Re: Verification of download pages and links

2024-05-01 Thread sebb
On Wed, 1 May 2024 at 18:14, tison wrote: > > Hi Sebb, > > I'm trying to make a reference in the next OpenDAL release, that staging > the download pages and links. > > Here is a technical issue. Before the release is out, which link should be > used? IIUC we can use the link at [1]. But it

Re: Verification of download pages and links

2024-05-01 Thread tison
Hi Sebb, I'm trying to make a reference in the next OpenDAL release, that staging the download pages and links. Here is a technical issue. Before the release is out, which link should be used? IIUC we can use the link at [1]. But it doesn't follow the INFRA policy for formal releases. However,

Re: Verification of download pages and links

2024-04-28 Thread tison
> The link to the source download and keys and hashes is broken. Thank you! File [1] to fix it. [1] https://github.com/apache/opendal/pull/4547 This shows the necessity of previewing the download page. At least since we provide it, we should ensure its correctness. Best, tison. Justin Mclean

Re: Verification of download pages and links

2024-04-28 Thread Justin Mclean
Hi, > Here is a patch to cover a minimal download page [1], which is derived from > OpenDAL's download page [2]. Welcome to leave comments if you find any > issues or things we can improve on. > > [1] https://github.com/apache/datafusion/pull/10271 > [2] https://opendal.apache.org/download The

Re: Verification of download pages and links

2024-04-28 Thread sebb
On Sun, 28 Apr 2024 at 19:46, Dave Fisher wrote: > > > > > On Apr 28, 2024, at 9:58 AM, tison wrote: > > > > Thank you! > > > > I found there is no rationale for these links, and thus, it's quite a bit > > challenging in memory. > > > > IIRC the closer.lua script is for selecting the most proper

Re: Verification of download pages and links

2024-04-28 Thread Dave Fisher
Tison, > On Apr 28, 2024, at 12:10 PM, tison wrote: > > Thank you, Dave :D > > It gives a reason. I'm OK with this explanation now so that I won't bring > it to the INFRA. > > Back to the original purpose of this thread, I suggest: > > 1. Go through our Incubator Guide and find if we have

Re: Verification of download pages and links

2024-04-28 Thread tison
Thank you, Dave :D It gives a reason. I'm OK with this explanation now so that I won't bring it to the INFRA. Back to the original purpose of this thread, I suggest: 1. Go through our Incubator Guide and find if we have some references to this release distribution policy (maybe in [1][2]).

Re: Verification of download pages and links

2024-04-28 Thread Dave Fisher
> On Apr 28, 2024, at 9:58 AM, tison wrote: > > Thank you! > > I found there is no rationale for these links, and thus, it's quite a bit > challenging in memory. > > IIRC the closer.lua script is for selecting the most proper CDN for > source/binary bundles in use. They can, technically,

Re: Verification of download pages and links

2024-04-28 Thread tison
Thank you! I found there is no rationale for these links, and thus, it's quite a bit challenging in memory. IIRC the closer.lua script is for selecting the most proper CDN for source/binary bundles in use. They can, technically, work for SHASUM and signatures also. Why do we use

Re: Verification of download pages and links

2024-04-28 Thread sebb
On Sun, 28 Apr 2024 at 15:38, tison wrote: > > Yeah. I support that we always need to release sources on our platform. > > Given the links to downloads.apache.org, archive.apache.org, > https://www.apache.org/dyn/closer.lua, can be unintuitive for users, I > agree that we can have a simple

Re: Verification of download pages and links

2024-04-28 Thread tison
Yeah. I support that we always need to release sources on our platform. Given the links to downloads.apache.org, archive.apache.org, https://www.apache.org/dyn/closer.lua, can be unintuitive for users, I agree that we can have a simple Download page for such library-only projects. Here is a

Re: Verification of download pages and links

2024-04-27 Thread Justin Mclean
Hi, Projects need to make source releases on ASF infrastructure and have a download page for good reasons. Some users need a place to verify and download a trusted release. Having it hosted on ASF infrastructure means people can 100% trust it, unlike 3rd party providers. 3rd party providers

Re: Verification of download pages and links

2024-04-25 Thread tison
On the one hand, we release sources. We should ensure that the source code is released (on our platform), or else the pivotal character "open source" is challenged. On the other hand, library users do always "download" the software with a dependency manager, and thus, a heavy download page may

Re: Verification of download pages and links

2024-04-25 Thread tison
Here is a new situation for projects that may not carry source releases heavily: [1] [1] https://github.com/apache/datafusion/pull/10233/files#r1579895024 They may not even maintain a download page, but just leverage https://downloads.apache.org/. I suggest we rethink the form of distribution

Re: Verification of download pages and links

2024-04-23 Thread sebb
On Tue, 23 Apr 2024 at 13:25, tison wrote: > > Thanks for starting this thread and thanks a lot for verifying the download > page for a lot of podlings! > > For previewing a staging website, with .asf.yaml config, there is a way [1] > to do so self-served by any (P)PMC. And I agree that we should

Re: Verification of download pages and links

2024-04-23 Thread tison
Thanks for starting this thread and thanks a lot for verifying the download page for a lot of podlings! For previewing a staging website, with .asf.yaml config, there is a way [1] to do so self-served by any (P)PMC. And I agree that we should spread such practices to every project. [1]

Verification of download pages and links

2024-04-23 Thread sebb
The primary mission of the ASF is to produce open SOURCE, so it seems to me that an essential part of a release is a download page with properly constituted links to the source bundle and the associated download verification files. However, no attention seems to be given to this aspect of a