Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.javascript > #25467 > unrolled thread

Images and pdf in a div

Back to article view | Back to comp.lang.javascript

Contents

  Images and pdf in a div danca <cyberdanny@tiscalinet.it> - 2014-07-19 00:31 +0200
    Re: Images and pdf in a div "Christoph M. Becker" <cmbecker69@arcor.de> - 2014-07-19 13:34 +0200
      Re: Images and pdf in a div danca <cyberdanny@tiscalinet.it> - 2014-07-19 15:34 +0200
        Re: Images and pdf in a div "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2014-07-19 15:51 +0200
          Re: Images and pdf in a div danca <cyberdanny@tiscalinet.it> - 2014-07-19 18:30 +0200
            Re: Images and pdf in a div Andrew Poulos <ap_prog@hotmail.com> - 2014-07-20 15:23 +1000
              Re: Images and pdf in a div Thomas 'PointedEars' Lahn <PointedEars@web.de> - 2014-07-20 12:42 +0200
                Re: Images and pdf in a div danca <cyberdanny@tiscalinet.it> - 2014-07-20 20:07 +0200
                  Re: Images and pdf in a div Andrew Poulos <ap_prog@hotmail.com> - 2014-07-21 14:10 +1000
                    Re: Images and pdf in a div danca <cyberdanny@tiscalinet.it> - 2014-07-21 14:19 +0200
                Re: Images and pdf in a div Andrew Poulos <ap_prog@hotmail.com> - 2014-07-21 08:14 +1000
                  Re: Images and pdf in a div Thomas 'PointedEars' Lahn <PointedEars@web.de> - 2014-07-21 02:00 +0200
                    Re: Images and pdf in a div "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2014-07-21 07:15 +0200
                      Re: Images and pdf in a div Osmo Saarikumpu <osmo@weppipakki.com> - 2014-07-24 18:26 +0300
                        Re: Images and pdf in a div "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2014-07-24 17:48 +0200
                          Re: Images and pdf in a div Ben Bacarisse <ben.usenet@bsb.me.uk> - 2014-07-24 22:32 +0100
                            Re: Images and pdf in a div "Evertjan." <exxjxw.hannivoort@inter.nl.net> - 2014-07-24 23:58 +0200
                        Re: Images and pdf in a div "Christoph M. Becker" <cmbecker69@arcor.de> - 2014-07-24 18:40 +0200
                          Re: Images and pdf in a div Thomas 'PointedEars' Lahn <PointedEars@web.de> - 2014-07-25 12:31 +0200
                            Re: Images and pdf in a div John Harris <niam@jghnorth.org.uk.invalid> - 2014-07-25 15:52 +0100

#25467 — Images and pdf in a div

Hi all,
I have a page with thumbnails. When the user clicks on a thumbnail a 
script does the following:
- creates a div
- populates the div with some buttons (next,previous, close, change 
resolution...)
-creates a image node, sets src to the big image the thumbnail 
correspond to, ad inserts it in the div
- finally appends the div to document.
So far, so good.
When the thumbnails points to a PDF, however, I have to change the div, 
particularly I:
-create a new embed node
-use replaceChild to switch the node image with the newly created node 
embed.
I could anyway go another route:
create the two nodes and simply give display:none to alternatively the 
image and the embed node, without the use of replaceChild.

Has someone an idea about cross-browsers issues with the first approach? 
It works with all the (main, modern) browsers on a pc. Android browser 
is OK too, but I'm unsure about iPads and iPhones (and I have no one to 
test with).
Thanks
Dan

[toc] | [next] | [standalone]

#25475

danca wrote:

> When the thumbnails points to a PDF, however, I have to change the div,
> particularly I:
> -create a new embed node
> -use replaceChild to switch the node image with the newly created node
> embed.
> I could anyway go another route:
> create the two nodes and simply give display:none to alternatively the
> image and the embed node, without the use of replaceChild.
> 
> Has someone an idea about cross-browsers issues with the first approach?
> It works with all the (main, modern) browsers on a pc. Android browser
> is OK too, but I'm unsure about iPads and iPhones (and I have no one to
> test with).

I don't see any problems regarding the replacement of the node (i.e.
your first approach).

However, the embed element surely requires a PDF reader/renderer to be
available in the browser (e.g. the Adobe Acrobat plugin).  If that is
not available, the browser likely will render an empty space or maybe an
error message.  Perhaps it is reasonable to display a preview image
instead of the PDF document.

