On 28/07/2025 06:23, Brian Inglis via Cygwin-apps wrote:
On 2025-07-27 13:18, Jon Turney via Cygwin-apps wrote:
On 05/07/2025 15:16, Andrew Schulman via Cygwin-apps wrote:
On Wed, 2 Jul 2025 22:06:45 +0100, Jon Turney via Cygwin-apps wrote:
Having developers build executable packages locally and then upload
them
doesn't really meet contemporary standards.
Given my druthers, I'd just disable sftp package uploads right now, and
make you all use the janky "build service" I hacked together in a few
spare weekends.
Are there any Cygwin packages that aren't fully cygport-scriptable
yet? For
example, are there any that rely on (free) software not yet available in
Cygwin for their builds? If so, then they'll need to keep being built
locally
and manully uploaded.
So, to be clear, the answer is "yes" to the first part, and "no" to
the second part.
Instances of the first case that I'm aware of are:
* all(?) of Jari Aalto's packages use g-b-s for build and packaging.
(This is probably supportable with some sort of extension to declare
the build-requires)
* mintty assembles the package archives itself and uploads those along
with a set of pre-written hint files. (I'm trying to work with Thomas
to make this a bit more like regular usage)
I don't think there are any instances of the second case, and such a
thing would probably be unacceptable as a package.
A number of packages have some failing test cases which prevent deploy,
or hang under CI requiring cancel: sometimes getting the tests to run is
like playing whack-a-mole, and some upstreams don't even respond to
patch submissions, although they are sometimes incorporated; others tell
you not to hold your breath, as it could be months. ;^>
failing some tests or taking too long on build/test is not uncommon
on same of my packages
Regards
Marco
