Don Spam's Reckless Son wrote:
> Dirk Munk wrote:
>> Don Spam's Reckless Son wrote:
>>> Mike Spencer wrote:
>>>> frg <frgrahl@gmx.net> writes:
>>>>
>>>>> Well decades-long is 2015 till today and I am not the only one
>>>>> working on it :) In any case you will need a second browser for bad
>>>>> sites for the foreseeable future. This will not go away any time
>>>>> soon unless we get more contributors and less whining about the bad
>>>>> world out there. Don't like it either.
>>>>
>>>> Just where does thwe problem lie withing Seamonkey?  Failure to parse
>>>> and execute some locutions of javascript?   Failure to correctly
>>>> interpret STYLE blocks?
>>>>
>>>> I encounter many sites that are awkwardly rendered or even to too much
>>>> of a hash to be readable.  Choosing "None" from the style menu usually
>>>> fixes that, albeit at the cost of having SVG crap rendered as
>>>> full-screen.
>>>>
>>>> I occasionally hit a failure of js but as I almost always disable js,
>>>> that isn't a common problem.
>>>>
>>>> There are only a couple of sites I log in to and, with images, style
>>>> and scripts enabled, they work as expected.  I don't do Facebook or
>>>> anything similar where, I assume, there is a vast block of js,
>>>> directly from FB and from other sites, that has to work perfectly for
>>>> the UX to have all the intended features.
>>>>
>>>> In any case, I really like Seamonkey, using 2.53.10.2.
>>>>
>>>
>>> This has been discussed here before, and I don't know JS so this is
>>> going to be generic.  There are two problems, one relatively minor
>>> and one anything but:
>>> 1 - Browser sniffing.  Sites deciding to support particular versions
>>> of particular browsers, and nothing else.
>>> 2 - Google love adding new features to Javascript.  Other browsers
>>> such as Firefox have to follow suit because websites - in particular
>>> ones with dynamically generated pages like using these wonderful (or
>>> not) new commands.  Firefox has the resources to handle the new
>>> syntax, Seamonkey does not and is much slower implementing the commands.
>>> Firefox rewrote their rendering engine a year or two ago and cut
>>> Thunderbird adrift as part of the process.  Seamonkey = Firefox +
>>> Thunderbird (+ Composer + other-bits) so FF and TB using different
>>> rendering engines is a problem.  Firefox has also made some other
>>> changes but those are not as relevant to this discussion.
>>
>> So, apart from the browser sniffing, the problem lies with Javascript.
>>
>> I've tried to understand what the problem is, and it seems that the
>> most recent version of the Javascript language is EcmaScript 2022, if
>> I understand that correctly.
>>
>> Am I to understand that the Javascript engine in Seamonkey does not
>> confirm to this standard?
>
> FRG is one person and can only do so much.  Google is adding syntax to
> JS faster than one person can implement it.

I'm not blaming Frank, I just want to know the cause of the problems.

In the good old days of Microsoft Explorer, Microsoft added functions
that were not part of the official HTML language.

At the moment Chrome is blamed for adding new functions in JavaScript.
But it does make a difference if these functions are part of the
official JavaScript language or not. If they are, then it is a Seamonkey
problem. If they are not, then web builders are using inofficial
JavaScript functions.

Don Spam's Reckless Son wrote:
> Mike Spencer wrote:
>> frg <frgrahl@gmx.net> writes:
>>
>>> Well decades-long is 2015 till today and I am not the only one
>>> working on it :) In any case you will need a second browser for bad
>>> sites for the foreseeable future. This will not go away any time
>>> soon unless we get more contributors and less whining about the bad
>>> world out there. Don't like it either.
>>
>> Just where does thwe problem lie withing Seamonkey?  Failure to parse
>> and execute some locutions of javascript?   Failure to correctly
>> interpret STYLE blocks?
>>
>> I encounter many sites that are awkwardly rendered or even to too much
>> of a hash to be readable.  Choosing "None" from the style menu usually
>> fixes that, albeit at the cost of having SVG crap rendered as
>> full-screen.
>>
>> I occasionally hit a failure of js but as I almost always disable js,
>> that isn't a common problem.
>>
>> There are only a couple of sites I log in to and, with images, style
>> and scripts enabled, they work as expected.  I don't do Facebook or
>> anything similar where, I assume, there is a vast block of js,
>> directly from FB and from other sites, that has to work perfectly for
>> the UX to have all the intended features.
>>
>> In any case, I really like Seamonkey, using 2.53.10.2.
>>
>
> This has been discussed here before, and I don't know JS so this is
> going to be generic.  There are two problems, one relatively minor and
> one anything but:
> 1 - Browser sniffing.  Sites deciding to support particular versions of
> particular browsers, and nothing else.

