Closed Bug 120033 Opened 23 years ago Closed 23 years ago

gzipped file when transfer encoding is gzip.

Categories

(Core Graveyard :: File Handling, defect)

defect
Not set
major

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: yasuo_ohgaki, Assigned: law)

References

()

Details

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.7+) Gecko/20020111
BuildID:    2002011108

Mozilla saves file as gzipped file when transfer encoding is gzip.

Reproducible: Always
Steps to Reproduce:
1. Open the URL and locate some patch/file to download
2. Download file
3. execute "file <downloaded-file>"

Actual Results:  File is gzipped.

Expected Results:  Decoded text should be saved just like IE does.

I'm not saying decode gzipped file, but decode gzip transfer encoding :)
file handling...?
Assignee: trudelle → law
Component: XP Apps → File Handling
Yasuo, how are you downloading the file/patch?
If you see this page, you'll see "Download" links at right hand side.
http://cvs.php.net/cvs.php/php4/ext/pgsql/pgsql.c
Confirming.  Clicking on the link and then selecting "save" from the resulting
dialog does not decode based on the content-encoding.  This is distinct from
doing right click and then save as, which uses the persistance object....
Status: UNCONFIRMED → NEW
Ever confirmed: true
OS: Linux → All
Hardware: PC → All
Note that the gzip is a _content_ encoding, not a _transfer_ encoding.  So
arguably we should _not_ be unzipping....
Boris,
Are you saying the file(content) is actually gzipped?
If so, I have to complain about it to horde guys :)
They said files are not gzipped.

IEs decodes files from cvs.php.net, while gzipped files remain gzipped, though...


> Are you saying the file(content) is actually gzipped?

I'm saying that servers send the _same_ headers in both cases, unfortunately...
so we have to try to guess whether it's "actually gzipped" or "just
content-encoded"
Then I'll take a look at server settings, actual headers and code when I have time.
*** Bug 123192 has been marked as a duplicate of this bug. ***
I haven't spent time to take a look at actual packets, yet.
I noticed Mozilla decodes zgip/bz2 files in some cases.

Try http://snaps.php.net/


 
This makes life difficult for me.

http://hosted.barrysworld.net/basiliskii/rail/photos.html is where I am usually
bitten by it...
From comments in this bug, it sounds like it's a problem with content-encoding 
rather than transfer-encoding.  Update summary?
When saving a file Mozilla obviously forgets to parse the received HTTP header
for the presence of the "Content-Encoding:" header (and to decode it before
saving if necessary) as specified in RFC 2616 section 14.11 (HTTP 1.1) and RFC
1945 section 10.3 (HTTP 1.0). This is also necessary if the file is in the cache
and Mozilla is offline!

Here are some wget logfiles for
http://hosted.barrysworld.net/basiliskii/rail/photos.html:

without the accept-content header:
---request begin---
GET /basiliskii/rail/photos.html HTTP/1.0
User-Agent: Wget/1.7
Host: hosted.barrysworld.net
Accept: */*
Connection: Keep-Alive

---request end---
HTTP request sent, awaiting response... HTTP/1.1 200 OK
Date: Sun, 17 Feb 2002 08:26:56 GMT
Server: Apache
Last-Modified: Thu, 07 Feb 2002 19:43:16 GMT
ETag: "3c199-1805-3c62d8d4"
Accept-Ranges: bytes
Content-Length: 6149
Connection: close
Content-Type: text/html


Length: 6,149 [text/html]

and now with the header:
---request begin---
GET /basiliskii/rail/photos.html HTTP/1.0
User-Agent: Wget/1.7
Host: hosted.barrysworld.net
Accept: */*
Connection: Keep-Alive
Accept-Encoding: gzip, deflate, compress;q=0.9

---request end---
HTTP request sent, awaiting response... HTTP/1.1 200 OK
Date: Sun, 17 Feb 2002 08:27:26 GMT
Server: Apache
Last-Modified: Thu, 07 Feb 2002 19:43:16 GMT
ETag: "3c199-1805-3c62d8d4"
Accept-Ranges: bytes
Connection: close
Content-Type: text/html
Content-Encoding: gzip
Content-Length: 2057


Length: 2,057 [text/html]
*** Bug 126007 has been marked as a duplicate of this bug. ***
> and is only decoded before rendering or analogous usage.

It's not clear that saving is analogous usage.  The problem with content
encoding is that the encoding is an integral part of the data, unlike with
transfer encoding.  So we _should_ save with the content encoding.  Servers
often use content encoding when they mean transfer encoding, however.  Please
see my comments in bug 121001

For example, type application/tar content-encoding gzip should not be
decompressed before saving.  IE gets this wrong because it always decodes.

All that said, I am working on a possible fix.
Depends on: 69306
fixed by checkin for bug 69306
Status: NEW → RESOLVED
Closed: 23 years ago
Resolution: --- → FIXED
QA Contact: sairuh → tever
*** Bug 127954 has been marked as a duplicate of this bug. ***
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.