-- 
Christoph M. Becker

[toc] | [prev] | [next] | [standalone]

#25477

Il 19/07/2014 13:34, Christoph M. Becker ha scritto:
> danca wrote:
>
>> When the thumbnails points to a PDF, however, I have to change the div,
>> particularly I:
>> -create a new embed node
>> -use replaceChild to switch the node image with the newly created node
>> embed.
>> I could anyway go another route:
>> create the two nodes and simply give display:none to alternatively the
>> image and the embed node, without the use of replaceChild.
>>
>> Has someone an idea about cross-browsers issues with the first approach?
>> It works with all the (main, modern) browsers on a pc. Android browser
>> is OK too, but I'm unsure about iPads and iPhones (and I have no one to
>> test with).
>
> I don't see any problems regarding the replacement of the node (i.e.
> your first approach).
>
> However, the embed element surely requires a PDF reader/renderer to be
> available in the browser (e.g. the Adobe Acrobat plugin).  If that is
> not available, the browser likely will render an empty space or maybe an
> error message.  Perhaps it is reasonable to display a preview image
> instead of the PDF document.
>
It is unclear to me how to detect preliminarly the availability of the 
plugin, or which syntax to use to have a fallback to a preview image.
I am a bit puzzled with this idea, it sounds good but I would greatly 
appreciate a hint.
My basic code is something like this:
...
var oldNode=document.getElementById("doc")
var newNode=document.createElement("embed")
newNode.type="application/pdf"
newNode.src="http://"+location.host+"/images/test.pdf"
oldNode.parentNode.replaceChild(newNode,oldNode)
newNode.id="doc" // ready for another substitution - same id as previous
...
then I resize the div container etc.
So... what kind of feature must I look for?
Many thanks
Dan

[toc] | [prev] | [next] | [standalone]

#25478

danca <cyberdanny@tiscalinet.it> wrote on 19 jul 2014 in 
comp.lang.javascript:

> Il 19/07/2014 13:34, Christoph M. Becker ha scritto:
>> danca wrote:
>>
>>> When the thumbnails points to a PDF, however, I have to change the div,
>>> particularly I:
>>> -create a new embed node
>>> -use replaceChild to switch the node image with the newly created node
>>> embed.
>>> I could anyway go another route:
>>> create the two nodes and simply give display:none to alternatively the
>>> image and the embed node, without the use of replaceChild.
>>>
>>> Has someone an idea about cross-browsers issues with the first approach?
>>> It works with all the (main, modern) browsers on a pc. Android browser
>>> is OK too, but I'm unsure about iPads and iPhones (and I have no one to
>>> test with).
>>
>> I don't see any problems regarding the replacement of the node (i.e.
>> your first approach).
>>
>> However, the embed element surely requires a PDF reader/renderer to be
>> available in the browser (e.g. the Adobe Acrobat plugin).  If that is
>> not available, the browser likely will render an empty space or maybe an
>> error message.  Perhaps it is reasonable to display a preview image
>> instead of the PDF document.
>>
> It is unclear to me how to detect preliminarly the availability of the 
> plugin, or which syntax to use to have a fallback to a preview image.
> I am a bit puzzled with this idea, it sounds good but I would greatly 
> appreciate a hint.
> My basic code is something like this:
> ...
> var oldNode=document.getElementById("doc")
> var newNode=document.createElement("embed")
> newNode.type="application/pdf"
> newNode.src="http://"+location.host+"/images/test.pdf"
> oldNode.parentNode.replaceChild(newNode,oldNode)
> newNode.id="doc" // ready for another substitution - same id as previous
> ...
> then I resize the div container etc.
> So... what kind of feature must I look for?

Google is your friend, I hope:

<http://thecodeabode.blogspot.nl/2011/01/detect-adobe-reader-plugin.html>

I did not test this, btw.


-- 
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[toc] | [prev] | [next] | [standalone]

#25480

Il 19/07/2014 15:51, Evertjan. ha scritto:
> danca <cyberdanny@tiscalinet.it> wrote on 19 jul 2014 in
> comp.lang.javascript:
>
>> Il 19/07/2014 13:34, Christoph M. Becker ha scritto:
...
>>>
>>> I don't see any problems regarding the replacement of the node (i.e.
>>> your first approach).
>>>
>>> However, the embed element surely requires a PDF reader/renderer to be
>>> available in the browser (e.g. the Adobe Acrobat plugin).  If that is
>>> not available, the browser likely will render an empty space or maybe an
>>> error message.  Perhaps it is reasonable to display a preview image
>>> instead of the PDF document.
>>>
>> It is unclear to me how to detect preliminarly the availability of the
>> plugin, or which syntax to use to have a fallback to a preview image.
>> I am a bit puzzled with this idea, it sounds good but I would greatly
>> appreciate a hint.
>> My basic code is something like this:
>> ...
>> var oldNode=document.getElementById("doc")
>> var newNode=document.createElement("embed")
>> newNode.type="application/pdf"
>> newNode.src="http://"+location.host+"/images/test.pdf"
>> oldNode.parentNode.replaceChild(newNode,oldNode)
>> newNode.id="doc" // ready for another substitution - same id as previous
>> ...
>> then I resize the div container etc.
>> So... what kind of feature must I look for?
>
> Google is your friend, I hope:
>
> <http://thecodeabode.blogspot.nl/2011/01/detect-adobe-reader-plugin.html>
>
> I did not test this, btw.
>
I feared an answer like this. Thanks, no.

