For details see:
http://www.squid-cache.org/Versions/v4/changesets/
In July 2018, 'squid 4' was "released for production use", see:
https://wiki.squid-cache.org/Squid-4
"The features have been set and large code changes are reserved for later versions."
I've tested almost all 4.x-versions and patch series before with good results.
Right now, 4.4 is running here with no seen problems together with
'squidclamav', 'squidguard' and 'privoxy'.
I too would declare this version stable.
Best,
Matthias
Signed-off-by: Matthias Fischer <matthias.fischer@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Download the requested rules tarball into a temporay file
and if every thing is fine, replace the old by the
downloaded one.
In addition with the previously implemented file size check,
we are saved now from a corrupt rules tarball on disk.
Signed-off-by: Stefan Schantl <stefan.schantl@ipfire.org>
The downloader now requests the html header for the rulestarball
and obtain the size of the file bevore downloading it.
After success the size of the downloaded file will be compared with
the requested one before. If they do not match, an error will be gained.
Signed-off-by: Stefan Schantl <stefan.schantl@ipfire.org>
This reverts commit 6174b7b1c7.
This had absolutely no effect on build time or rather made it
slower. So this is being reverted to save ourselves the RAM.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
This is an experimental change that I want to trial to speed up
the nightly builds. The build environment will be mounted in a
ramdisk and the build will be performed in there.
This will hopefully reduce IO on the (slow) replicated disks.
If there is no significant performance gain from this, this
commit will be reverted.
To enable this, USE_RAMDISK must be set to 1 in .config.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
It doesn't need to be re-installed because no system with
that configuration should exist right now.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
This change is necessary to make sure that the script prefers
are link with internet access. That would usually be red (after
the second boot) or eth* (on the first boot).
That allows (and ensures) that we can install packages in
the user-data script.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
This variable was not initialized on systems where EFI was not
in use. Therefore the generated parted command line was not
valid and caused the installation to abort.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
The display should displayed always except the linux-pae
packages is planned to be installed after this update.
Signed-off-by: Arne Fitzenreiter <arne_f@ipfire.org>
This is useful when the user-data script is installing
packages. For that it will need valid keys for course.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
