1875646 - https://sk-data.special-k.info/SpecialK.7z page is treated like plain text instead of archive that should be downloaded

Status: RESOLVED WORKSFORME

(Web Compatibility :: Site Reports, defect, P3)

Description

[soredake]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:121.0) Gecko/20100101 Firefox/121.0

Steps to reproduce:

https://sk-data.special-k.info/SpecialK.7z page is treated like plain text instead of archive that should be downloaded.

Actual results:

https://sk-data.special-k.info/SpecialK.7z page is treated like plain text instead of archive that should be downloaded.

Expected results:

https://sk-data.special-k.info/SpecialK.7z should download archive, in edge this url do this.

Comment 1

[soredake]

Attachment: Firefox displays this, nothing is downloaded

Comment 2

[soredake]

Attachment: Edge correctly triggers download for this url

Comment 3

[Daniel Bodea [:danibodea]]

I confirm this issue in Windows 10 and Ubuntu 22 as the link opens a corrupted text page instead of downloading an archive. This does not occut in Chrome.

Comment 4

[:Gijs (he/him)]

The site sends no content-type header. I'd expect the networking code to sniff. That part of the web is under-specified in terms of compatibility, so I don't know if there's just a sniffer difference between chromium/blink (Edge) and gecko in terms of how they determine text/plain vs. non-text here - but that's somewhere in networking code, so moving over there.

Comment 5

[Kershaw Chang [:kershaw]]

(In reply to :Gijs (he/him) from comment #4)

The site sends no content-type header. I'd expect the networking code to sniff. That part of the web is under-specified in terms of compatibility, so I don't know if there's just a sniffer difference between chromium/blink (Edge) and gecko in terms of how they determine text/plain vs. non-text here - but that's somewhere in networking code, so moving over there.

It seems to me that the content-type header from the site is text/plain.

HTTP/2 200 
date: Fri, 02 Feb 2024 16:30:00 GMT
content-type: text/plain
last-modified: Sun, 10 Sep 2023 03:12:44 GMT
x-rgw-object-type: Normal
etag: W/"9819140a060bdd01f0914e899d003aa6"
x-amz-request-id: tx000003cbc630e7525cf49-0065b6cd26-4f1170e1-nyc3b
vary: Origin, Access-Control-Request-Headers, Access-Control-Request-Method, Accept-Encoding
strict-transport-security: max-age=15552000; includeSubDomains; preload
x-do-cdn-uuid: a1be59ba-b6d2-43b8-ad4f-f4c143529f9b
cache-control: max-age=3600
x-envoy-upstream-healthchecked-cluster: 
cf-cache-status: HIT
set-cookie: __cf_bm=ZpIvml9wu6QhI6lxz0mCk4_09FOkbkKXLwfoK4Wd9.s-1706891400-1-AY4kKir7WwmRNnveB8am/dF3WuQ3yl1hGjnm4irKyPhftiBYXe28pGHY4foup7qGaQmMTvYfFmunsnsqiSyk3nI=; path=/; expires=Fri, 02-Feb-24 17:00:00 GMT; domain=.sk-data.special-k.info; HttpOnly; Secure; SameSite=None
server: cloudflare
cf-ray: 84f3d0f1ee0d9249-FRA
content-encoding: gzip

I think this is a web compat issue.

Comment 6

[:Gijs (he/him)]

FWIW I see this in the network monitor, so the 304 has no content type, which is what confused me:

HTTP/2 304 
date: Fri, 02 Feb 2024 16:52:16 GMT
last-modified: Sun, 10 Sep 2023 03:12:44 GMT
x-rgw-object-type: Normal
etag: "9819140a060bdd01f0914e899d003aa6"
x-amz-request-id: tx000003cbc630e7525cf49-0065b6cd26-4f1170e1-nyc3b
vary: Origin, Access-Control-Request-Headers, Access-Control-Request-Method, Accept-Encoding
strict-transport-security: max-age=15552000; includeSubDomains; preload
x-do-cdn-uuid: a1be59ba-b6d2-43b8-ad4f-f4c143529f9b
cache-control: max-age=3600
x-envoy-upstream-healthchecked-cluster: 
cf-cache-status: HIT
set-cookie: __cf_bm=Uu28LPnwRsnPkLBm.8as9bqM1d8bVtWCGT07PTrPwGE-1706892736-1-AcUwNw4ZQLscVUeyjr++3S8yuXSAlymbWgHrmrUN5/3Vu2FCDirtZZ1oj0J1ypIk9OkTnGIOrn+s8IEZQY8pcNo=; path=/; expires=Fri, 02-Feb-24 17:22:16 GMT; domain=.sk-data.special-k.info; HttpOnly; Secure; SameSite=None
server: cloudflare
cf-ray: 84f3f19429d47785-LHR
X-Firefox-Spdy: h2

Kershaw, do you know why chromium does produce a download here? Is there a spec issue / chromium issue?

Comment 7

[Kershaw Chang [:kershaw]]

Kershaw, do you know why chromium does produce a download here? Is there a spec issue / chromium issue?

I assume it's because the file name SpecialK.7z, but I am not really sure.
I'll search chromium issues and fetch spec to see if I can find something.

Comment 8

[Dennis Schubert [:denschub]]

The URL in question appears to a) no longer redirect, b) no longer send text/plain:

# curl -I https://sk-data.special-k.info/SpecialK.7z
HTTP/2 200
...
content-type: application/octet-stream