198252 - Networking Preferences - index of preferences

Status: RESOLVED FIXED

(Developer Documentation Graveyard :: General, defect, P1)

Description

[benc]

I'm going to put a master document for networking related-prefs. I don't
actually QA all these features, so I am not going to cover everything on the
list, but I will own this document, and keep the links accurate.

Comment 1

[benc]

I just finished documenting most of the prefs I cover as QA.

http://www.mozilla.org/quality/networking/docs/netprefs.html

At some later date, hopefully via people suggesting itesm that I do not QA, this
list could encompass all networking behaviors, not just my assigned areas.

Comment 2

[benc]

I've added all the missing misc. stuff + the beginnings of the HTTP write-up.

Comment 3

[bomfog]

Attachment: Patch for a few typo-type nits.

Comment 4

[Daniel Wang]

Comment on attachment 125989 [details] [diff] [review]
Patch for a few typo-type nits.

Warner's patch in

Comment 5

[benc]

cc: darin as an FYI.

Thanks for the corrections!

The document still trails the trunk, but I've added partial info on
external-handler, as well as the https disk caching change darin made.

I'm going to give this extra attention in the future, because it is one of the
few really important network docs that is also a moving target.

Comment 6

[benc]

-> mine, P1

I know I'm behind on this, so anyone who spots a netpref that is not documented,
tell me here!

Comment 7

[benc]

Added info about the general offline/online prefs (nothing module-specific, like
MailNews).

Comment 8

[benc]

kerberos prefs! I'm behind on bugmail... bug 243850 

Comment 9

[OstGote!]

The description for the pref browser.cache.memory.capacity on
http://www.mozilla.org/quality/networking/docs/netprefs.html#cache is wrong. 
1) The values are in KiB not in B. See
https://bugzilla.mozilla.org/show_bug.cgi?id=105344.
2) Not any value. Pref not present or < 0 == use default, = 0 memory cache is
DISABLED, > 0 value is used. See
http://lxr.mozilla.org/mozilla/source/netwerk/cache/src/nsCacheService.cpp#1245. 

Comment 10

[Mark Nottingham]

Where possible, it would be nice to document what version prefs became avaialable in, as well as when defaults change. 

For example, the default for network.dnsCacheExpiration is listed as "60", but AFAICT by browsing bugzilla, it was infinite before ~2002, and 5 minutes until 1.6 (please correct me if I'm wrong).

Cheers,

Comment 11

[u7736]

network.http.keep-alive.timeout is documented as having a default of "30", but a clean install of FF (Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.8.0.4) Gecko/20060508 Firefox/1.5.0.4) has "300".

Comment 12

[Nickolay_Ponomarev]

http://kb.mozillazine.org/About:config_entries has a fuller list of preferences, suggest referring people to that doc instead.

Comment 13

[Nickolay_Ponomarev]

Didn't mean to reassign.. benc are you still working on this?

Comment 14

[Gabriel]

Hello,

I can't seem to find pref to set the proxy timeout. In FF 1.0.x you could set a proxy up and if it wasn't found it would switch to direct connection very fast (maybe 2-5 sec). Since FF 1.5.x the timeout has become very long, so I had to install an extension to switch from proxied to non-proxied mode. Is there a way to get the pre-1.5 mode/timeout back via a pref? This problem has gotten to be unbearable since 2.0.x since aperently FF tried to contact the proxy before the UI is up, so if I go from a proxied network to a non-proxied network (and vice-versa), I have to wait about 5 minute (a colleague says he times it and it took him 11 minutes) to launch FF. I tried to turn off the update feature, since I though it was it that tried to connect, but that changed nothing. If I launch FF without the network it starts right up.So to summarize my questions :

1) Is there a pref (or can one be added) to get the FF < 1.5 mode back, where if a proxy is set, if it can't be contacted then after a short timeout it switches to non-proxied mode. This would be nice so that an extention isn't needed to switch from proxied-to non-proxied mode and vice versa.

2) Is there a pref (or is it a bug?) so that FF doesn't try to connect to the internet before the UI is up, so that I don't have to wait more than 5 minutes for FF to launch. (I have 2 colleagues that have the same problem).

Thanks,
Gabriel

Comment 15

[Nickolay_Ponomarev]

(In reply to comment #14)
> Hello,
This post totally doesn't belong in this bug. Ask in the relevant forum/newsgroup/mailing list instead.

Comment 16

[benc]

Nickolay: I'm only making appearences when I am unemployed. So for another week or so, yes :)

Gabriel: I'm not aware of such a timeout. If you find a bug number or other reference, please email me, and we can see about documenting it.

If you want it added as a feature, ask in the group Nickolay mentioned.

I'm going to review more recent comments and try to do some updating tonight.

Comment 17

[benc]

Gabriel:

now that I think about this more, what your are describing is two problems:

1- ff starts up and sends an http request to the proxy. ff has no offline-online controls, but you can find that in the plug-icon in seamonkey.

2- how long does this take, to timeout? necko uses the system default. time your failure from the line command (date;telnet proxy port;date), and compare the times.

Comment 18

[benc]

Checking in mozilla-org/html/quality/networking/docs/netprefs.html;
/www/mozilla-org/html/quality/networking/docs/netprefs.html,v  <--  netprefs.html
new revision: 1.21; previous revision: 1.20
done

moved feedback link.
made dns entries separate from cache.
removed old cookie prefs. (pointer to mconners stuff, which is current.

Comment 19

[benc]

RESOLVED/fixed.

This was migrated to:

http://developer.mozilla.org/en/docs/Mozilla_Networking_Preferences#Audience

I'm porting the late changes, then I will hack the top of the page to make it historical.

http://kb.mozillazine.org/About:config_entries is a very impressive article. I don't have edit access there, so I'm going to re-think the content of the MDC document to avoid duplicating info.