Getting attention to unowned modules, especially Core::Editor, Core::Serializers and Core::Spelling checker

[Jörg Knobloch]

Dear Mitchell,

I am writing to you to voice the following concern:

Many users not only use Firefox to "consume" web content (incl. audio
and video) but also to participate in social networks (Facebook,
Twitter), write e-mail (for example using Gmail) or do online shopping
and online banking.

For these activities they enter/type/paste information into text fields
of a web page displayed in Firefox. The core modules which handle the
text input are the modules mentioned in the subject: Core::Editor,
Core::Serializers and Core::Spelling checker.

These modules are currently unowned and have a plethora of obvious,
reproducible and annoying bugs which puzzle users and make Firefox look
bad, also in comparison with Chrome [1].

In October 2015 Ehsan stepped down as the owner of Core::Editor [2] and
this role has been vacant ever since. I don't understand the internals
of Mozilla well enough, but IMHO it seems unfortunate that these *vital*
modules receive no visible attention. What puzzles me is that some
internal refactoring does take place, at least in Core::Editor [3], but
that does not deliver improvements to the users.

I request that an owner be assigned or hired for these modules and that
these bugs get properly triaged and the more significant ones receive
some attention.

Kind regards from Berlin, Germany,
Jorg K (Mozilla contributor - www.jorgk.com/misc/Mozilla-work.pdf)

[1]
https://bugzilla.mozilla.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&component=Editor&component=Serializers&component=Spelling%20checker&list_id=12876575&product=Core&query_format=advanced&resolution=---&order=bug_status%2Cpriority%2Cassigned_to%2Cbug_id&limit=0
(1105 bugs, of which perhaps 5% are worth fixing)
[2]
https://groups.google.com/forum/#!searchin/mozilla.dev.platform/ehsan$20editor/mozilla.dev.platform/GKZX7CrUbnY/b1sxG9ZpBgAJ
[3] https://bugzilla.mozilla.org/show_bug.cgi?id=1156062
https://bugzilla.mozilla.org/show_bug.cgi?id=1190172
https://bugzilla.mozilla.org/show_bug.cgi?id=1191354
https://bugzilla.mozilla.org/show_bug.cgi?id=1191356

[Jet Villegas]

Thanks for posting this!

I share your concern about un-owned Core components. Ehsan Akhgari, Andrew
Overholt, and I are meeting next week to make plans for Core::Editor. I'll
make a note to also discuss Serializer and SpellChecker.

I think our first priority is to establish a clear Module Owner for Editor,
then triage the bugs and draft a project roadmap for that component. We'll
then need to determine the list of contributors who can help put code
behind those plans. Can we add you to that list?

--Jet

> _______________________________________________
> governance mailing list
> gover...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/governance
>

[Jörg Knobloch]

On 26/02/2016 19:01, Jet Villegas wrote:
> I think our first priority is to establish a clear Module Owner for Editor,
> then triage the bugs and draft a project roadmap for that component. We'll
> then need to determine the list of contributors who can help put code
> behind those plans. Can we add you to that list?
>
> --Jet

Sure.

As you can see from the link in my signature, I have already fixed a few
bugs in these areas, look for the boxes "M-C / Gecko: Editor problems"
and "M-C / Gecko: Spell checker clean-up". Sadly my skills are limited,
so I have to rely on guidance from others.

It would be good to find a module owner and responsive reviewer(s) like
Ehsan and Robert O'Callahan have been in the past (and hopefully still
are during the transition).

On a different issue:
Some of the code in the editor/serialisers/etc. dates back to the
Netscape days, when the program had a browser and integrated e-mail
client. If you're drafting a roadmap, you might consider clearly
identifying the parts only maintained for what has become Thunderbird
and transfer those into the responsibility of the Thunderbird
maintainers, if possible. Just some examples from the top of my head:
grep -i mail editor/libeditor/*.cpp
grep -i mail dom/base/nsIDocumentEncoder.idl
grep -i mail dom/base/nsPlainTextSerializer.cpp
mozTXTToHTMLConv::ScanHTML - OK, different module, but you get the idea.

[Ting-Yu Lin]

FWIW, Core::Selection also has no owner. Please consider find a owner for
this module.

Ting-Yu

[Jörg Knobloch]

Hello Jet,

May I ask whether you've made any progress on this issue? Ting-Yu Lin
rightly pointed out that Core::Selection is also important and unowned.

Kind regards, Jörg.

