git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [WIP 0/7] CDN offloading of fetch response

From
Jonathan Nieder <jrnieder@gmail.com>
Date
Feb 25, 2019, 23:45 UTC
Message-ID
<20190225234528.GD16965@google.com>
In-Reply-To
<CAP8UFD19521P=-R1XEdRK++HPatt3BJaxWPvpPRo8LHr3eisjA@mail.gmail.com>
Hi,
Christian Couder wrote:
> On Sun, Feb 24, 2019 at 12:39 AM Jonathan Tan <jonathantanmy@google.com> wrote:
Show 17 quoted lines
>> There are probably some more design discussions to be had:
>
> [...]
>
>>  - Client-side whitelist of protocol and hostnames. I've implemented
>>    whitelist of protocol, but not hostnames.
>
> I would appreciate a more complete answer to my comments in:
>
> https://public-inbox.org/git/CAP8UFD16fvtu_dg3S_J9BjGpxAYvgp8SXscdh=TJB5jvAbzi4A@mail.gmail.com/
>
> Especially I'd like to know what should the client do if they find out
> that for example a repo that contains a lot of large files is
> configured so that the large files should be fetched from a CDN that
> the client cannot use? Is the client forced to find or setup another
> repo configured differently if the client still wants to use CDN
> offloading?
The example from that message:
  For example I think the Great Firewall of China lets people in China
  use GitHub.com but not Google.com. So if people start configuring
  their repos on GitHub so that they send packs that contain Google.com
  CDN URLs (or actually anything that the Firewall blocks), it might
  create many problems for users in China if they don't have a way to
  opt out of receiving packs with those kind of URLs.

But the same thing can happen with redirects, with embedded assets in web pages, and so on. I think in this situation the user would likely (and rightly) blame the host (github.com) for requiring access to a separate inaccessible site, and the problem could be resolved with them.

The beauty of this is that it's transparent to the client: the fact that packfile transfer was offloaded to a CDN is an implementation detail, and the server takes full responsibility for it.

This doesn't stop a hosting provider from using e.g. server options to allow the client more control over how their response is served, just like can be done for other features of how the transfer works (how often to send progress updates, whether to prioritize latency or throughput, etc).

What the client *can* do is turn off support for packfile URLs in a request completely. This is required for backward compatibility and allows working around a host that has configured the feature incorrectly.

Thanks for an interesting example, Jonathan

Previous: Christian CouderNext: Christian Couder
Message 16 of 45 in “[WIP 0/7] CDN offloading of fetch response”
  1. Jonathan TanFeb 23, 2019
  2. 1/7 http: use --stdin and --keep when downloading packJonathan Tan, Feb 23, 2019
  3. 2/7 http: improve documentation of http_pack_requestJonathan Tan, Feb 23, 2019
  4. 3/7 http-fetch: support fetching packfiles by URLJonathan Tan, Feb 23, 2019
  5. 4/7 Documentation: order protocol v2 sectionsJonathan Tan, Feb 23, 2019
  6. 5/7 Documentation: add Packfile URIs design docJonathan Tan, Feb 23, 2019
  7. 6/7 upload-pack: refactor reading of pack-objects outJonathan Tan, Feb 23, 2019
  8. 7/7 upload-pack: send part of packfile response as uriJonathan Tan, Feb 23, 2019
  9. Junio C HamanoFeb 24, 2019
  10. Christian CouderFeb 25, 2019
  11. Jonathan NiederFeb 26, 2019
  12. Christian CouderFeb 26, 2019
  13. Josh SteadmonMar 1, 2019
  14. Jonathan TanMar 1, 2019
  15. Christian CouderFeb 25, 2019
  16. Jonathan NiederFeb 25, 2019
  17. Christian CouderFeb 26, 2019
  18. Ævar Arnfjörð BjarmasonFeb 26, 2019
  19. Christian CouderMar 4, 2019
  20. Jonathan NiederFeb 28, 2019
  21. Christian CouderMar 4, 2019
  22. 0/8 CDN offloading of fetch responseJonathan Tan, Mar 8, 2019
  23. 1/8 http: use --stdin when getting dumb HTTP packJonathan Tan, Mar 8, 2019
  24. 2/8 http: improve documentation of http_pack_requestJonathan Tan, Mar 8, 2019
  25. 3/8 http-fetch: support fetching packfiles by URLJonathan Tan, Mar 8, 2019
  26. 4/8 Documentation: order protocol v2 sectionsJonathan Tan, Mar 8, 2019
  27. 5/8 Documentation: add Packfile URIs design docJonathan Tan, Mar 8, 2019
  28. Jeff KingApr 23, 2019
  29. Jonathan TanApr 23, 2019
  30. Ævar Arnfjörð BjarmasonApr 23, 2019
  31. Jonathan NiederApr 23, 2019
  32. Ævar Arnfjörð BjarmasonApr 23, 2019
  33. Jonathan NiederApr 23, 2019
  34. Jonathan NiederApr 23, 2019
  35. Ævar Arnfjörð BjarmasonApr 23, 2019
  36. Jonathan NiederApr 23, 2019
  37. Ævar Arnfjörð BjarmasonApr 24, 2019
  38. Junio C HamanoApr 24, 2019
  39. 6/8 upload-pack: refactor reading of pack-objects outJonathan Tan, Mar 8, 2019
  40. 7/8 fetch-pack: support more than one pack lockfileJonathan Tan, Mar 8, 2019
  41. 8/8 upload-pack: send part of packfile response as uriJonathan Tan, Mar 8, 2019
  42. Josh SteadmonMar 19, 2019
  43. Jeff KingApr 23, 2019
  44. Jonathan TanApr 23, 2019
  45. Ævar Arnfjörð BjarmasonApr 24, 2019

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.