(excerpt)

   var getBrowserName = function() {
     return this.name = this.name || function() {
       var userAgent = navigator ? navigator.userAgent.toLowerCase() : 
"other";

       if(userAgent.indexOf("chrome") > -1)        return "chrome";
       else if(userAgent.indexOf("safari") > -1)   return "safari";
       else if(userAgent.indexOf("msie") > -1)     return "ie";
       else if(userAgent.indexOf("firefox") > -1)  return "firefox";
       return userAgent;

ugh.
btw weren't we all negative about browser sniffing? I'm doing my best to 
follow best practices and this NG is a big resource in this matter.
(Still learning, as all my life is about).

Uhm no, I'll not support people not able to correctly display a pdf in 
their browser. They can download the file and see it offline.

Apart the fact that this all is about a CMS-like system, the script part 
being a fragment of the total, the publisher decides what to publish and 
his audience, so if one comes and has no pdf support it is not my 
problem at the end of the tale, if not in a very indirect way.

Anyway I hope my question is of interest for someone else.
When it comes to "particular" questions this is the only place to go. 
Thanks!
Dan

[toc] | [prev] | [next] | [standalone]

#25481

On 20/07/2014 2:30 AM, danca wrote:
> Il 19/07/2014 15:51, Evertjan. ha scritto:

>> <http://thecodeabode.blogspot.nl/2011/01/detect-adobe-reader-plugin.html>
>>
>> I did not test this, btw.
>>
> I feared an answer like this. Thanks, no.
>
> (excerpt)
>
>    var getBrowserName = function() {
>      return this.name = this.name || function() {
>        var userAgent = navigator ? navigator.userAgent.toLowerCase() :
> "other";
>
>        if(userAgent.indexOf("chrome") > -1)        return "chrome";
>        else if(userAgent.indexOf("safari") > -1)   return "safari";
>        else if(userAgent.indexOf("msie") > -1)     return "ie";
>        else if(userAgent.indexOf("firefox") > -1)  return "firefox";
>        return userAgent;
>
> ugh.

How about changing the code to something like:

var AcrobatInfo = (function() {
   var isInstalled = false,
       vern = null;
		
   function findPlugin(nme) {
     var p;
     for (key in navigator.plugins) {
       p = navigator.plugins[key];
       if (p.name == nme) return p;
     }
     return null;
   }

   (function() {
     var plugin;
		
     if (typeof window.ActiveXObject != "undefined") {
       // AcroPDF.PDF is used by version 7 and later
       // PDF.PdfCtrl is used by version 6 and earlier
       try {
         plugin = new ActiveXObject("AcroPDF.PDF");
         isInstalled = true;
       } catch(e) {}
       try {
         plugin = new ActiveXObject("PDF.PdfCtrl");
         isInstalled = true;
       } catch(e) {}
			
       if (plugin) {
         vern = plugin.GetVersions().split(',')[0].split('=')[1];
       }
     } else {
       plugin = findPlugin('Adobe Acrobat')
               || findPlugin('Chrome PDF Viewer')
               || findPlugin('WebKit built-in PDF');
			
       if (plugin) isInstalled = true;
       if (plugin && plugin.version) vern = plugin.version;
     }
       plugin = null;
     })();
	
     return {
       installed: isInstalled,
       version: vern
     };
})();


Usage

var a = "Acrobat is "
         + (AcrobatInfo.installed?"":"not")
         + " installed. "
         + AcrobatInfo.version;

One shortcoming is that while FF will return the full version string (eg 
10.1.7.27) IE 9 only returns part of the version string (eg 10.1.)

Andrew Poulos

[toc] | [prev] | [next] | [standalone]

#25483

Andrew Poulos wrote:

> How about changing the code to something like:
> 
> var AcrobatInfo = (function() {

Not a constructor; by convention, the name should start lowercase.  
Preferably, you would put it in a private namespace, like “ap.acrobatInfo”.  
Creating a user-defined object that would retrieve the information about all 
plugins, and then would allow to retrieve information for a specific plugin, 
is indicated.

>    var isInstalled = false,
>        vern = null;
> 
>    function findPlugin(nme) {
>      var p;
>      for (key in navigator.plugins) {

Do not use the for-in statement with host objects and array-like objects; 
their properties may not be enumerable, not in a defined order or have 
enumerable properties that you do not want to iterate over (like “refresh”); 
use the “for” statement instead.  Exception: Array instances known to 
encapsulate a sparse array; in that case for-in is more efficient, and you 
can filter and sort the properties later, if necessary.