On 26/02/2016 19:01, Jet Villegas wrote:

> Thanks for posting this!
>
> I share your concern about un-owned Core components. Ehsan Akhgari, Andrew
> Overholt, and I are meeting next week to make plans for Core::Editor. I'll
> make a note to also discuss Serializer and SpellChecker.
>

> I think our first priority is to establish a clear Module Owner for Editor,
> then triage the bugs and draft a project roadmap for that component. We'll
> then need to determine the list of contributors who can help put code
> behind those plans. Can we add you to that list?
>
> --Jet
>

> On Fri, Feb 26, 2016 at 6:51 AM, Jörg Knobloch <jo...@jorgk.com> wrote:
>
>> Dear Mitchell,
>>
>> I am writing to you to voice the following concern:
>>
>> Many users not only use Firefox to "consume" web content (incl. audio and
>> video) but also to participate in social networks (Facebook, Twitter),
>> write e-mail (for example using Gmail) or do online shopping and online
>> banking.
>>
>> For these activities they enter/type/paste information into text fields of
>> a web page displayed in Firefox. The core modules which handle the text
>> input are the modules mentioned in the subject: Core::Editor,
>> Core::Serializers and Core::Spelling checker.
>>
>> These modules are currently unowned and have a plethora of obvious,
>> reproducible and annoying bugs which puzzle users and make Firefox look
>> bad, also in comparison with Chrome [1].
>>
>> In October 2015 Ehsan stepped down as the owner of Core::Editor [2] and
>> this role has been vacant ever since. I don't understand the internals of
>> Mozilla well enough, but IMHO it seems unfortunate that these *vital*
>> modules receive no visible attention. What puzzles me is that some internal
>> refactoring does take place, at least in Core::Editor [3], but that does
>> not deliver improvements to the users.
>>
>> I request that an owner be assigned or hired for these modules and that
>> these bugs get properly triaged and the more significant ones receive some
>> attention.
>>
>> Kind regards from Berlin, Germany,

>> Jorg K (Mozilla contributor - www.jorgk.com/misc/Mozilla-work.pdf)
>>

>> [1]
>> https://bugzilla.mozilla.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&component=Editor&component=Serializers&component=Spelling%20checker&list_id=12876575&product=Core&query_format=advanced&resolution=---&order=bug_status%2Cpriority%2Cassigned_to%2Cbug_id&limit=0
>> (1105 bugs, of which perhaps 5% are worth fixing)
>> [2]
>> https://groups.google.com/forum/#!searchin/mozilla.dev.platform/ehsan$20editor/mozilla.dev.platform/GKZX7CrUbnY/b1sxG9ZpBgAJ
>> [3] https://bugzilla.mozilla.org/show_bug.cgi?id=1156062
>> https://bugzilla.mozilla.org/show_bug.cgi?id=1190172
>> https://bugzilla.mozilla.org/show_bug.cgi?id=1191354
>> https://bugzilla.mozilla.org/show_bug.cgi?id=1191356

[Daniel Glazman]

Le vendredi 26 février 2016 19:02:02 UTC+1, Jet Villegas a écrit :
> Thanks for posting this!
>
> I share your concern about un-owned Core components. Ehsan Akhgari, Andrew
> Overholt, and I are meeting next week to make plans for Core::Editor. I'll
> make a note to also discuss Serializer and SpellChecker.
>
> I think our first priority is to establish a clear Module Owner for Editor,
> then triage the bugs and draft a project roadmap for that component. We'll
> then need to determine the list of contributors who can help put code
> behind those plans. Can we add you to that list?

I concur to everything Jörg said, and Ting-Yu is perfectly right
mentioning Selection. A lot of bugs filed into the editor basket
are really related to Selection, and that's the reason why the
original editor team at netscape had one expert of the selection
and caret codes.

Since you mentioned plans, and since embedding is not a blocker
any more as it used to be eons ago, I have a recommendation: move
the whole machinery to JavaScript and make it extensible. I have
something in mind wrt extensibility but this is not the right place
or moment to discuss it.
I think this is totally doable, at minimal perf expense, and it would
allow MANY MORE contributors to the editor. I understand it's an
investment but I think it's a necessary one, and has been a necessary
one for a decade and a half.

The serializer is, IMHO, in worse condition than the editor. Its
pretty-printing option - among others - is plagued with so many issues
most of my changes to mozilla-central are done there. Its entity
output options are not enough. It was already a semi-abandoned module
back in 2002...
It could be a bit harder to move it to JS, but I think we would still
greatly benefit from such a move.

</Daniel>