You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@myfaces.apache.org by "Werner Punz (JIRA)" <de...@myfaces.apache.org> on 2010/08/04 12:58:16 UTC
[jira] Created: (MYFACES-2858) pointless oamsubmit inline rendering
pointless oamsubmit inline rendering
------------------------------------
Key: MYFACES-2858
URL: https://issues.apache.org/jira/browse/MYFACES-2858
Project: MyFaces Core
Issue Type: Bug
Affects Versions: 2.0.2-SNAPSHOT
Reporter: Werner Punz
We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
prevents that the affected scripts can be browser cached.
A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (MYFACES-2858) pointless oamsubmit inline
rendering
Posted by "Werner Punz (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12895230#action_12895230 ]
Werner Punz commented on MYFACES-2858:
--------------------------------------
I have isolated most of the scripts, from what I can see is there is no direct event handling messed up, but what could be is that the autoscroll feature does something in that case, do you have a bugreport on the issue?
But there is one section from what I can gather which never worked, that was that one
//this never worked in our code because version never was defined, I will drop it for now, since no one uses
//it anyway
//if (agentString.indexOf('msie') != -1) {
// if (!(agentString.indexOf('ppc') != -1 && agentString.indexOf('windows ce') != -1 && version >= 4.0) && window.external && window.external.AutoCompleteSaveForm) {
// window.external.AutoCompleteSaveForm(form);
// }
//}
(I commented that out) it sort of was a special context param which when enabled triggered autosave in older ie versions on input params, probably absolutely no one has used it, otherwise they would have stepped on following error:
gentString.indexOf('windows ce') != -1 && version >= 4.0)
the version var never was defined and probably would have caused an error. But the good thing is this code only was activated via a context param.
Nevertheless externalizing all this should be prio #1 it reduces the page size by about 8kbyte on every request!!! and also makes the code more maintainable.
> pointless oamsubmit inline rendering
> ------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Bug
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (MYFACES-2858) pointless oamsubmit inline
rendering (patch available)
Posted by "Jakob Korherr (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12899399#action_12899399 ]
Jakob Korherr commented on MYFACES-2858:
----------------------------------------
As with MYFACESTEST-24, we can remove the setUpFactories() workaround for the PartialViewContextFactory. I will also do a snapshot deploy of MyFaces test to verify that this is updated in the repo too and that no-one has build problems because of that.
> pointless oamsubmit inline rendering (patch available)
> ------------------------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Improvement
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
> Assignee: Jakob Korherr
> Fix For: 2.0.2-SNAPSHOT
>
> Attachments: https___issues_apache_org_jira_browse_MYFACES-2858_externalizing_oamsubmit__.patch
>
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (MYFACES-2858) pointless oamsubmit inline rendering
(patch available)
Posted by "Jakob Korherr (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jakob Korherr updated MYFACES-2858:
-----------------------------------
Status: Patch Available (was: Reopened)
> pointless oamsubmit inline rendering (patch available)
> ------------------------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Improvement
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
> Assignee: Jakob Korherr
> Fix For: 2.0.2-SNAPSHOT
>
> Attachments: https___issues_apache_org_jira_browse_MYFACES-2858_externalizing_oamsubmit__.patch, MYFACES-2858-2.patch
>
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (MYFACES-2858) pointless oamsubmit inline
rendering (patch available)
Posted by "Jakob Korherr (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12899353#action_12899353 ]
Jakob Korherr commented on MYFACES-2858:
----------------------------------------
Acutally that's not true. It first was on api (in Werner's patch) and I moved it to impl, because it is an implementation detail.
However you're right. I did not think about using tomahawk standalone with Mojarra. This will currently not work. Thus putting it on shared may solve the problem, but I have to try this first.
If you put them on jsf.js, the scenario tomahawk + mojarra also won't work. I will try some scenarios and provide a patch afterwards!
> pointless oamsubmit inline rendering (patch available)
> ------------------------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Improvement
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
> Fix For: 2.0.2-SNAPSHOT
>
> Attachments: https___issues_apache_org_jira_browse_MYFACES-2858_externalizing_oamsubmit__.patch
>
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (MYFACES-2858) pointless oamsubmit inline
rendering
Posted by "Jan-Kees van Andel (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12895224#action_12895224 ]
Jan-Kees van Andel commented on MYFACES-2858:
---------------------------------------------
No opinion regarding externalizing the scripts. I think it's a good idea.
What I did notice a while ago, was that those scripts are very annoying when doing AJAX. I didn't look at it in detail, but it looked like they messed up the event handling.
> pointless oamsubmit inline rendering
> ------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Bug
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (MYFACES-2858) pointless oamsubmit inline
rendering (patch available)
Posted by "Leonardo Uribe (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12899278#action_12899278 ]
Leonardo Uribe commented on MYFACES-2858:
-----------------------------------------
Note that on myfaces core there are two copies of oamSubmit.js: one on api and the other one on impl.
