X
點擊此處開啟此網站的行動版。

技術支援討論區

Firefox 49 won't start after installation

已張貼

OS : Windows 7 x64 I was using Firefox 48.0.2 (x64) - french version. I manually downloaded and installed Firefox 49.0 (x64). After installation, Firefox does not start : - the firefox.exe process is running, using something between 5000 and 65000 Mb of memory - nothing else happens : the process does nothing, using 0% CPU

I've seen this issue : - on several PC's of my organization - after automatic update or manual update - after having deleted my Firefox profile - after having disabled Kasperky Antivirus

I've noticed that the last version of Firefox Developer Edition has the same behaviour : install OK, won't start.

Going back to previous version is OK : Firefox starts again.

OS : Windows 7 x64 I was using Firefox 48.0.2 (x64) - french version. I manually downloaded and installed Firefox 49.0 (x64). After installation, Firefox does not start : - the firefox.exe process is running, using something between 5000 and 65000 Mb of memory - nothing else happens : the process does nothing, using 0% CPU I've seen this issue : - on several PC's of my organization - after automatic update or manual update - after having deleted my Firefox profile - after having disabled Kasperky Antivirus I've noticed that the last version of Firefox Developer Edition has the same behaviour : install OK, won't start. Going back to previous version is OK : Firefox starts again.

額外的系統細節

已安裝的外掛程式

  • Shockwave Flash 23.0 r0

應用程式

  • Firefox 48.0.2
  • 使用者代理:Mozilla/5.0 (Windows NT 6.0; Win64; x64; rv:48.0) Gecko/20100101 Firefox/48.0
  • 技術支援網址:https://support.mozilla.org/1/firefox/48.0.2/WINNT/fr/

擴充套件

  • BehindTheOverlay 0.1.1.1-signed.1-signed (jid1-Y3WfE7td45aWDw@jetpack)
  • Firefox Hello Beta 1.4.4 (loop@mozilla.org)
  • Firefox Hotfix 20160826.01 (firefox-hotfix@mozilla.org)
  • I don't care about cookies 2.6.0 (jid1-KKzOGWgsW3Ao4Q@jetpack)
  • LeechBlock 1.0.5 ({a95d8332-e4b4-6e7f-98ac-20b733364387})
  • MeasureIt 0.4.15 ({75CEEE46-9B64-46f8-94BF-54012DE155F0})
  • Multi-process staged rollout 1.2 (e10srollout@mozilla.org)
  • Pocket 1.0.4 (firefox@getpocket.com)
  • Priv8 0.1.1 (id@baku.priv8)
  • Stop AutoPlay Next for YouTube™ 0.1.2.1 (jid1-nbzaq9ol2hyY64gGe@jetpack)
  • uBlock Origin 1.9.6 (uBlock0@raymondhill.net)

JavaScript

  • incrementalGCEnabled: True

圖形

  • adapterDescription: Intel(R) HD Graphics 4600
  • adapterDescription2:
  • adapterDeviceID: 0x0412
  • adapterDeviceID2:
  • adapterDrivers: igdumdim64 igd10iumd64 igd10iumd64 igdumdim32 igd10iumd32 igd10iumd32
  • adapterDrivers2:
  • adapterRAM: Unknown
  • adapterRAM2:
  • adapterSubsysID: 06121028
  • adapterSubsysID2:
  • adapterVendorID: 0x8086
  • adapterVendorID2:
  • direct2DEnabled: True
  • directWriteEnabled: True
  • directWriteVersion: 6.2.9200.17568
  • driverDate: 3-16-2015
  • driverDate2:
  • driverVersion: 10.18.14.4170
  • driverVersion2:
  • info: {u'AzureCanvasAccelerated': 0, u'AzureCanvasBackend': u'direct2d 1.1', u'AzureFallbackCanvasBackend': u'cairo', u'AzureContentBackend': u'direct2d 1.1'}
  • isGPU2Active: False
  • numAcceleratedWindows: 1
  • numTotalWindows: 1
  • supportsHardwareH264: Yes; Using D3D9 API
  • webglRenderer: Google Inc. -- ANGLE (Intel(R) HD Graphics 4600 Direct3D11 vs_5_0 ps_5_0)
  • windowLayerManagerRemote: True
  • windowLayerManagerType: Direct3D 11

修改過的偏好設定

其他

  • User JS: 否
  • 輔助功能: 否
philipp
  • Top 25 Contributor
  • Moderator
5287 個解決方法 23362 個答案

many thanks again, so this testing narrows it down to one single code-change that has apparently introduced this issue. i have put that info in the associated bug report and hope that it is now possible for developers to come up with a fix soon.

many thanks again, so this testing narrows it down to one single code-change that has apparently introduced this issue. i have put that info in the associated bug report and hope that it is now possible for developers to come up with a fix soon.

提出問題者

Thanks Philipp, tell me if you need more information or build testing.

