for dependencies of > |other packages on the machine. I don't want to go into that now. > > I see. Yes, that is unfortunate. If you have a compiler you > could compile it locally for yourself, but that requires the -dev > packages of all the libraries that we need, however. At least > openssl, terminfo, hm. > > |> There are almost a thousand commits, and we do have unicode tests, > |> and they succeed. So i do not think it happens in a newer > |> version. > |> > |> I see there is a Unicode character in this line, and i would > |> really be interested in seeing the message as the ArchLinux list > |> posted it, unfortunately the Mailman archive does not have the > |> full message? Would you send it to me? It is possible the > |> easiest if you `resend' it ("? resend MSG-NO stef...@sdaoden.eu"). > | > |Unfortunately, I deleted the messages that started this discussion. > |But I think I am able to reproduce the problematic situation. > |Which is easy. Just have a message with unicode character. For example, > > No, it is surely not that easy, there must be an invalid Unicode > character. We do test Unicode messages, in the unit test. It was > a bit smaller in v14.9.11, but a bit of testing there was by then. > > | The next line should have a single right arrow, U+2192, → > | (Here comes the right arrow symbol) > | The previous line had a single right arrow, U+2192, → > | > |I can arrange for such a message to be sent to, if you like. > | > |No other comments on my part. I will consider upgrading to Debian > |testing whenever I will have an opportunity. > |Thank you. > > Ciao. >
14.9.19 back ported to Debian Buster, which is currently Debian stable. In case someone is not aware of that. https://lists.debian.org/debian-backports-changes/2020/11/msg00031.html https://packages.debian.org/buster-backports/s-nail -- u34
