summaryrefslogtreecommitdiff
path: root/hurd/translator/httpfs.mdwn
diff options
context:
space:
mode:
Diffstat (limited to 'hurd/translator/httpfs.mdwn')
-rw-r--r--hurd/translator/httpfs.mdwn100
1 files changed, 100 insertions, 0 deletions
diff --git a/hurd/translator/httpfs.mdwn b/hurd/translator/httpfs.mdwn
new file mode 100644
index 0000000..dc4a62f
--- /dev/null
+++ b/hurd/translator/httpfs.mdwn
@@ -0,0 +1,100 @@
+[[!meta copyright="Copyright © 2013 Free Software Foundation, Inc."]]
+
+[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable
+id="license" text="Permission is granted to copy, distribute and/or modify this
+document under the terms of the GNU Free Documentation License, Version 1.2 or
+any later version published by the Free Software Foundation; with no Invariant
+Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license
+is included in the section entitled [[GNU Free Documentation
+License|/fdl]]."]]"""]]
+
+`httpfs` is a virtual filesystem allowing you to access web pages and files.
+
+
+# Source
+
+<http://www.nongnu.org/hurdextras/#httpfs>
+
+
+# Documentation
+
+## IRC, freenode, #hurd, 2013-09-03
+
+[[!tag open_issue_documentation]]
+
+ <congzhang> hi, why I can't cd to /http:/richtlijn.be/~larstiq/hurd/ to do
+ <congzhang> grep?
+ <pinotree> this is not ftp
+ <congzhang> it works for other file
+ <pinotree> ?
+ <congzhang> I can't cd to ~larstiq, I don't know why
+ <pinotree> http is not a filesystem protocol
+ <pinotree> while httpfs could try in representing it as such, it is not
+ something reliable
+ <congzhang> ok, it's not reliable
+ <congzhang> I expect it can expose dir like browser
+ <congzhang> so, the translator just know href from home page, and one by
+ one
+ <pinotree> uh?
+ <congzhang> if ...:80/a/b/c.png exits, but not has a href in homepage, so I
+ can't cd to a, right?
+ <pinotree> you are looking things from the wrong point of view
+ <pinotree> a web server can do anything with URLs, including redirecting,
+ URL rewriting and whatever else
+ <congzhang> so, how to understand httpfs's idea?
+ <congzhang> how httpfs list dir?
+ <pinotree> check its code
+ <congzhang> en, no need it's not reliable
+ <congzhang> it's not work, it's enough
+ <congzhang> I have an idea, for the file system, we explore dir level by
+ level, but for http, we change full path one
+ <congzhang> once time
+ <congzhang> maybe can allow user to cd any directory, and just mark as some
+ special color to make user know the translator was not sure, file exist
+ or not
+ <congzhang> once the file exits, mark all the parent directory as normal
+ color?
+ <rekado> congzhang: you can find more info about httpfs here:
+ http://nongnu.org/hurdextras/
+ <pinotree> congzhang: you're still looking at http from the wrong point of
+ view
+ <pinotree> there are no directories nor files
+ <pinotree> you start a request for a URL, and you get some content back (in
+ case there's no error)
+ <congzhang> you mean httpfs just for kidding?
+ <pinotree> that the content is a web page listing a directory on the
+ filesystem of the web server machine, or a file sent back via the
+ connection, or a complex web page, it's the same
+ <rekado> congzhang: you can only get a list of files if the web server
+ responds with an index of files
+ <pinotree> "files"
+ <rekado> The readme explains how httpfs does its thing:
+ http://cvs.savannah.gnu.org/viewvc/*checkout*/httpfs/README?root=hurdextras
+ <congzhang> if I can't cd to /http:/host/a/b how to get
+ /http:/host/a/b/c.html, even the file exist?
+ <pinotree> you don't cd in http
+ <pinotree> cd is for changing directory, which does not apply to a protocol
+ like http which is not fs-oriented
+ <congzhang> yes, I agree with you, http was not fs-oriented
+ <congzhang> so httpfs was not so useful
+ <rekado> You can access the document directly, though, can't you?
+ <congzhang> rekado: I try once more
+ <congzhang> I can't
+ <congzhang> so, the httpfs need some extend, http protocol was not fs
+ oriented, so need some extend to make it work with file system
+ <pinotree> http is not designed for file system usage, so extending it is
+ useless
+ <congzhang> or, httpfs was not so useful
+ <pinotree> there are many other protocols for file systems
+ <congzhang> I don't think so
+ <pinotree> i do
+ <congzhang> if we can't make it more useful, remove it from hurd rep, or
+ extend it more useful
+ <congzhang> add some more rule, to make it work with file system
+ <pinotree> no
+ <congzhang> some paradox in it
+ <pinotree> which paradox?
+ <congzhang> for http vs file system
+ <pinotree> ???
+ <congzhang> tree oriented and star topology oriented?
+ <pinotree> you don't make any sense