This one may or may not be scripting related. Certainly, a script can
sniff out a UA string and announce an objection if it doesn't find a
defined minimum version.

But that doesn't have to be done with scripts.

On the small page that I maintain, I can check UA strings via the
server's .htaccess file. I do some measure of UA checking, both for
browsers or platforms that are *really* old, or repeated visits showing
UA strings whose syntax is unusual (e.g., odd spacing, caps use, or
other evidence of hand-crafting). These are all evidence of bot
activity, and dropping connections that show those kinds of UAs is a
relatively easy way of discarding at least some bot traffic hitting my site.

It's interesting to see just how old some of the stuff is, including
things like Firefox 1.x, IE 5, Windows 98, and I think I even see a few
hits from bots showing MacOS 9.x. As for Google Chrome, they update
frequently enough and their processes are automatic enough that if a
visitor is showing a UA string that is more than a handful of 1.x
versions older than the current version, chances are pretty high that
that's a bot. Right now, I don't have a problem with dropping
connections that claim Chrome versions earlier than 100.x.

One of my favorites is one bot that claims to be "Firefox 30.1", which
was never a valid release number.

Knowing that there are legitimate users of trailing-edge browsers, I'm
much more conservative about ones I know about, especially Gecko
browsers, but for how this page is used, it's going to be pretty
unlikely that we see a legitimate user running something older than
Firefox 45.x ESR

> 2 - Google love adding new features to Javascript.  Other browsers such
> as Firefox have to follow suit because websites - in particular ones
> with dynamically generated pages like using these wonderful (or not) new
> commands.  Firefox has the resources to handle the new syntax, Seamonkey
> does not and is much slower implementing the commands.
> Firefox rewrote their rendering engine a year or two ago and cut
> Thunderbird adrift as part of the process.  Seamonkey = Firefox +
> Thunderbird (+ Composer + other-bits) so FF and TB using different
> rendering engines is a problem.  Firefox has also made some other
> changes but those are not as relevant to this discussion.

That's the bigger challenge, of what Google is doing with JavaScript.
Some of that is capacities that Chrome is able to support, but a lot of
it comes from the scripting that is actually being done on web pages
(and assuming Google support from Chrome). And for good measure, a lot
of the scripting is outsourced to external hosting providers rather than
necessarily a lot of work that is being done by web page developers.

I've seen a few sites that have more than 20 scripting hosts that they
call, and not all of that scripting is related to tracking and
advertising. Some is related to content delivery (especially content
delivery networks), and some is also related to page functionality,
ranging from things like login and authentication processes, display
(especially animation) and validation of data that goes into forms

For the most part Firefox seems to be keeping up, but I think there are
some number of developers that have the attitude of "if it works in
Chrome, it works", and where Firefox (even current versions) has a small
enough number of users to be considered to be irrelevant, and not worth
the effort of making sure it all works. However, I've seen a few sites
that will not work with any version of Firefox, and where the only
option is to use Chrome, or at least some sort of Chromium-derived
browser, such as Edge, Opera or Epic. I haven't spent enough time with
Safari to see if they have issues with those kinds of sites or not.

Smith

On 24/05/2023 11:11, Dirk Munk wrote:
> Don Spam's Reckless Son wrote:
>> Mike Spencer wrote:
>...>
>>
>> 2 - Google love adding new features to Javascript.
>
> So, apart from the browser sniffing, the problem lies with Javascript.
>
> I've tried to understand what the problem is, and it seems that the most
> recent version of the Javascript language is EcmaScript 2022, if I
> understand that correctly.
>
> Am I to understand that the Javascript engine in Seamonkey does not
> confirm to this standard?

From <https://kangax.github.io/compat-table/es2016plus/> (etc), SM is on
a par with major desktop browsers for ES6/2015 and fully supports all
the new features in ES2016. Many features from 2017-2021 are also
supported, but 2022+ support is absent. For 2016+ overall, SM vs latest
FF/Chrome is a bit like the height of your HD TV vs its width. However
SM is 55 times better than IE11!

Why is this? The cretins/monopolists who design changes to the JS
specification introduce new syntax instead of extending the library
system and with no standard way of negotiating JS language version or
use of syntax extensions. This ensures that any browser implementer who
was lucky enough to get one year's crop of changes working is screwed
the next year.

