Re: [Evolution] Version information when building from git - Was: Many strange errors after upgrading to Evolution 3.24.2?

[Milan Crha <mcrha redhat com>]

On Wed, 2017-04-26 at 14:13 +0200, Ralf Mardorf wrote:
> don't expect that it will be fixed for Arch packages.

        Hi,
I see.

> "Comment by Jan de Groot (JGC) - Wednesday, 26 April 2017, 10:21 GMT
> IMHO this is only cosmetic. Distributions package software with
> shitloads of patches and changes to default enabled/disabled options
> most of the time, so even when evolution -v shows 3.24.1 there's no
> warranty that it's the same 3.24.1 you would get when building from a
> released tarball with default options.

Right. Still it's clear that it's 3.24.1 with added things, rather than
unreleased next version. It confuses people even more when you tell
them: "hey, you reported the issue against 3.24.2, aha, and that issue
*will be* fixed in *the next* 3.24.2 release". Who does understand
this? People reading archives? I doubt it. Just nobody but package
maintainer of that distro knows how this anomaly could happen.

> If upstream cares about versions in evolution -v output,

Misunderstanding, upstream cares of Help->About too.

> then upstream should not commit "Post-release version bump" but do
> that on release.

Well, hmm, rather nothing.

> We build from git because of 2 main reasons:
> - easier to add upstream committed patches

Okay, so Arch is bleeding bleeding bleeding edge distro. Noted. And
it's the only distribution (I know of) doing it this way. Why would one
make changes for one ditro, not the opposite? Distro is for users, no?

> - release tarballs are often generated with broken tools

Often? Give me a proof, please. Or not. I do not want to argue. The way
the reply had been written doesn't sound like leading to anything
usable.

> I removed it from my comment, since actually Claws' "About" version
> string was reformatted by the PKGBUILD,

I do not know how they do it, maybe some sort of hook when downloading
snapshot or whatever. It doesn't worth to do it in evolution* due to
one odd distribution. If they want to make chaos, then it's their call.

Anyway, thanks for the follow up Ralph. I consider this issue
closed/rejected by the Arch package maintainer.
        Bye,
        Milan