The debian-l10n-devel@lists.alioth.debian.org mailing list is getting
regular mails sent from franck, every 6 hours at H+52:

Date: Mon, 11 May 2015 19:52:22 +0000
From: Debian FTP Masters <ftpmaster@ftp-master.debian.org>
To: debian-l10n-devel@lists.alioth.debian.org
Subject: [Debian-l10n-devel] Lisa, if you don't like your job you
don't strike. You just go in every day and do it really half-assed.
That's the American way.
X-CRM114-Status: Good  ( pR: 39.5418 )

Arf, Arf, Arf, missing the timestamp 201505022135 directory, not
updating i18n, arf, arf, arf


These mails are usually related to the package description i18n magic
but they have been cryptic for years.... and I can't indeed really
find if the problem lies on franck or on i18n.debian.org

At each dinstall, as always, short after a release.

Yes. Come on, it fits. The i18n job is currently done half - only
one part of it works, the other fails. :)

There it even tells you why. Clear and concise and directly to the
point! :)

i18n.

The function thet emits it (shortened a little) is the following (put
some comments at the right place, marked with >)

# The first i18n one, syncing new descriptions
function i18n1() {
    rsync -aq --delete --delete-after ddtp-sync:/does/not/matter . || true
    if $(gpgv --keyring /srv/ftp-master.debian.org/s3kr1t/dot-gnupg/pubring.gpg timestamp.gpg timestamp); then
        TSTAMP=$(cat timestamp)
        if [ -d ${scriptdir}/i18n/${TSTAMP} ]; then
            if ${scriptsdir}/ddtp-i18n-check.sh . ${scriptdir}/i18n/${TSTAMP}; then

            else
                echo "Arf, Arf, Arf, bad guys, wrong files, arf, arf, arf" | mail -a "X-Debian: DAK" -s "Don't you kids take anything. I'm watching you. I've got eye implants in the back of my head." -a "From: Debian FTP Masters <ftpmaster@ftp-master.debian.org>" debian-l10n-devel@lists.alioth.debian.org
            fi
        else
            echo "Arf, Arf, Arf, missing the timestamp ${TSTAMP} directory, not updating i18n, arf, arf, arf" | mail -a "X-Debian: DAK" -s "Lisa, if you don't like your job you don't strike. You just go in every day and do it really half-assed. That's the American way." -a "From: Debian FTP Masters <ftpmaster@ftp-master.debian.org>" debian-l10n-devel@lists.alioth.debian.org

            fi
    else
        echo "Arf, Arf, Arf, could not verify our timestamp signature, arf. Don't mess with our files, i18n guys, arf, arf, arf" | mail -a "X-Debian: DAK" -s "You can't keep blaming yourself. Just blame yourself once, and move on." -a "From: Debian FTP Masters <ftpmaster@ftp-master.debian.org>" debian-l10n-devel@lists.alioth.debian.org
    fi
}

Thing is, our generation of your input happens as

    cd ${scriptdir}/i18n/${STAMP}
    for suite in stable testing unstable; do
        codename=$(dak admin s show ${suite}|grep '^Codename')
        codename=${codename##* }
        echo "Codename is ${codename}"
        dak control-suite -l ${suite} >${codename}
    done
    echo "${STAMP}" > timestamp

so we always give you the current three suites as above, using their
codename. If your site doesn't update after a release, your processing
fails, then there is nothing to read from us and you get to read "Arf,
Arf, Arf". :)

At each dinstall, as always, short after a release.

Yes. Come on, it fits. The i18n job is currently done half - only
one part of it works, the other fails. :)

There it even tells you why. Clear and concise and directly to the
point! :)

i18n.

The function thet emits it (shortened a little) is the following (put
some comments at the right place, marked with >)