By patching JS code with user scripts, SM's 2016+ compliance score can
be pushed from 55% to 63%, but further compliance requires actual
interpreter changes.

Refer to the linked page above for links to the definitions of the
features listed below.

Patchable (SM scripts, plundered and hacked from the Web, available):
* Symbol.prototype.description (2019): fails empty Symbol() test
* String.prototype.matchAll (2020)
* Promise.any (2021)
* .at() method for built-in indexables (2022)
* Object.hasOwn (2022)
* Error.cause property (2022): hard work
* Array find from last (2023).

Not patchable:
* shared memory and atomics (2017)
* RegExp Perly update (2018)
* BigInt (2020, at least for literals and operators)
* WeakRef (2021)
* Logical assignment (2021)
* numeric separators (2021)
* static and instance class fields (2022)
* private class methods and field checks (2022)
* class static initialization blocks (2022)
* RegExp Match Indices (2023).

/df

--
London
UK

Don Spam's Reckless Son <hyperspace.flyover@vogon.gov.invalid> writes:

> Mike Spencer wrote:
>
>> frg <frgrahl@gmx.net> writes:
>>
>>> Well decades-long is 2015 till today and I am not the only one
>>> working on it :) In any case you will need a second browser for bad
>>> sites for the foreseeable future. This will not go away any time
>>> soon unless we get more contributors and less whining about the bad
>>> world out there. Don't like it either.
>>
>> Just where does the problem lie within Seamonkey? Failure to parse
>> and execute some locutions of javascript? Failure to correctly
>> interpret STYLE blocks?
>
> This has been discussed here before, and I don't know JS so this is
> going to be generic. There are two problems, one relatively minor and
> one anything but:
>
> 1 - Browser sniffing. Sites deciding to support particular versions of
> particular browsers, and nothing else.

Yes, I knew about that but haven't had (don't have) a clear picture of
how extensive it is.

> 2 - Google love adding new features to Javascript.

Ow. Okay, I suspected it was (rapid?) changes in js. I didn't know
that it was somehow attributable to Google. I thought there was a
reference standard (ECMAScript?) for js. But I suppose that, with
so very many people using gmail and the other kewl Google services
together with the popularity of Chrome, they can get away with that.

> Other browsers such as Firefox have to follow suit because websites
> - in particular ones with dynamically generated pages like using
> these wonderful (or not) new commands.

Jeez, think what hacking would have been like in the 80s & 90s if C
had morphed and mutated at some company's whim.

> Firefox has the resources to handle the new syntax, Seamonkey
> does not and is much slower implementing the commands.

Dang. Too bad. Firefox has eliminated features I depend on and has
added stuff (AFAICT) that I don't want.

> Firefox rewrote their rendering engine a year or two ago and cut
> Thunderbird adrift as part of the process.

I wasn't aware that the multiple products within Mozilla were walled
apart like that. Apparently, I don't keep up.

> Seamonkey = Firefox + Thunderbird (+ Composer + other-bits) so FF
> and TB using different rendering engines is a problem.

I can see that. I didn't know all this. I use Seamonkey only as a
browser and occasionally to read files I've saved as HTML. I do
email and news in Emacs and write HTML 4 by hand for my simple web
page.

Thanks for the enlightenment. I'm too old and increasingly slow to
learn to write js, never mind alone contributing to a
parser/compiler/interpreter.

> Firefox has also made some other changes but those are not as
> relevant to this discussion.

--
Mike Spencer Nova Scotia, Canada

Barryedwin1 schrieb:
>> https://addons.thunderbird.net/thunderbird/downloads/file/1012011/open_with-6.8.6-sm+tb.xpi

Thank you for the hint, that add-on indeed eases things a little.

CU

Rainer

Dirk Fieldhouse wrote:

> From <https://kangax.github.io/compat-table/es2016plus/> (etc), SM is on
> a par with major desktop browsers for ES6/2015 and fully supports all
> the new features in ES2016. Many features from 2017-2021 are also
> supported, but 2022+ support is absent.