You have not declared “key”; this code will throw a ReferenceError exception 
in strict mode.

> […]
>    (function() {
>      var plugin;
> 
>      if (typeof window.ActiveXObject != "undefined") {

Check the “ActiveXObject” property of the global object instead.

>        // AcroPDF.PDF is used by version 7 and later
>        // PDF.PdfCtrl is used by version 6 and earlier

You should use multi-line comments for documentation instead, in order to be 
able to tell them apart from deactivating comments easily.

>        try {
>          plugin = new ActiveXObject("AcroPDF.PDF");
>          isInstalled = true;
>        } catch(e) {}
>        try {
>          plugin = new ActiveXObject("PDF.PdfCtrl");
>          isInstalled = true;
>        } catch(e) {}

You should nest those try-catch blocks, leaving at most one empty “catch” 
block.

“isInstalled” is superfluous because you can test against the value of 
“plugin” instead, which would be either not “null” or not “undefined” if 
successful, depending on how you initialize it.

>      })();
> […]
> })();

I suggest you always use the same indentation for opening and closing 
delimiters.

And “}())” instead in order to make clear what makes up the right-hand side 
expression.  This will also become handy in editors like Vim or Eclipse JSDT 
with Vrapper plugin, where, which I discovered yesterday, you can type “vi)” 
to select everything between the outer parentheses, for refactoring.
 
-- 
PointedEars
FAQ: <http://PointedEars.de/faq> | SVN: <http://PointedEars.de/wsvn/>
Twitter: @PointedEars2 | ES Matrix: <http://PointedEars.de/es-matrix>
Please do not Cc: me. / Bitte keine Kopien per E-Mail.

[toc] | [prev] | [next] | [standalone]

#25484

Il 20/07/2014 12:42, Thomas 'PointedEars' Lahn ha scritto:
> Andrew Poulos wrote:
>
>> How about changing the code to something like:
>>
>> var AcrobatInfo = (function() {
>
> Not a constructor; by convention, the name should start lowercase.
> Preferably, you would put it in a private namespace, like
> “ap.acrobatInfo”. Creating a user-defined object that would retrieve
> the information about all plugins, and then would allow to retrieve
> information for a specific plugin, is indicated.
>
>> var isInstalled = false, vern = null;
>>
>> function findPlugin(nme) { var p; for (key in navigator.plugins) {
>
> Do not use the for-in statement with host objects and array-like
> objects; their properties may not be enumerable, not in a defined
> order or have enumerable properties that you do not want to iterate
> over (like “refresh”); use the “for” statement instead.  Exception:
> Array instances known to encapsulate a sparse array; in that case
> for-in is more efficient, and you can filter and sort the properties
> later, if necessary.
>
> You have not declared “key”; this code will throw a ReferenceError
> exception in strict mode.
>
>> […] (function() { var plugin;
>>
>> if (typeof window.ActiveXObject != "undefined") {
>
> Check the “ActiveXObject” property of the global object instead.
>
>> // AcroPDF.PDF is used by version 7 and later // PDF.PdfCtrl is
>> used by version 6 and earlier
>
> You should use multi-line comments for documentation instead, in
> order to be able to tell them apart from deactivating comments
> easily.
>
>> try { plugin = new ActiveXObject("AcroPDF.PDF"); isInstalled =
>> true; } catch(e) {} try { plugin = new
>> ActiveXObject("PDF.PdfCtrl"); isInstalled = true; } catch(e) {}
>
> You should nest those try-catch blocks, leaving at most one empty
> “catch” block.
>
> “isInstalled” is superfluous because you can test against the value
> of “plugin” instead, which would be either not “null” or not
> “undefined” if successful, depending on how you initialize it.
>
>> })(); […] })();
>
> I suggest you always use the same indentation for opening and
> closing delimiters.
>
> And “}())” instead in order to make clear what makes up the
> right-hand side expression.  This will also become handy in editors
> like Vim or Eclipse JSDT with Vrapper plugin, where, which I
> discovered yesterday, you can type “vi)” to select everything between
> the outer parentheses, for refactoring.
>
>
Uhm... something like this?
getInfo={}
getInfo.acrobatInfo = (function() { var vern = null;
function findPlugin(nme) { var p;
for (var key=0; key<navigator.plugins.length; ++key) {
p = navigator.plugins[key];
if (p.name == nme) return p; }
return  null; }
(function() { var plugin;
if (typeof ActiveXObject != "undefined") { /* AcroPDF.PDF is used by 
version 7 and later - PDF.PdfCtrl is used by version 6 and earlier*/
try { plugin = new ActiveXObject("AcroPDF.PDF")}
catch(e) { try { plugin = new ActiveXObject("PDF.PdfCtrl")}
catch(e) {}
}
if (plugin) { vern = plugin.GetVersions().split(',')[0].split('=')[1]; }
} else { plugin = findPlugin('Adobe Acrobat') || findPlugin('Chrome PDF 
Viewer') || findPlugin('WebKit built-in PDF');
if (plugin && plugin.version) vern = plugin.version; }
plugin = null; })();
return { installed: vern!=null, version: vern }; })();


