Closed ESPWarren closed 4 years ago
Oh dear, The zip file does not include the git metadata to be able to determine the version. I have no clue about Manjaro, Can you use git clone instead of the zip file? That would solve your problem.
I will think about how to make the zip file compile.
I have changed the way the version file is created. It is now part of the zip file and so should build on Manjaro
Thanks for the responsiveness - this really helps...
Ok that is obviously a move in the right direction - it no longer throws an MD5 error with the main zip file.
But it still has an issue:
Preparing... Cloning usbrelay build files... Checking usbrelay dependencies... Resolving dependencies... Checking inter-conflicts...
Building usbrelay... ==> Making package: usbrelay 20190629-1 (Sat 16 May 2020 14:06:08 BST) ==> Checking runtime dependencies... ==> Checking buildtime dependencies... ==> Retrieving sources... -> Found master.zip ==> Validating source files with md5sums... master.zip ... Passed ==> Removing existing $srcdir/ directory... ==> Extracting sources... -> Extracting master.zip with bsdtar ==> Starting prepare()... ==> Removing existing $pkgdir/ directory... ==> Starting build()... make: *** No rule to make target '.git/HEAD', needed by 'gitversion.c'. Stop. ==> ERROR: A failure occurred in build(). Aborting...
There is some old code somewhere. There is no reference to git in the makefile anymore.
On Sat., 16 May 2020, 10:12 pm Warren, notifications@github.com wrote:
Thanks for the responsiveness - this really helps...
Ok that is obviously a move in the right direction - it no longer throws an MD5 error with the main zip file.
But it still has an issue:
Preparing... Cloning usbrelay build files... Checking usbrelay dependencies... Resolving dependencies... Checking inter-conflicts...
Building usbrelay... ==> Making package: usbrelay 20190629-1 (Sat 16 May 2020 14:06:08 BST) ==> Checking runtime dependencies... ==> Checking buildtime dependencies... ==> Retrieving sources... -> Found master.zip ==> Validating source files with md5sums... master.zip ... Passed ==> Removing existing $srcdir/ directory... ==> Extracting sources... -> Extracting master.zip with bsdtar ==> Starting prepare()... ==> Removing existing $pkgdir/ directory... ==> Starting build()... make: *** No rule to make target '.git/HEAD', needed by 'gitversion.c'. Stop. ==> ERROR: A failure occurred in build(). Aborting...
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/darrylb123/usbrelay/issues/52#issuecomment-629636181, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABTSUVNMOQ357STF677N6I3RRZ7J3ANCNFSM4NBVFVOA .
Hi, I don't think it is a good idea for a distribution to build the package directly from master. First, as @ESPWarren points out, there will be problems with checksums whenever a new commit is pushed. Second, some versions might not be stable. For example, few commits back, one of my relays stopped working. Fortunately, Darryl fixed it before I found time to file an issue :-).
So I think we should do the following (I'll send PR if @darrylb123 is not against):
gitversion.c
is missing during build, Makefile should generate one with version ???
or something like this.gitversion.c
, which corresponds to the commit. These will be distributed as release assets. Darryl already did that for 0.7. I think I'll extend the Makefile to generate these release zip/tar, for example by running make release
. And uploading the assets can be automated. So maybe, we can also have make upload-assets
. Alternatively, this all could by automated with GitHub Action.AUR package is fixed.
Thanks Chris
Worked like a charm!
Nick
On 08/06/2020 01:43, Chris Severance wrote:
AUR package is fixed.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/darrylb123/usbrelay/issues/52#issuecomment-640306359, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE5INRWFVKDNRAEO4TH5NQDRVQX47ANCNFSM4NBVFVOA.
It has multiple failures. First it was the MDS for master.zip - I edited the build script.
That sorted the mds but then it failed further on - log:
GNU nano 4.9.2 g
Cloning usbrelay build files... Checking usbrelay dependencies... Resolving dependencies... Checking inter-conflicts...
Building usbrelay... ==> Making package: usbrelay 20190629-1 (Fri 15 May 2020 17:05:01 BST) ==> Checking runtime dependencies... ==> Checking buildtime dependencies... ==> Retrieving sources... -> Found master.zip ==> Validating source files with md5sums... master.zip ... Passed ==> Removing existing $srcdir/ directory... ==> Extracting sources... -> Extracting master.zip with bsdtar ==> Starting prepare()... ==> Starting build()... make: *** No rule to make target '.git/HEAD', needed by 'gitversion.c'. Stop. ==> ERROR: A failure occurred in build().
Here is the build script:
Maintainer: Your Name linuxboy@fel.hopto.org
pkgname=usbrelay pkgver=20190629 pkgrel=1 pkgdesc="Control usb relay - based on hidapi" url="https://github.com/darrylb123/usbrelay" license=('GPL-2.0') groups=('system') depends=('hidapi') arch=('i686' 'x86_64') makedepends=() checkdepends=() optdepends=() provides=() conflicts=() replaces=() backup=() options=() install= changelog= source=("https://github.com/darrylb123/usbrelay/archive/master.zip") noextract=() md5sums=('B42B221CC51B5FEA0D22661E379A92FE') validpgpkeys=()
prepare() { cd "$pkgname-master" }
build() { cd "$pkgname-master" make }
package() { cd "$pkgname-master" make DESTDIR="$pkgdir/" install }