ETag header is not double quoted

[zapb]

According to RFC 2616 (3.11 Entity Tags) the value of the ETag header must be a double
quoted string but the swift server always returns unquoted strings.

I think the code snippet (https://github.com/openstack/swift/blob/master/swift/obj/server.py#L647):

metadata = {
    'X-Timestamp': request.headers['x-timestamp'],
    'Content-Type': request.headers['content-type'],
    'ETag': etag,
    'Content-Length': str(upload_size),
}

should be changed like that:

metadata = {
    'X-Timestamp': request.headers['x-timestamp'],
    'Content-Type': request.headers['content-type'],
    'ETag': '"' + etag + '"',
    'Content-Length': str(upload_size),
}

There might be some other code snippets to change as well.

zapb

[John Dickinson (notmyname)]

When swift was originally written, it was to replace an existing system. The requirements for swift were, "must be better than the old system and customers can't notice". That meant we had to port over a couple of the bugs from the old system (like no quotes on the etag). So it comes down to "historical reasons", and now at this point (3+ years of production swift deployments), we can't assume that clients will still work if we change that. and so it stays. If we have a swift api v2, that's one of the things that would probably change

[John Dickinson (notmyname)]

Changing back to "new" from "won't fix" until this is discussed with other core devs.

It may be possible to allow an option config parameter to control this behavior. If that's an acceptable solution to all of the core devs, or if there is a better on, we can accept the bug and move forward.

However, if the core devs determine that this is the functionality for swift api v1 and it won't be changed, even with a config option, I will re-close this as "won't fix"

[Chmouel Boudjnah (chmouel)]

I think this has been decided that it something we want to live with for 1.0 API and maybe fix in a new API (1.1 or 2.0) ? We may be able to change this back to "Won't Fix"

[Samuel Merritt (torgomatic)]

It's been noted as something to fix in the next rev of the Swift API, whenever that is.

https://wiki.openstack.org/wiki/SwiftNextAPI

[Tim Burke (1-tim-z)]

Maybe we could make it an option somewhere in the proxy-server?

[Tim Burke (1-tim-z)]

https://review.opendev.org/#/c/695131/

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/swift 2.24.0 release.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: feature/losf
Review: https://review.opendev.org/713632

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/713632
Committed: https://git.openstack.org/cgit/openstack/swift/commit/?id=79bd2e59e5c15ee84814ec1c4f0893176ba79412
Submitter: Zuul
Branch: feature/losf

commit f2ffd900593db2829a39a073f0c033d82985c40f
Author: Clay Gerrard <email address hidden>
Date: Fri Feb 28 11:09:51 2020 -0600

    Apply limit to list versioned containers

    Change-Id: I28e062273d673c4f07cd3c5da088aa790b77a599
    Closes-Bug: #1863841

commit dc40779307095b8d0b2761b77b9cb2904ec721ae
Author: Clay Gerrard <email address hidden>
Date: Fri Feb 28 10:00:25 2020 -0600

    Use float consistently for proxy timeout settings

    Change-Id: I433c97df99193ec31c863038b9b6fd20bb3705b8

commit 55049beda5b9d7038a3604a87f28312d7702ccb2
Author: Tim Burke <email address hidden>
Date: Fri Feb 28 18:59:32 2020 -0800

    tests: Use timedelta to adjust dates, not string manipulations

    Change-Id: I8f65ccd7f2a79d5b877bfbef0274fb7857e21391

commit 3b65a5998cc921d2763cf1a9ec1e40b88491262d
Author: Tim Burke <email address hidden>
Date: Wed Jan 10 06:16:41 2018 +0000

    Fix up some Content-Type handling in account/container listings

    Update content type on 204 (not just 200) to properly handle HEAD
    requests from xml/txt listings.

    Add "Vary: Accept" header to listings, since otherwise, browsers may
    serve the wrong content type from cache (even though we *would have*
    sent the *right* type if it actually sent the request).

    Change-Id: Iaa333aaca36a8dc2df65d38ef2173e3a6e2000ee

commit ecca23eb806e11cf6517f0456483da7a065350a8
Author: Clay Gerrard <email address hidden>
Date: Fri Feb 21 15:33:21 2020 -0600

    Extend eventlet_debug logging to GreenAsyncPile

    Change-Id: Ibd9fe5c9a1e75b86eb7d540594d5cf516758e17e

commit 0fb3371484f1d0f629d0b0e33f6aafbff0e43ee9
Author: Sam Morrison <email address hidden>
Date: Tue Feb 18 10:17:50 2020 +1100

    Delay importing swiftclient until after monkey-patching

    Commit message below partly copied from nova:

    Eventlet monkey patching should be as early as possible

    We were seeing infinite recursion opening an ssl socket when running
    various combinations of python3, eventlet, and urllib3. It is not
    clear exactly what combination of versions are affected, but for
    background there is an example of this issue documented here:

    https://github.com/eventlet/eventlet/issues/371

    The immediate cause in swift's case was that we were calling
    eventlet.monkey_patch() after importing swiftclient (which imported
    requests, which finally imported urllib3).

    We only use the imported function in one place, however; hold off on
    importing until we actually need it to ensure that monkey patching
    happens first. Note that we *don't* want to monkey-patch at import time,
    as we've previously had bugs related to import-time side-effects.

    Change-Id: I24f4bcc3d62dc37fd9559032bfd25f5b15f98745
    Closes-bug: #1863680
    Related-bug: #1380815

commit a5afe767581d2cb97cf3690067e6d626c7682c2c
Author: Tim Burke <email address hidden>
Date: Wed Feb 19 10:09:49 2020 -0800

    Revert "Make roll...

[Tim Burke (1-tim-z)]

Fixed in https://review.opendev.org/c/openstack/swift/+/700056, released in Swift 2.24.0 as part of the Openstack Ussuri release.