Here's one I ran into recently:

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Nullish_coalescing_assignment

it says it's supported in FF 79, but it doesn't work in SM 2.53.16

it affects multiple websites because they all use popular JavaScript
libraries

Scratchpad:

cachedValue = cachedValue ?? []; // works

cachedValue ??= []; // syntax error

gerry 666uk wrote:
> Dirk Fieldhouse wrote:
>
>>  From <https://kangax.github.io/compat-table/es2016plus/> (etc), SM is on
>> a par with major desktop browsers for ES6/2015 and fully supports all
>> the new features in ES2016. Many features from 2017-2021 are also
>> supported, but 2022+ support is absent.
>
> Here's one I ran into recently:
>
> https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Nullish_coalescing_assignment
>
>
> it says it's supported in FF 79, but it doesn't work in SM 2.53.16
>
> it affects multiple websites because they all use popular JavaScript libraries
>
> Scratchpad:
>
> cachedValue = cachedValue ?? [];    // works
>
> cachedValue ??= [];         // syntax error
>

Try current 2.53.17b1 pre as of two days ago.

90% of the current problems seem to be because of missing support for dynamic
imports and older regexp interpreter. I am working my way thru the js engine
patches but it is a bit slow. Unfortunately needed and need to take out
support for some really obsolete syntax which broke and might still break some
add-ons.

IanN is busy updating gfx code among other things and we hope that we can one
day in the not so far future switch to the new stylo rust css parser which
should take care of missing later css features. If it were not for the Rust
language with it deprecations and everchanging syntax it would already be on.

We have help switching the build system to python 3 and this means we will not
be kicked out of Linux distributions because of it.

FRG

frg wrote:
> gerry 666uk wrote:
>> Dirk Fieldhouse wrote:
>>
>>>  From <https://kangax.github.io/compat-table/es2016plus/> (etc), SM
>>> is on
>>> a par with major desktop browsers for ES6/2015 and fully supports all
>>> the new features in ES2016. Many features from 2017-2021 are also
>>> supported, but 2022+ support is absent.
>>
>> Here's one I ran into recently:
>>
>> https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Nullish_coalescing_assignment
>>
>>
>> it says it's supported in FF 79, but it doesn't work in SM 2.53.16
>>
>> it affects multiple websites because they all use popular JavaScript
>> libraries
>>
>> Scratchpad:
>>
>> cachedValue = cachedValue ?? [];    // works
>>
>> cachedValue ??= [];         // syntax error
>>
>
> Try current 2.53.17b1 pre as of two days ago.
>
> 90% of the current problems seem to be because of missing support for
> dynamic imports and older regexp interpreter. I am working my way thru
> the js engine patches but it is a bit slow. Unfortunately needed and
> need to take out support for some really obsolete syntax which broke and
> might still break some add-ons.
>
> IanN is busy updating gfx code among other things and we hope that we
> can one day in the not so far future switch to the new stylo rust css
> parser which should take care of missing later css features. If it were
> not for the Rust language with it deprecations and everchanging syntax
> it would already be on.
>
> We have help switching the build system to python 3 and this means we
> will not be kicked out of Linux distributions because of it.
>
> FRG
Thanks for all your hard work Frank.

On 25/05/2023 18:33, gerry 666uk wrote:
>...>
>
> Here's one I ran into recently:
>
> https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Nullish_coalescing_assignment

That's "Logical assignment" in the list.

/df

--
London
UK

Ant wrote on 24/5/23 4:00 am:
> meagain <rick0.merrill@gmail.com> wrote: ...
>> Hey FRG, thanks for all you have done for SeaMonkey. I am sorry my
>> prodding came over as whining, but I've been a supporter since
>> Netscape.
>
> Ditto.
>
When I first connected to the Internet (1996), my brand-new ISP gave me
two 5.25 inch floppies. One had the Windows dial-up stuff and the other
had Netscape Navigator Ver 0.9.

On and on through to Netscape Communicator Ver 7.0, through to Mozilla
Suite and on to SeaMonkey Suite.

Along the way I have used FireFox and Thunderbird a little bit and been
forced to use MicroSoft stuff a little bit a work!!

Thanks SM Devs for all the work you guys do!! It *IS* appreciated!!
--
Daniel

Daniel65 <daniel47@nomail.afraid.org> wrote:
> Ant wrote on 24/5/23 4:00 am:
> > meagain <rick0.merrill@gmail.com> wrote: ...
> >> Hey FRG, thanks for all you have done for SeaMonkey. I am sorry my
> >> prodding came over as whining, but I've been a supporter since
> >> Netscape.
> >
> > Ditto.
> >
> When I first connected to the Internet (1996), my brand-new ISP gave me
> two 5.25 inch floppies. One had the Windows dial-up stuff and the other
> had Netscape Navigator Ver 0.9.

Wow, I didn't know they still used 5.25" floppies at that time. I
remember 3.5" floppies though. I don't remember anyone using 5.25"
during that year. Are you sure you're not thinking of 3.5"?