Usage

var a = "Acrobat is " + (getInfo.acrobatInfo.installed?"":"not") + 
"installed. " + getInfo.acrobatInfo.version||"";

Many thanks.
I tested in FF, Konqueror and IE 8/9 and it seems to work, (ok I don't 
have Acrobat plugin installed in Konqueror so it's unclear). In IE 8 it 
truncates the version number, and in Opera 12.16 and Safari 5.1.7 the 
plugin is not detected.
Dan

[toc] | [prev] | [next] | [standalone]

#25492

On 21/07/2014 4:07 AM, danca wrote:
> Il 20/07/2014 12:42, Thomas 'PointedEars' Lahn ha scritto:
>> Andrew Poulos wrote:
>>
>>> How about changing the code to something like:
>>>
>>> var AcrobatInfo = (function() {
>>
>> Not a constructor; by convention, the name should start lowercase.
>> Preferably, you would put it in a private namespace, like
>> “ap.acrobatInfo”. Creating a user-defined object that would retrieve
>> the information about all plugins, and then would allow to retrieve
>> information for a specific plugin, is indicated.
>>
>>> var isInstalled = false, vern = null;
>>>
>>> function findPlugin(nme) { var p; for (key in navigator.plugins) {
>>
>> Do not use the for-in statement with host objects and array-like
>> objects; their properties may not be enumerable, not in a defined
>> order or have enumerable properties that you do not want to iterate
>> over (like “refresh”); use the “for” statement instead.  Exception:
>> Array instances known to encapsulate a sparse array; in that case
>> for-in is more efficient, and you can filter and sort the properties
>> later, if necessary.
>>
>> You have not declared “key”; this code will throw a ReferenceError
>> exception in strict mode.
>>
>>> […] (function() { var plugin;
>>>
>>> if (typeof window.ActiveXObject != "undefined") {
>>
>> Check the “ActiveXObject” property of the global object instead.
>>
>>> // AcroPDF.PDF is used by version 7 and later // PDF.PdfCtrl is
>>> used by version 6 and earlier
>>
>> You should use multi-line comments for documentation instead, in
>> order to be able to tell them apart from deactivating comments
>> easily.
>>
>>> try { plugin = new ActiveXObject("AcroPDF.PDF"); isInstalled =
>>> true; } catch(e) {} try { plugin = new
>>> ActiveXObject("PDF.PdfCtrl"); isInstalled = true; } catch(e) {}
>>
>> You should nest those try-catch blocks, leaving at most one empty
>> “catch” block.
>>
>> “isInstalled” is superfluous because you can test against the value
>> of “plugin” instead, which would be either not “null” or not
>> “undefined” if successful, depending on how you initialize it.
>>
>>> })(); […] })();
>>
>> I suggest you always use the same indentation for opening and
>> closing delimiters.
>>
>> And “}())” instead in order to make clear what makes up the
>> right-hand side expression.  This will also become handy in editors
>> like Vim or Eclipse JSDT with Vrapper plugin, where, which I
>> discovered yesterday, you can type “vi)” to select everything between
>> the outer parentheses, for refactoring.
>>
>>
> Uhm... something like this?
> getInfo={}
> getInfo.acrobatInfo = (function() { var vern = null;
> function findPlugin(nme) { var p;
> for (var key=0; key<navigator.plugins.length; ++key) {
> p = navigator.plugins[key];
> if (p.name == nme) return p; }
> return  null; }
> (function() { var plugin;
> if (typeof ActiveXObject != "undefined") { /* AcroPDF.PDF is used by
> version 7 and later - PDF.PdfCtrl is used by version 6 and earlier*/
> try { plugin = new ActiveXObject("AcroPDF.PDF")}
> catch(e) { try { plugin = new ActiveXObject("PDF.PdfCtrl")}
> catch(e) {}
> }
> if (plugin) { vern = plugin.GetVersions().split(',')[0].split('=')[1]; }
> } else { plugin = findPlugin('Adobe Acrobat') || findPlugin('Chrome PDF
> Viewer') || findPlugin('WebKit built-in PDF');
> if (plugin && plugin.version) vern = plugin.version; }
> plugin = null; })();
> return { installed: vern!=null, version: vern }; })();
>
> Usage
>
> var a = "Acrobat is " + (getInfo.acrobatInfo.installed?"":"not") +
> "installed. " + getInfo.acrobatInfo.version||"";
>
> Many thanks.
> I tested in FF, Konqueror and IE 8/9 and it seems to work, (ok I don't
> have Acrobat plugin installed in Konqueror so it's unclear). In IE 8 it
> truncates the version number, and in Opera 12.16 and Safari 5.1.7 the
> plugin is not detected.

