Closed codesardine closed 6 years ago
Why is an epoch change needed here anyway? Why was one introduced back here: https://github.com/codesardine/packages-community/commit/a89f5737a1a21c2141cd87f02a25b88eeda31f2a#diff-88bc577e9acac9d5c27d19ac70e066b2 ?
There are too many things wrong with this PKGBUILD, e.g. empty variables.
Reading and following the Arch packaging standards would be a good thing for Manjaro contributors too. We need to have a proper discussion about this so we can maintain a consistent standard.
IIRC there's an Arch packaging class they do via IRC. Might be worth looking into.
@philmmanjaro introduced the epoch back then, how you guys recommend doing this? @Ste74 @jonathonf
One approach to avoid an epoch this time round:
$ vercmp 1 2
-1
$ vercmp 2.0.0 2.0.0.canary
-1
$ vercmp 2.0.0.0 2.0.0.canary
1
In future, add "pre-release" tags directly onto the pkgver
, i.e.
$ vercmp 2.1.0 2.1.0canary
1
so 2.1.0 will replace 2.1.0canary. This will avoid the need for an epoch.
I like your method :)
@jonathonf tried to build hyper with pkgver 2.0.0.0 but epoch array still remain the main version .. i will upload the 2.0.0 and remove the version with epoch . After we follow your suggestion example 2.1.0canary
There is a more recent canary version available, I was thinking you guys prefer stable @ste74
@oberon2007 @philmmanjaro @jonathonf @codesardine , i don't like use
epoch
here.. also because what if we push another canary version like 3.x.x.canary? When upstream release another major release we have to useepoch 3
.. Is possible removeepoch
and remove from boxit theepoch
version ? And why add https://github.com/manjaro/packages-community/pull/483/files#diff-88bc577e9acac9d5c27d19ac70e066b2R21 conflicts and replaces for hyper? In any case it update itself