I am writing some Elisp code that downloads files using url-copy-file, and most of the time it works fine, but sometimes the contents of the file end up being the http headers, e.g. the downloaded file has the following contents:
HTTP/1.1 200 OK
Server: GitHub.com
Date: Thu, 14 Nov 2013 20:54:41 GMT
Content-Type: text/plain; charset=utf-8
Transfer-Encoding: chunked
Status: 200 OK
Strict-Transport-Security: max-age=31536000
X-XSS-Protection: 1; mode=block
X-Content-Type-Options: nosniff
X-UA-Compatible: chrome=1
Access-Control-Allow-Origin: https://render.github.com
Or sometimes the following is appended to the end of an otherwise correctly-downloaded file:
0
- Peer has closed the GnuTLS connection
However, when these things occur, the function seems to return just fine, so there is no way for me to verify that the file has really been downloaded. Is there any more reliable way to download a file in elisp (without shelling out to wget/curl/whatever)?
As recommended, I have reported this as a bug: lists.gnu.org/archive/html/bug-gnu-emacs/2013-11/msg00758.html
Related
My initial response headers - notice the Accept-Ranges header
HTTP/1.1 200 OK
X-Powered-By: Express
Vary: Origin
Access-Control-Allow-Credentials: true
X-RateLimit-Limit: 1
X-RateLimit-Remaining: 0
Date: Thu, 08 Apr 2021 06:14:19 GMT
X-RateLimit-Reset: 1617862461
Accept-Ranges: bytes
Content-Length: 100000000
Content-Type: text/plain; charset=utf-8
Content-Disposition: attachment; filename="some_file.txt"
Connection: keep-alive
Keep-Alive: timeout=5
I then restart the server and click resume download in chrome, but chrome doesn't send back in Range request headers
I'm following the documentation on Mozilla's website
Am I missing a header or misunderstanding how this works, especially with chrome and other browsers? Is there another way I can manually support resuming downloads by sending the right response and understanding the right request? From a technical perspective, if chrome sends back which range it now needs I will be able to resume the download.
According to this article, chrome should support something like this. I just need to be pointed in the right direction.
Ty!
Chrome needs some way to know that the file it's trying to download at that URL is indeed the same file when it tries to resume.
If you add support for an ETag header, this will likely work.
I have a server using the google Drive API. I tried with a curl PUT request to upload a simple file (test.txt) at http://myserver/test.txt. As you can see, I did the PUT request at the root of my server. The response I get is the following:
HTTP/1.1 200 OK
X-GUploader-UploadID: AEnB2UqANa4Bj6ilL7z5HZH0wlQi_ufxDiHPtb2zq1Gzcx7IxAEcOt-AOlWsbX1q_lsZUwWt_hyKOA3weAeVpQvPQTwbQhLhIA
ETag: "6e809cbda0732ac4845916a59016f954"
x-goog-generation: 1548877817413782
x-goog-metageneration: 1
x-goog-hash: crc32c=jwfJwA==
x-goog-hash: md5=boCcvaBzKsSEWRalkBb5VA==
x-goog-stored-content-length: 6
x-goog-stored-content-encoding: identity
Content-Type: text/html; charset=UTF-8
Accept-Ranges: bytes
Via: 1.1 varnish
Content-Length: 0
Accept-Ranges: bytes
Date: Wed, 30 Jan 2019 19:50:17 GMT
Via: 1.1 varnish
Connection: close
X-Served-By: cache-bwi5139-BWI, cache-cdg20732-CDG
X-Cache: MISS, MISS
X-Cache-Hits: 0, 0
X-Timer: S1548877817.232336,VS0,VE241
Vary: Origin
Access-Control-Allow-Methods: POST,PUT,PATCH,GET,DELETE,OPTIONS
Access-Control-Allow-Headers: Cache-Control,X-Requested-With,Authorization,Content-Type,Location,Range
Access-Control-Allow-Credentials: true
Access-Control-Max-Age: 300
I know you're not supposed to use the API that way. I did that for testing purposes. I understand every headers returned but can't figure out if my file has been uploaded because I don't have enough knowledge of this API.
My question is very simple :
Just by looking at the response, can you tell me if my file has been uploaded ?
If yes can I retrieve it and how ?
The HTTP status code traditionally indicates, for any given request, if it was successful. The status code in the response is always on the first line:
HTTP/1.1 200 OK
200 type status codes mean success. You should take some time to familiarize yourself with HTTP status codes if you intend to work with HTTP APIs.
Edit: it seems the issue was caused by a dropped cookie. There should have been a session id cookie as well.