It truncates the version in Chrome as well.

It works for me in Opera 22 and Safari 5 (under Windows) but to get the 
version in Safari 5 I parsed plugin.description as plugin.version didn't 
seem to exist.

Andrew Poulos

[toc] | [prev] | [next] | [standalone]

#25497

Il 21/07/2014 06:10, Andrew Poulos ha scritto:
> On 21/07/2014 4:07 AM, danca wrote:

>> Many thanks.
>> I tested in FF, Konqueror and IE 8/9 and it seems to work, (ok I don't
>> have Acrobat plugin installed in Konqueror so it's unclear). In IE 8 it
>> truncates the version number, and in Opera 12.16 and Safari 5.1.7 the
>> plugin is not detected.
>
> It truncates the version in Chrome as well.
>
> It works for me in Opera 22 and Safari 5 (under Windows) but to get the
> version in Safari 5 I parsed plugin.description as plugin.version didn't
> seem to exist.
>
> Andrew Poulos
>
Ah...Ok for Safari, I had no time to investigate further, Opera 22 is 
similar to Chrome, AFAIK, while Opera 12.16 is the last one with Presto, 
a different beast. It should use the same plugin as Firefox, so for now 
it's unclear. Not that important: probably nobody uses Opera 12 anymore, 
except me ans a few other dinosaurs. But Opera mini should be still 
alive, I didn't test any mobile device.
THX,
Dan

[toc] | [prev] | [next] | [standalone]

#25487

On 20/07/2014 8:42 PM, Thomas 'PointedEars' Lahn wrote:
> Andrew Poulos wrote:
>
>> How about changing the code to something like:
>>
>> var AcrobatInfo = (function() {
>
> Not a constructor; by convention, the name should start lowercase.
> Preferably, you would put it in a private namespace, like “ap.acrobatInfo”.
> Creating a user-defined object that would retrieve the information about all
> plugins, and then would allow to retrieve information for a specific plugin,
> is indicated.

Yes, you're right.

>>     var isInstalled = false,
>>         vern = null;
>>
>>     function findPlugin(nme) {
>>       var p;
>>       for (key in navigator.plugins) {
>
> Do not use the for-in statement with host objects and array-like objects;
> their properties may not be enumerable, not in a defined order or have
> enumerable properties that you do not want to iterate over (like “refresh”);
> use the “for” statement instead.  Exception: Array instances known to
> encapsulate a sparse array; in that case for-in is more efficient, and you
> can filter and sort the properties later, if necessary.

How do I search through navigator.plugins using "for" instead of "for-in"?

> You have not declared “key”; this code will throw a ReferenceError exception
> in strict mode.

Dang, I missed it.

>> […]
>>     (function() {
>>       var plugin;
>>
>>       if (typeof window.ActiveXObject != "undefined") {
>
> Check the “ActiveXObject” property of the global object instead.
>
>>         // AcroPDF.PDF is used by version 7 and later
>>         // PDF.PdfCtrl is used by version 6 and earlier
>
> You should use multi-line comments for documentation instead, in order to be
> able to tell them apart from deactivating comments easily.
>
>>         try {
>>           plugin = new ActiveXObject("AcroPDF.PDF");
>>           isInstalled = true;
>>         } catch(e) {}
>>         try {
>>           plugin = new ActiveXObject("PDF.PdfCtrl");
>>           isInstalled = true;
>>         } catch(e) {}
>
> You should nest those try-catch blocks, leaving at most one empty “catch”
> block.

Ok.

> “isInstalled” is superfluous because you can test against the value of
> “plugin” instead, which would be either not “null” or not “undefined” if
> successful, depending on how you initialize it.

After searching for "plugin" I set it to null to clear any memory...

>>       })();
>> […]
>> })();
>
> I suggest you always use the same indentation for opening and closing
> delimiters.

Hmm, I think its a typo.

> And “}())” instead in order to make clear what makes up the right-hand side
> expression.  This will also become handy in editors like Vim or Eclipse JSDT
> with Vrapper plugin, where, which I discovered yesterday, you can type “vi)”
> to select everything between the outer parentheses, for refactoring.

Ok, 
<http://dmitrysoshnikov.com/ecmascript/chapter-5-functions/#question-about-surrounding-parentheses> 
says that both are correct so I just picked the first one.

Thanks for the advice.

