User Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:33.0) Gecko/20100101 Firefox/33.0 (Beta/Release) Build ID: 20140710030200 Steps to reproduce: 1. Start FF Nightly 33.0a1 (2014-07-10) [no addons] 2. Open webpage can reproduce on following pages: www.anandtech.com www.techspot.com www.heise.de www.tomshardware.com -> Favicon in tab is shown 3. Right Click -> Reload -> Favicon vanishes from tab (tab gets smaller) or 3. Open new tab, open same page -> Favicon is not shown P.S. www.tomshardware.com is especially confusing, 1st open = normal, 2nd open/reload = shows a differend icon, 3rd open/reload = shows no icon anymore Actual results: The Favicon is only shown once after restart. After a reload or reopen the Favicon is not shown any more. Expected results: The Favicon should be displayed.

I can't reproduce this. Can you reproduce on a clean profile?

Ok, I can reproduce on a fresh install (Nightly) with following settings: do not track : enabled override automatic cache management: 0 mb [to explain, I use a Firefox Portable, so I reenabled those settings that could be the culprid] able to reproduce on heise, tomshardware, techspot not able to reproduce on anandtech

cannot reproduce Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:33.0) Gecko/20100101 Firefox/33.0 Build ID: 20140713030204 [testday-20140714]

(In reply to Jörn Albert from comment #2) > Ok, I can reproduce on a fresh install (Nightly) Did you try http://support.mozilla.org/en-US/kb/Managing-profiles#w_starting-the-profile-manager ? I couldn't also reproduce on 33.0a1 (2014-07-20), Win 7 x64

I was able to reproduce this on latest Nightly (2014-07-20), Windows 7 32bit and 64bit. Also this seems to be a regression: Last good revision: 58c5a3427997 (2014-05-16) First bad revision: 2893f60d5903 (2014-05-17) Pushlog: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=58c5a3427997&tocha nge=2893f60d5903 Maybe from this: fb6ced6ea7c9 Margaret Leibovic — bug 1010237 - Remove unused data sent with Link:Favicon message. r=bnicholson

Adding a gif to show the issue.

(In reply to Bogdan Maris, QA [:bogdan_maris] from comment #5) > I was able to reproduce this on latest Nightly (2014-07-20), Windows 7 32bit > and 64bit. > Also this seems to be a regression: > > Last good revision: 58c5a3427997 (2014-05-16) > First bad revision: 2893f60d5903 (2014-05-17) > Pushlog: > https://hg.mozilla.org/mozilla-central/ > pushloghtml?fromchange=58c5a3427997&tocha > nge=2893f60d5903 > > Maybe from this: > fb6ced6ea7c9 Margaret Leibovic — bug 1010237 - Remove unused data sent with > Link:Favicon message. r=bnicholson If this is a bug in desktop Firefox, there is no way that this changeset is responsible for the regression. This patch only affects Firefox for Android.

(In reply to :Margaret Leibovic from comment #7) > (In reply to Bogdan Maris, QA [:bogdan_maris] from comment #5) > > I was able to reproduce this on latest Nightly (2014-07-20), Windows 7 32bit > > and 64bit. > > Also this seems to be a regression: > > > > Last good revision: 58c5a3427997 (2014-05-16) > > First bad revision: 2893f60d5903 (2014-05-17) > > Pushlog: > > https://hg.mozilla.org/mozilla-central/ > > pushloghtml?fromchange=58c5a3427997&tocha > > nge=2893f60d5903 > > > > Maybe from this: > > fb6ced6ea7c9 Margaret Leibovic — bug 1010237 - Remove unused data sent with > > Link:Favicon message. r=bnicholson > > If this is a bug in desktop Firefox, there is no way that this changeset is > responsible for the regression. This patch only affects Firefox for Android. Yes. But http cache v2 was turned on, and that seems like a more likely culprit here. Honza?

Just try to reproduce with cache2 off by setting prefs: browser.cache.use_new_backend = 0 browser.cache.use_new_backend_temp = false

(In reply to Honza Bambas (:mayhemer) from comment #9) > Just try to reproduce with cache2 off by setting prefs: > > browser.cache.use_new_backend = 0 > browser.cache.use_new_backend_temp = false Bogdan, as you could reproduce this... can you check this is at issue here and set tracking/status flags accordingly? Thank you!

(In reply to :Gijs Kruitbosch from comment #10) > (In reply to Honza Bambas (:mayhemer) from comment #9) > > Just try to reproduce with cache2 off by setting prefs: > > > > browser.cache.use_new_backend = 0 > > browser.cache.use_new_backend_temp = false > > Bogdan, as you could reproduce this... can you check this is at issue here > and set tracking/status flags accordingly? Thank you! Sure thing. Disabling cache2 I was not able to reproduce this issue anymore on latest Aurora and Nightly.

[Tracking Requested - why for this release]: because a regression, and favicon-less tabs and bookmarks make it harder to find what you want - pretty visible, too. Honza, can you investigate this further?

Please confirm you can reproduce with a clean profile ("fresh install" doesn't mean a clean profile!). Thanks.

Cannot reproduce.

(In reply to Honza Bambas (:mayhemer) from comment #13) > Please confirm you can reproduce with a clean profile ("fresh install" > doesn't mean a clean profile!). Thanks. I confirm that I can reproduce with a clean profile.

Tracking, for the reason Gijs stated.

(In reply to Bogdan Maris, QA [:bogdan_maris] from comment #15) > (In reply to Honza Bambas (:mayhemer) from comment #13) > > Please confirm you can reproduce with a clean profile ("fresh install" > > doesn't mean a clean profile!). Thanks. > > I confirm that I can reproduce with a clean profile. OK, since I cannot reproduce, I will ask for an NSPR log: NSPR_LOG_MODULES=timestamp,nsHttp:4,cache2:5 NSPR_LOG_FILE=path/to/log Thanks.

(In reply to Honza Bambas (:mayhemer) from comment #17) > (In reply to Bogdan Maris, QA [:bogdan_maris] from comment #15) > > (In reply to Honza Bambas (:mayhemer) from comment #13) > > > Please confirm you can reproduce with a clean profile ("fresh install" > > > doesn't mean a clean profile!). Thanks. > > > > I confirm that I can reproduce with a clean profile. > > OK, since I cannot reproduce, I will ask for an NSPR log: > > NSPR_LOG_MODULES=timestamp,nsHttp:4,cache2:5 > NSPR_LOG_FILE=path/to/log > > Thanks. Sure, here you go https://db.tt/esSioXQB. If there is anything else I can help with please let me know.

It shows up the problem is related to bug 1064408 but slightly different. This is more similar to bug 1048348, at least the solution is based on the same fix.