[studxml-users] Hash mismatch wehn building from cppget.org
Hollenbach, Lars
lars.hollenbach at secunet.com
Wed Nov 20 11:34:21 EST 2024
The checksum matches when I download the archive directly:
$ wget
https://pkg.cppget.org/1/beta/libstudxml/libstudxml-1.1.0-b.10+2.tar.gz
$ sha256sum libstudxml-1.1.0-b.10+2.tar.gz
2f2cdab781da70ba6c214f73937145a86f53a24f55d276cdd2563b6d889fc2e4
libstudxml-1.1.0-b.10+2.tar.gz
But, if I use the same command that bpkg shows me with --verbose 3:
fetching libstudxml-1.1.0-b.10+2.tar.gz from pkg:cppget.org/beta
curl -L -A "bpkg/0.17.0 (GNU/Linux; +https://build2.org) libbpkg/0.17.0
libbutl/0.17.0 curl" -i
https://pkg.cppget.org/1/beta/libstudxml/libstudxml-1.1.0-b.10+2.tar.gz
I get a different checksum each time.
Which makes sense, because curl -i includes the response headers in the
output.
This is the same behaviour as with bpkg - a new checksum each time.
But the headers should get removed/ignored by bpkg, right?
On 20.11.24 13:46, Boris Kolpackov wrote:
> Hollenbach, Lars <lars.hollenbach at secunet.com> writes:
>
>> error: checksum mismatch for libstudxml 1.1.0-b.10+2
>> info: pkg:cppget.org/beta has 2f2cdab781da70ba6c214f73937145a86f53a24f55d276cdd2563b6d889fc2e4
>> info: fetched archive has dda2021e32e2a73b8b8f7f01aa5681f1dc388482da9482d6e6c2f73be5a57af7
>>
>> The problem persists, even after running "bpkg fetch".
>>
>> I would assume this is an issue with the current package on cppget, correct?
> Hm, I just tried to build it and things seem to work fine:
>
> $ bpkg create -d cfg cc
> $ cd cfg
> $ bpkg build libstudxml at https://pkg.cppget.org/1/beta
>
> I've also checked the checksum and it matches what your diagnostics
> says is on cppget.org:
>
> $ sha256sum libstudxml-1.1.0-b.10+2.tar.gz
> 2f2cdab781da70ba6c214f73937145a86f53a24f55d276cdd2563b6d889fc2e4 libstudxml-1.1.0-b.10+2.tar.gz
>
> I am not sure what's going on. Can you try to download the archive
> directly and see what checksum you get:
>
> https://pkg.cppget.org/1/beta/libstudxml/
More information about the studxml-users
mailing list