Thanks Philipp, tell me if you need more information or build testing.
philipp
  • Top 25 Contributor
  • Moderator
5287 個解決方法 23362 個答案

the issue is quite puzzling. in the stacktrace log that we got, the networking team identified a problem with a hanging event loop that could totally explain the symptoms you were seeing - however the build-by-build testing that we did here in this support.mozilla.org thread with your help points to a totally different code-change that first introduced the problem of firefox not starting.

so now we have two avenues to explore: first we manually created a build based on the current beta branch with just the one code-change backed out (that's the one we suspect based on the manual regression testing). it would be https://archive.mozilla.org/pub/firefox/try-builds/archaeopteryx@coole-files.de-73d579235407a1cd599845897d18cea974fb5a50/try-win32/firefox-50.0.en-US.win32.installer.exe & i'd be glad if you can test if it's opening/connecting again...

in parallel the networking team is patching the problem they see in the log and might come up with another try build containing a fix for that seperate issue a little bit later today.

the issue is quite puzzling. in the stacktrace log that we got, the networking team identified a problem with a hanging event loop that could totally explain the symptoms you were seeing - however the build-by-build testing that we did here in this support.mozilla.org thread with your help points to a totally different code-change that first introduced the problem of firefox not starting. so now we have two avenues to explore: first we manually created a build based on the current beta branch with just the one code-change backed out (that's the one we suspect based on the manual regression testing). it would be https://archive.mozilla.org/pub/firefox/try-builds/archaeopteryx@coole-files.de-73d579235407a1cd599845897d18cea974fb5a50/try-win32/firefox-50.0.en-US.win32.installer.exe & i'd be glad if you can test if it's opening/connecting again... in parallel the networking team is patching the problem they see in the log and might come up with another try build containing a fix for that seperate issue a little bit later today.

提出問題者

Philipp : same symptom here, no Firefox UI, process running.

Philipp : same symptom here, no Firefox UI, process running.
philipp
  • Top 25 Contributor
  • Moderator
5287 個解決方法 23362 個答案

a build with the presumptive networking fix just arrived - do we get lucky this time (does it launch/connect)?: https://archive.mozilla.org/pub/firefox/try-builds/dd.mozilla@gmail.com-51836a9c48ad0fb657dadcc3f79444ea0d73425f/try-win32/firefox-52.0a1.en-US.win32.installer.exe

a build with the presumptive networking fix just arrived - do we get lucky this time (does it launch/connect)?: https://archive.mozilla.org/pub/firefox/try-builds/dd.mozilla@gmail.com-51836a9c48ad0fb657dadcc3f79444ea0d73425f/try-win32/firefox-52.0a1.en-US.win32.installer.exe

有幫助的回覆

Good news Philipp : this build works fine : Firefox UI OK + connectivity OK. Well done ! Can we expect a quick fix ?

Good news Philipp : this build works fine : Firefox UI OK + connectivity OK. Well done ! Can we expect a quick fix ?
kevinherrmann 0 個解決方法 9 個答案

The release notes claim to have fixed the start issue in 49.0.2. I installed this over 48 again and, so far, 49.0.2 does seem to be OK and does not stall on start up. However, the other issue I reported, that it breaks the HTML Validator plugin on startup, makes the C++ library inaccessible and causes it to route to an error page remains. I put blank lines either side of the relevant lines in the dump and added the error at the top. Something seems to have been changed in Firefox access to C++ that has not been fixed as well as the networking issue between 48 and 49 that has been fixed. Nothing else in my setup has changed and this issue is fixed by reverting to 48 again, so definitely a 49 issue still. Whether Firefox or the plugin or both need fixing is another matter.

The release notes claim to have fixed the start issue in 49.0.2. I installed this over 48 again and, so far, 49.0.2 does seem to be OK and does not stall on start up. However, the other issue I reported, that it breaks the HTML Validator plugin on startup, makes the C++ library inaccessible and causes it to route to an error page remains. I put blank lines either side of the relevant lines in the dump and added the error at the top. Something seems to have been changed in Firefox access to C++ that has not been fixed as well as the networking issue between 48 and 49 that has been fixed. Nothing else in my setup has changed and this issue is fixed by reverting to 48 again, so definitely a 49 issue still. Whether Firefox or the plugin or both need fixing is another matter.
Truckerman 1 個解決方法 6 個答案

Yesterday I updated Firefox to version 49 via the Help about Firefox button, where it allows one to check for updates. That resulted in 5 hours of forcibly removing Firefox , re-installing Firefox, If I had not backed up the profile, deleted the entire install directory where Firefox is located I would still be having issues. Note, I finally had to get a non-corrupted download of Firefox off the web, not from Mozilla's website to fix the problems. There is documentation somewhere that indicates to try running in safe mode. Apparently that option is no longer available for version 49. It seems that what is downloaded via updates, is a corrupt file. If and when there is a resolution, please advise.

Yesterday I updated Firefox to version 49 via the Help about Firefox button, where it allows one to check for updates. That resulted in 5 hours of forcibly removing Firefox , re-installing Firefox, If I had not backed up the profile, deleted the entire install directory where Firefox is located I would still be having issues. Note, I finally had to get a non-corrupted download of Firefox off the web, not from Mozilla's website to fix the problems. There is documentation somewhere that indicates to try running in safe mode. Apparently that option is no longer available for version 49. It seems that what is downloaded via updates, is a corrupt file. If and when there is a resolution, please advise.