The functions on oamSubmit.js:
function oamSetHiddenInput(formname, name, value)
function oamClearHiddenInput(formname, name, value)
function oamSubmitForm(formName, linkId, target, params)
Shouldn't we just put them on jsf.js, but change its names to something like <someothernamespace>.oamSetHiddenInput ? In theory, those methods are myfaces specific. If by some reason someone needs them defined, why don't just put a resource with the functions calling the ones with namespace like this:
function oamSetHiddenInput(formname, name, value){
<someothernamespace>.oamSetHiddenInput(formname, name, value)
}
Or why don't create a variable that allows choose between render it inline or use the namespaced functions (default one)?
> pointless oamsubmit inline rendering (patch available)
> ------------------------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Improvement
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
> Fix For: 2.0.2-SNAPSHOT
>
> Attachments: https___issues_apache_org_jira_browse_MYFACES-2858_externalizing_oamsubmit__.patch
>
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (MYFACES-2858) pointless oamsubmit inline
rendering
Posted by "Werner Punz (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12895286#action_12895286 ]
Werner Punz commented on MYFACES-2858:
--------------------------------------
Ok the isolation so far works here, but I also want a compressed uncompressed version like we have for jsf.js, it is a 60% filesize difference between
those two.
The second issue is, I have the same problem as with jsf.js if inline rendered already and included in a ppr case we do not habe any suppression
on the server side hence the script gets aloaded anew in auto eval cases (mozilla)
> pointless oamsubmit inline rendering
> ------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Bug
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (MYFACES-2858) pointless oamsubmit inline rendering
(patch available)
Posted by "Jakob Korherr (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jakob Korherr updated MYFACES-2858:
-----------------------------------
Status: Resolved (was: Patch Available)
Resolution: Fixed
Patch applied, this should work now in every possible scenario.
> pointless oamsubmit inline rendering (patch available)
> ------------------------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Improvement
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
> Assignee: Jakob Korherr
> Fix For: 2.0.2-SNAPSHOT
>
> Attachments: https___issues_apache_org_jira_browse_MYFACES-2858_externalizing_oamsubmit__.patch, MYFACES-2858-2.patch
>
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (MYFACES-2858) pointless oamsubmit inline
rendering
Posted by "Werner Punz (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12895242#action_12895242 ]
Werner Punz commented on MYFACES-2858:
--------------------------------------
Ok I just ran a test since I got the first basics working, the actual savings is 2-3 kbyte per page, which is not too shabby I guess.
> pointless oamsubmit inline rendering
> ------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Bug
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Reopened: (MYFACES-2858) pointless oamsubmit inline
rendering (patch available)
Posted by "Leonardo Uribe (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Leonardo Uribe reopened MYFACES-2858:
-------------------------------------
We have to review first what happen if tomahawk + mojarra is used. Note in this case oamSubmit.js is not on the classpath.
I think if myfaces is used, we could include oamSubmit.js inside our jsf.js, but if mojarra is used, include it on a separate file as part of tomahawk and add it inline as shown.
> pointless oamsubmit inline rendering (patch available)
> ------------------------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Improvement
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
> Fix For: 2.0.2-SNAPSHOT
>
> Attachments: https___issues_apache_org_jira_browse_MYFACES-2858_externalizing_oamsubmit__.patch
>
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (MYFACES-2858) pointless oamsubmit inline
rendering (patch available)
Posted by "Jakob Korherr (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jakob Korherr resolved MYFACES-2858.
------------------------------------
Fix Version/s: 2.0.2-SNAPSHOT
Resolution: Fixed
committed a slightly modified version of Werner's patch.
> pointless oamsubmit inline rendering (patch available)
> ------------------------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Improvement
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
> Fix For: 2.0.2-SNAPSHOT
>
> Attachments: https___issues_apache_org_jira_browse_MYFACES-2858_externalizing_oamsubmit__.patch
>
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (MYFACES-2858) pointless oamsubmit inline
rendering
Posted by "Jan-Kees van Andel (JIRA)" <de...@myfaces.apache.org>.
[ https://issues.apache.org/jira/browse/MYFACES-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12895258#action_12895258 ]
Jan-Kees van Andel commented on MYFACES-2858:
---------------------------------------------
Never mind. I'll report my thingy in a separate issue. It's not related to your externalization.
Sorry for the unclear comment. :-)
> pointless oamsubmit inline rendering
> ------------------------------------
>
> Key: MYFACES-2858
> URL: https://issues.apache.org/jira/browse/MYFACES-2858
> Project: MyFaces Core
> Issue Type: Bug
> Affects Versions: 2.0.2-SNAPSHOT
> Reporter: Werner Punz
>
> We have had several functions rendered inline for ages, namely appendHiddenInput oamSubmit, the autoscroll stuff etc...
> I personally think the rendering of those functions as inline scripts is pointless, blows up the browsers tremendously and
> prevents that the affected scripts can be browser cached.
> A quick look at the code revealed that there is basically nothing which would prevent to externalize the scripts. My main problem is where to we handle the auto append code.
> My personal guess is we probably simply should add it as a resource definitions to the commandLink, Button etc.. renderers, any ideas regarding this?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.