For posterity, here's the original question
When sending a request formed as this
GET https://<url>?<parameters>
Cache-Control: max-age=0
Connection: keep-alive
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Charset: iso-8859-1,utf-8,UTF-8
Accept-Encoding: gzip, x-gzip, deflate, x-bzip2
Accept-Language: en-US,en;q=0.5
If-None-Match: "6eb7d55abfd0546399e3245ad3a76090"
User-Agent: Mozilla/5.0 libwww-perl/6.13
Cookie: auth_token=<blah>; __cfduid=<blah>
Cookie2: $Version="1"
I receive the following response
response-type: text/html
charset=utf-8
HTTP/1.1 406 Not Acceptable
Cache-Control: no-cache
Connection: keep-alive
Date: Fri, 12 Feb 2016 18:34:00 GMT
Server: cloudflare-nginx
Content-Type: text/html; charset=utf-8
CF-RAY: 273a62969a9b288e-SJC
Client-Date: Fri, 12 Feb 2016 18:34:00 GMT
Client-Peer: <IP4>:443
Client-Response-Num: 10
Client-SSL-Cert-Issuer: /C=GB/ST=Greater Manchester/L=Salford/O=COMODO CA Limite
d/CN=COMODO ECC Domain Validation Secure Server CA 2
Client-SSL-Cert-Subject: /OU=Domain Control Validated/OU=PositiveSSL Multi-Domai
n/CN=ssl<blah>.cloudflaressl.com
Client-SSL-Cipher: <some value>
Client-SSL-Socket-Class: IO::Socket::SSL
Client-SSL-Warning: Peer certificate not verified
Client-Transfer-Encoding: chunked
Status: 406 Not Acceptable
X-Runtime: 9
I'm not entirely sure why the response is 406 Not Acceptable. When
downloaded with firefox, the file in question in 996 KB (as reported
by Windows 8.1's explorer). It looks like I have a partially
transferred file from my perl script at 991 KB (again, windows
explorer size), so it got MOST of the file before throwing the Not
Acceptable response. Using the same URL pattern and request style, I
was able to successfully download a 36 MB file from the server with
this perl library and request form, so the size of the file should not
be magically past some max (chunk) size. As these files are being
updated on approximately 15-minute intervals, I suppose it's possible
that a write was performed on the server, invalidating the ETag before
all chunks were complete on this file?
I tried adding chunked to Accept-Encoding, but that's not for
Transfer encoding and it appears to have no affect on the server's behavior. Additionally, as I've been able to download larger files
(same format) from the same server, that alone shouldn't be the cause
of my woes. LWP is supposed to be able to handle chunked data
returned by a response to GET (as per this newsgroup post).
The server in question is running nginx with Rack::Lint. The
particular server configuration (which I in no way control), throws
500 errors on its own attempts to send 304: not modified. This
caused me to write a workaround for File::Download (sub
lintWorkAround here), so I'm not above putting blame on the
server in this instance also, if warranted. I don't believe I buggered
up the chunk-handling code from File::Download 0.3 (see diff),
but I suppose that's also possible. Is it possible to request a
particular chunk size from the server?
I'm using LWP and libwww versions 6.13 in perl 5.18.2.
File::Download version is my own 0.4_050601.
So, what else could the 406 error mean? Is there a way to request that
the server temporarily cache/version control the entire file so that I
can download a given ETag'd file once the transfer begins?
I'm trying to upload a file to Skydrive where I don't a-priori know the Content-Length. With other storage services I can do this with chunked HTTP upload, but Skydrive always complains about Content-Length being invalid.
Here are the full headers I'm sending:
PUT /v5.0/me/skydrive/files/skydrive_test.js?overwrite=ChooseNewName HTTP/1.1
Authorization: Bearer <TOKEN_REDACTED>
host: apis.live.net
content-type: application/javascript
Connection: keep-alive
Transfer-Encoding: chunked
Here's the response I get back:
cache-control: private, no-cache, no-store, must-revalidate
transfer-encoding: chunked
content-type: application/json; charset=UTF-8
server: Live-API/16.4.1731.327 Microsoft-HTTPAPI/2.0
x-content-type-options: nosniff
x-http-live-request-id: API.c6afda25-2d9f-4248-9f49-001ccb3a9007
x-http-live-server: BAYMSG1010836
date: Wed, 15 May 2013 14:33:00 GMT
{ "error": { "code": "request_invalid_content_length",
"message": "The value for the Content-Length header isn't valid." }}
Is there any way I can do this without setting Content-Length (i.e. using chunked encoding)?
I'm using node.js to do this, but it should apply equally with any language using the REST API, hence I haven't tagged this with a particular language.