# The first i18n one, syncing new descriptions
function i18n1() {
    rsync -aq --delete --delete-after ddtp-sync:/does/not/matter . || true
    if $(gpgv --keyring /srv/ftp-master.debian.org/s3kr1t/dot-gnupg/pubring.gpg timestamp.gpg timestamp); then
        TSTAMP=$(cat timestamp)
        if [ -d ${scriptdir}/i18n/${TSTAMP} ]; then
            if ${scriptsdir}/ddtp-i18n-check.sh . ${scriptdir}/i18n/${TSTAMP}; then

            else
                echo "Arf, Arf, Arf, bad guys, wrong files, arf, arf, arf" | mail -a "X-Debian: DAK" -s "Don't you kids take anything. I'm watching you. I've got eye implants in the back of my head." -a "From: Debian FTP Masters <ftpmaster@ftp-master.debian.org>" debian-l10n-devel@lists.alioth.debian.org
            fi
        else
            echo "Arf, Arf, Arf, missing the timestamp ${TSTAMP} directory, not updating i18n, arf, arf, arf" | mail -a "X-Debian: DAK" -s "Lisa, if you don't like your job you don't strike. You just go in every day and do it really half-assed. That's the American way." -a "From: Debian FTP Masters <ftpmaster@ftp-master.debian.org>" debian-l10n-devel@lists.alioth.debian.org

            fi
    else
        echo "Arf, Arf, Arf, could not verify our timestamp signature, arf. Don't mess with our files, i18n guys, arf, arf, arf" | mail -a "X-Debian: DAK" -s "You can't keep blaming yourself. Just blame yourself once, and move on." -a "From: Debian FTP Masters <ftpmaster@ftp-master.debian.org>" debian-l10n-devel@lists.alioth.debian.org
    fi
}

Thing is, our generation of your input happens as

    cd ${scriptdir}/i18n/${STAMP}
    for suite in stable testing unstable; do
        codename=$(dak admin s show ${suite}|grep '^Codename')
        codename=${codename##* }
        echo "Codename is ${codename}"
        dak control-suite -l ${suite} >${codename}
    done
    echo "${STAMP}" > timestamp

so we always give you the current three suites as above, using their
codename. If your site doesn't update after a release, your processing
fails, then there is nothing to read from us and you get to read "Arf,
Arf, Arf". :)

It may be related to the fact that the i18n check is failing on
ddtp.debian.net with the message:

./Translation-files_to-check/dists/jessie/main/i18n/Translation-da contains packages which are not in sid

along with a list of 50 packages or so. No idea why that is a problem.
The code in question can be found at:

https://anonscm.debian.org/cgit/debian-l10n/ddtp.git/tree/ddtp-dinstall/checks/ddtp_i18n_check.sh#n379

but I'm not understanding what the code is actually trying to do...

Have a nice day,

   -- Arthur Schopenhauer

I've updated all the code in that file to deal with the $STABLE since
in the past we only ever got $TESTING and $UNSTABLE.  At least in local
testing that appears to stop the errors occurring. However, I don't
think this will completely fix the problem since AIUI the server that
retrieves the Translation files will run the same tests and fail in the
same way. But we'll see...

Have a nice day,

   -- Arthur Schopenhauer

I've tracked down the root cause of the problem. It lies in the recent
commit to dak which switches stable to jessie. (I can't find a gitweb
for dak so can't provide a link). The repo is here:
https://ftp-master.debian.org/git/dak.git/

commit 19ffc1de995fd165d9438a8b7769cd7c45bbce60
Author: Ansgar Burchardt <ansgar@debian.org>
Date:   Sat Apr 25 10:49:14 2015 +0200

    Switch to "jessie" as stable.


In some places jessie was replaced with stretch and in some places
stretch was added, inconsistantly.

To fix this, dak needs to be patched to either apply the rest of the
changes to the ddtp_i18n_check.sh script in this commit:

https://anonscm.debian.org/cgit/debian-l10n/ddtp.git/commit/?id=ffd41dc00a0d3e10cbb0d455a4d3ca0236fb64a3

Or revert the change to dak that would allow updated translations to
stable. That would be the following patch I beleive (though I don't
really understand what the code is doing here):

    (cd $base/ftp/
-            for dist in sid jessie stretch; do
*            for dist in sid stretch; do
                    find ./dists/$dist/main/i18n/ \! -type d | sort -u | gzip -9 > $base/ftp/indices/files/components/translation-$dist.list.gz
            done
    )

Personally I would like it if we could provide updated translations for
stable, but in the past this has been rejected.

Hope this helps,

   -- Arthur Schopenhauer

Quoting Felipe Augusto van de Wiel (faw) (faw@funlabs.org):


We have daily meaningless "arf arf" messages that are received at
debian-l10n-devel@lists.alioth.debian.org mail addresse.

Martijn tried to have a look at this (see debian-l10n-devel mailing
list archive)....and I reported #785092....