Andrew Poulos

[toc] | [prev] | [next] | [standalone]

#25488

Andrew Poulos wrote:

> On 20/07/2014 8:42 PM, Thomas 'PointedEars' Lahn wrote:
>> Andrew Poulos wrote:
>>>     var isInstalled = false,
>>>         vern = null;
>>>
>>>     function findPlugin(nme) {
>>>       var p;
>>>       for (key in navigator.plugins) {
>>
>> Do not use the for-in statement with host objects and array-like objects;
>> their properties may not be enumerable, not in a defined order or have
>> enumerable properties that you do not want to iterate over (like
>> “refresh”);
>> use the “for” statement instead.  Exception: Array instances known to
>> encapsulate a sparse array; in that case for-in is more efficient, and
>> you can filter and sort the properties later, if necessary.
> 
> How do I search through navigator.plugins using "for" instead of "for-in"?

  for (var i = 0, plugins = navigator.plugins, len = plugins.length;
       i < len; ++i)
  {
    … plugins[i] …
  }

(One of the properties you get when you use for-in is the “length” property 
from which you could have inferred that.)
 
>> “isInstalled” is superfluous because you can test against the value of
>> “plugin” instead, which would be either not “null” or not “undefined” if
>> successful, depending on how you initialize it.
> 
> After searching for "plugin" I set it to null to clear any memory...

Unnecessary if you do not return it, may even be harmful.
 
>> And “}())” instead in order to make clear what makes up the right-hand
>> side expression.  This will also become handy in editors like Vim or
>> Eclipse JSDT with Vrapper plugin, where, which I discovered yesterday,
>> you can type “vi)” to select everything between the outer parentheses,
>> for refactoring.
> 
> Ok,
> <http://dmitrysoshnikov.com/ecmascript/chapter-5-functions/#question-about-surrounding-parentheses>
> says that both are correct so I just picked the first one.

Where you put the parantheses in this case is a matter of code style.

However, there are several factually wrong or misleading statements in that 
text.  I find it interesting that someone who has demonstrated here 
(approximately at the time of writing that article, or shortly after) that 
he neither understood ECMAScript nor that he was willing to learn, now 
thinks he can teach others about it.  I find it even more interesting that 
some people believe those statements as if he would be some kind of 
authority on the subject.

> Thanks for the advice.

You’re welcome.

-- 
PointedEars
FAQ: <http://PointedEars.de/faq> | SVN: <http://PointedEars.de/wsvn/>
Twitter: @PointedEars2 | ES Matrix: <http://PointedEars.de/es-matrix>
Please do not Cc: me. / Bitte keine Kopien per E-Mail.

[toc] | [prev] | [next] | [standalone]

#25493

Thomas 'PointedEars' Lahn <PointedEars@web.de> wrote on 21 jul 2014 in 
comp.lang.javascript:

> for (var i = 0, plugins = navigator.plugins, len = plugins.length;
>        i < len; ++i)

Some semantic plugging-in:

, var plugins = navigator.plugins, var len = plugins.length; 

methinks.

-- 
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[toc] | [prev] | [next] | [standalone]

#25551

On 21.7.2014 8:15, Evertjan. wrote:
> Thomas 'PointedEars' Lahn <PointedEars@web.de> wrote on 21 jul 2014 in
> comp.lang.javascript:
>
>> for (var i = 0, plugins = navigator.plugins, len = plugins.length;
>>         i < len; ++i)
>
> Some semantic plugging-in:
>
> , var plugins = navigator.plugins, var len = plugins.length;
>
> methinks.

I've never seen that suggested or used before. Are you sure?

-- 
Best wishes, Osmo

[toc] | [prev] | [next] | [standalone]

#25553

Osmo Saarikumpu <osmo@weppipakki.com> wrote on 24 jul 2014 in 
comp.lang.javascript:

> On 21.7.2014 8:15, Evertjan. wrote:
>> Thomas 'PointedEars' Lahn <PointedEars@web.de> wrote on 21 jul 2014 in
>> comp.lang.javascript:
>>
>>> for (var i = 0, plugins = navigator.plugins, len = plugins.length;
>>>         i < len; ++i)
>>
>> Some semantic plugging-in:
>>
>> , var plugins = navigator.plugins, var len = plugins.length;
>>
>> methinks.
> 
> I've never seen that suggested or used before. Are you sure?

Can you give us any reason why 'i' should be varred,
and 'plugins' and 'len' not?

It could be [I did not think of it] that:

var i = 0, plugins = navigator.plugins, len = plugins.length

implicitly vars all 3, like it does in a simple statement.

That could or could not be true, I see no way of testing that,
because of the forhiving behavour of js implementations?