For example Dropbox offers the Chunked Upload command: https://www.dropbox.com/developers/core/docs#chunked-upload
And Google Drive, even though it says it wants Content-Length, doesn't need it for it's resumable upload API: https://developers.google.com/drive/manage-uploads#resumable
Is there an API I'm missing?
Edit: Things I've tried: Setting Content-Length: 0 results in it working, but the file is zero bytes. Setting Content-length:0 and Transfer-Encoding: chunked, results in the original error above.
Try setting a dummy content-length to see if it is acceptable. Otherwise set the file size.
If you are using node.js you can get the file size you are trying to upload and set the size to content-length in OCTETs. You can get the file size requiring the fs (filesystem) module in node.
var fs = require('fs');
fs.watchFile('some.file', function () {
fs.stat('some.file', function (err, stats) {
console.log(stats.size);
});
});
#Resource
I'm trying to implement file upload functionality in the iPhone app. Server code is tested and works when files are uploaded from the desktop browser, so I moved to implementing the Objective-C client code. I'm assembling HTTP requests body manually, and despite that it looks correct, it is rejected by the server (server handler unable to extract the parts from multipart content). In desperation I've simplified the form to having only one parameter, but it still does not work.
I've captured the network traffic and I could see that Wireshark could not parse my multipart content as well (have a look at screenshots: Firefox request, iPhone request). I'm pasting it below in hope that you could see the errors I can't see.
Thanks in advance.
Firefox:
POST /cubepaint/actions/gallery/post HTTP/1.1
Host: [...]
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-GB; rv:1.9.1.3) Gecko/20090824 Firefox/3.5.3
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language: en-gb,en;q=0.5
Accept-Encoding: gzip,deflate
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive: 300
Connection: keep-alive
Authorization: Basic [...]
Content-Type: multipart/form-data; boundary=---------------------------20072377098235644401115438165
Content-Length: 180
-----------------------------20072377098235644401115438165
Content-Disposition: form-data; name="deviceId"
12345
-----------------------------20072377098235644401115438165--
HTTP/1.1 200 OK
Date: Sat, 17 Oct 2009 22:09:21 GMT
Server: Apache/2.2.3 (Debian) DAV/2 SVN/1.4.2 mod_python/3.2.10 Python/2.4.4 mod_ssl/2.2.3 OpenSSL/0.9.8c
Keep-Alive: timeout=15, max=100
Connection: Keep-Alive
Transfer-Encoding: chunked
Content-Type: text/html; charset=UTF-8
iPhone:
POST /cubepaint/actions/gallery/post HTTP/1.1
Host: [...]
User-Agent: Copenhagen/1.0 CFNetwork/459 Darwin/9.8.0
Content-Type: multipart/form-data; boundary=----------0E7B16E6-CD3D-4213-9B42-07DA30822C74
Accept: */*
Accept-Language: en-us
Accept-Encoding: gzip, deflate
Authorization: Basic [...]
Content-Length: 187
Connection: keep-alive
----------0E7B16E6-CD3D-4213-9B42-07DA30822C74
Content-Disposition: form-data; name="deviceId"
00000000-0000-1000-8000-0016CBCC0B61
----------0E7B16E6-CD3D-4213-9B42-07DA30822C74--
HTTP/1.1 200 OK
Date: Sat, 17 Oct 2009 22:04:07 GMT
Server: Apache/2.2.3 (Debian) DAV/2 SVN/1.4.2 mod_python/3.2.10 Python/2.4.4 mod_ssl/2.2.3 OpenSSL/0.9.8c
Keep-Alive: timeout=15, max=100
Connection: Keep-Alive
Transfer-Encoding: chunked
Content-Type: text/html; charset=UTF-8
Your iPhone version indicates keep-alive but doesn't specify a length. Not sure that's enough to cause trouble.
Also, is it possible your server is checking for user-agent strings it recognizes (say, for backward-compatibility mode)?
I'd also compare the two in a text editor that shows CR/LF characters to make sure you're getting proper line endings.
Another thing you could try is create a simple web-page that does a multipart POST and run it from the iPhone browser (instead of the Mac one) then check the headers that go across the wire. Or you could snag a toolkit like ASIHTTPRequest and see what kind of output it generates for multi-part posts (or just use the toolkit instead of trying to write your own).
Good luck
Solved by reading RFC 2046 (MIME specification): boundary between parts of multipart message should contain two leading '-'s, and last boundary should additionally contain two trailing '-'s. The boundary in the request header and request body in the Firefox request differ:
---------------------------20072377098235644401115438165
and
-----------------------------20072377098235644401115438165
The last boundary looks like this:
-----------------------------20072377098235644401115438165--
You really could not see this with the eye when there are so many leading '-'s in the original boundary.