> On and on through to Netscape Communicator Ver 7.0, through to Mozilla
> Suite and on to SeaMonkey Suite.

> Along the way I have used FireFox and Thunderbird a little bit and been
> forced to use MicroSoft stuff a little bit a work!!

> Thanks SM Devs for all the work you guys do!! It *IS* appreciated!!

Ditto.
--
"Whenever Aaron enters the Holy Place, he will bear the names of the sons of Israel over his heart on the breastpiece of decision as a continuing memorial before the LORD." --Exodus 28:29. Will this week be slammy & cold week like last week? So behind & much 2 do! Let's remember USA's dead military peep. Finally, Heat burned Celtics away althought that should had been a sweep.
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org.
/ /\ /\ \ Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

Ant wrote on 30/5/23 4:31 pm:
> Daniel65 <daniel47@nomail.afraid.org> wrote:
>> Ant wrote on 24/5/23 4:00 am:
>>> meagain <rick0.merrill@gmail.com> wrote: ...
>>>> Hey FRG, thanks for all you have done for SeaMonkey. I am sorry
>>>> my prodding came over as whining, but I've been a supporter
>>>> since Netscape.
>>>
>>> Ditto.
>>>
>> When I first connected to the Internet (1996), my brand-new ISP
>> gave me two 5.25 inch floppies. One had the Windows dial-up stuff
>> and the other had Netscape Navigator Ver 0.9.
>
> Wow, I didn't know they still used 5.25" floppies at that time. I
> remember 3.5" floppies though. I don't remember anyone using 5.25"
> during that year. Are you sure you're not thinking of 3.5"?

Yeap! Big Black things. I'm sure, if I looked, I'd still have them
somewhere here-abouts!!

>> On and on through to Netscape Communicator Ver 7.0, through to
>> Mozilla Suite and on to SeaMonkey Suite.
>
>> Along the way I have used FireFox and Thunderbird a little bit and
>> been forced to use MicroSoft stuff a little bit a work!!
>
>> Thanks SM Devs for all the work you guys do!! It *IS*
>> appreciated!!
>
> Ditto.
> --
Daniel

Daniel65 wrote:

> Ant wrote:
>
>> Wow, I didn't know they still used 5.25" floppies at that time. I
>> remember 3.5" floppies though. I don't remember anyone using 5.25"
>> during that year. Are you sure you're not thinking of 3.5"?
>
> Yeap! Big Black things. I'm sure, if I looked, I'd still have them
> somewhere here-abouts!!

ITYM medium-sized black things, fairly sure I got rid off all my 8"
floppies.

Andy Burns <usenet@andyburns.uk> wrote:
> Daniel65 wrote:

> > Ant wrote:
> >
> >> Wow, I didn't know they still used 5.25" floppies at that time. I
> >> remember 3.5" floppies though. I don't remember anyone using 5.25"
> >> during that year. Are you sure you're not thinking of 3.5"?
> >
> > Yeap! Big Black things. I'm sure, if I looked, I'd still have them
> > somewhere here-abouts!!

> ITYM medium-sized black things, fairly sure I got rid off all my 8"
> floppies.

When I was in high school, I didn't know they existed until I saw them in person with an IBM PC. Wow, they were so big (TWSS). :)
--
"Remember your leaders, who spoke the word of God to you. Consider the outcome of their way of life and imitate their faith." --Hebrews 13:7. Itchy, tiredy, slammy, and allergy Tuesday.
Note: A fixed width font (Courier, Monospace, etc.) is required to see this signature correctly.
/\___/\ Ant(Dude) @ http://aqfl.net & http://antfarm.home.dhs.org.
/ /\ /\ \ Please nuke ANT if replying by e-mail.
| |o o| |
\ _ /
( )

ant@zimage.comANT (Ant) wrote:

> Andy Burns wrote:
>
>> fairly sure I got rid off all my 8" floppies.
>
> When I was in high school, I didn't know they existed until I saw
> them in person with an IBM PC. Wow, they were so big (TWSS). :)
Similarly most phones now use micro/nano SIMs so people think mini-SIMs
are the full-sized version, rather than credit card-sized ones my first
GSM phone used.

frg wrote:

>> Scratchpad:
>>
>> cachedValue = cachedValue ?? [];    // works
>>
>> cachedValue ??= [];         // syntax error
>>
>
> Try current 2.53.17b1 pre as of two days ago.

I have now tested the same code in Seamonkey 2.53.17b1pre, and the
assignment operators are WORKING.

This is good news, because nearly every website uses minified
JavaScript, and this shorthand is commonly used in minification.