-- 
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[toc] | [prev] | [next] | [standalone]

#25555

"Evertjan." <exxjxw.hannivoort@inter.nl.net> writes:
<snip>
> It could be [I did not think of it] that:
>
> var i = 0, plugins = navigator.plugins, len = plugins.length
>
> implicitly vars all 3, like it does in a simple statement.

Yes, that's what the ECMA language specification says should happen.

> That could or could not be true, I see no way of testing that,
> because of the forhiving behavour of js implementations?

  (function () {
      for (var i = 0, j = 0; i < 1; i++);
  })();
  console.log(typeof i, typeof j);

You should see i and j undefined after the function execution.  If you
remove the "var", you will see that both i and j are now shown as
"number".  If the var applied to only one, you would see "undefined" and
"number".

-- 
Ben.

[toc] | [prev] | [next] | [standalone]

#25556

Ben Bacarisse <ben.usenet@bsb.me.uk> wrote on 24 jul 2014 in 
comp.lang.javascript:

> "Evertjan." <exxjxw.hannivoort@inter.nl.net> writes:
> <snip>
>> It could be [I did not think of it] that:
>>
>> var i = 0, plugins = navigator.plugins, len = plugins.length
>>
>> implicitly vars all 3, like it does in a simple statement.
> 
> Yes, that's what the ECMA language specification says should happen.
> 
>> That could or could not be true, I see no way of testing that,
>> because of the forhiving behavour of js implementations?
> 
>   (function () {
>       for (var i = 0, j = 0; i < 1; i++);
>   })();
>   console.log(typeof i, typeof j);
> 
> You should see i and j undefined after the function execution.  If you
> remove the "var", you will see that both i and j are now shown as
> "number".  If the var applied to only one, you would see "undefined" and
> "number".

tnx.



-- 
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[toc] | [prev] | [next] | [standalone]

#25554

Osmo Saarikumpu wrote:

> On 21.7.2014 8:15, Evertjan. wrote:
>> Thomas 'PointedEars' Lahn <PointedEars@web.de> wrote on 21 jul 2014 in
>> comp.lang.javascript:
>>
>>> for (var i = 0, plugins = navigator.plugins, len = plugins.length;
>>>         i < len; ++i)
>>
>> Some semantic plugging-in:
>>
>> , var plugins = navigator.plugins, var len = plugins.length;
>>
>> methinks.
>
> I've never seen that suggested or used before. Are you sure?

According to the ECMAScript specification, Edition 5.1[1], Evertjan.'s
suggestion would be a syntax error.  Besides, Thomas' example already
declares all three variables.

<http://www.ecma-international.org/ecma-262/5.1/#sec-12.6.3>

-- 
Christoph M. Becker

[toc] | [prev] | [next] | [standalone]

#25563

Christoph M. Becker wrote:

> Osmo Saarikumpu wrote:
>> On 21.7.2014 8:15, Evertjan. wrote:
>>> Thomas 'PointedEars' Lahn <PointedEars@web.de> wrote on 21 jul 2014 in
>>> comp.lang.javascript:

Attribution *line*, _not_ attribution novel.

>>>> for (var i = 0, plugins = navigator.plugins, len = plugins.length;
>>>>         i < len; ++i)
>>>
>>> Some semantic plugging-in:
>>>
>>> , var plugins = navigator.plugins, var len = plugins.length;
>>>
>>> methinks.
>>
>> I've never seen that suggested or used before. Are you sure?
> 
> According to the ECMAScript specification, Edition 5.1[1], Evertjan.'s
> suggestion would be a syntax error.  […]

Yes, of course, but simply *testing* the code with any ECMAScript-based 
script engine would show that (all major browsers have a built-in script 
console, there are add-ons/bookmarklets and “javascript:” URIs for the 
others [1]).  One should acquire a habit of testing code before posting it, 
or indicating that it is untested.  Posting untested code can be excused as 
its complexity increases.

[1] <http://PointedEars.de/faq#debugging>

-- 
PointedEars
FAQ: <http://PointedEars.de/faq> | SVN: <http://PointedEars.de/wsvn/>
Twitter: @PointedEars2 | ES Matrix: <http://PointedEars.de/es-matrix>
Please do not Cc: me. / Bitte keine Kopien per E-Mail.

[toc] | [prev] | [next] | [standalone]

#25564

On Fri, 25 Jul 2014 12:31:56 +0200, Thomas 'PointedEars' Lahn
<PointedEars@web.de> wrote:

  <snip>

>-- 
>PointedEars
>FAQ: <http://PointedEars.de/faq> | SVN: <http://PointedEars.de/wsvn/>
>Twitter: @PointedEars2 | ES Matrix: <http://PointedEars.de/es-matrix>
>Please do not Cc: me. / Bitte keine Kopien per E-Mail.

Please do not post an essay in your "signature".

  John

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.javascript

csiph-web