DownThemAll! Add-on ID: {DDC359D1-844A-42a7-9AA1-88A850A938A8} Contact info: https://addons.mozilla.org/en-US/firefox/user/nils-maier/ or www.downthemall.net Assignee: Jim Mathies Last update: July 2015 How well does it work?: ~90% - very feature rich addon, hard to say if it is working 100% - downloading all files appears to work - download file pane works - sound feedback works Any obvious performance problems? none - no cpow use in about:perf Chromium version: does not appear to exist.

also, sliding notifications work and right-click page options appear to work

This addon required a restart, so I'm assuming it's a xul addon.

Reopening this bug. The install is flawless and also the icons pop up in the customize section but the functionality is completely dead. You can not download anything and the buttons that appear in the customize section are not firing off any actions.

This is not e10s specific as we get the same functionality with it active and disabled. We encountered this issue on Win 7, Mac 10.11, and Ubuntu 14.04.

Tested on Nightly: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.11; rv:48.0) Gecko/20100101 Firefox/48.0 Mozilla/5.0 (Windows NT 6.1; WOW64; rv:48.0) Gecko/20100101 Firefox/48.0 Mozilla/5.0 (X11; Linux x86_64; rv:48.0) Gecko/20100101 Firefox/48.0

Last good revision: 0814f2d30eda6175980be9318b8ed4a7fccaf435 First bad revision: 733d3f5c161975dcd8ed476a7c923dad42c117eb Pushlog: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=0814f2d30eda6175980be9318b8ed4a7fccaf435&tochange=733d3f5c161975dcd8ed476a7c923dad42c117eb

Confirming also fails in Aurora: Version 47.0a2 Build ID 20160323004040 User Agent Mozilla/5.0 (X11; Linux x86_64; rv:47.0) Gecko/20100101 Firefox/47.0

For what it's worth, I've had e10s force-enabled since 46 (I run the release version on both Windows and OS X), and DTA has been working without any noticeable issue since 46's release. about:support shows "1/1 (Enabled by user)", so I know e10s is enabled.

Just tried downloading random stuff with Nightly 50.0a1. Downloaded files will not have their modified date set to the the server's last modified date even with the feature checked.

(In reply to bluejedi from comment #10) > Just tried downloading random stuff with Nightly 50.0a1. Downloaded files > will not have their modified date set to the the server's last modified date > even with the feature checked. That's a bug reported even by non-e10s users. From our regular feedback channels so far, we heard next to no complaining about e10s, so the shims might be working for now (or there are just no e10s users). However, our trunk nightly uses frame scripts since some time. It may or may not be e10s compatible already. https://code.downthemall.net/nightly/ But it is a major rewrite of sorts too however, going bootstrap and doing off-the-main-thread I/O (for the most part). If anybody wants to test that too... With the nightly (to be DownThemAll! 3.0), so far I have been waiting for e10s to settle (and being a little disgruntled with mozilla in general for various reasons). But now that e10s, and everything that goes along with it, seems to be somewhat stable, it's time to finish and release that version I guess.

(In reply to Nils Maier [:nmaier] from comment #11) > (In reply to bluejedi from comment #10) > > Just tried downloading random stuff with Nightly 50.0a1. Downloaded files > > will not have their modified date set to the the server's last modified date > > even with the feature checked. > That's a bug reported even by non-e10s users. > > > From our regular feedback channels so far, we heard next to no complaining > about e10s, so the shims might be working for now (or there are just no e10s > users). > > However, our trunk nightly uses frame scripts since some time. It may or may > not be e10s compatible already. > https://code.downthemall.net/nightly/ > But it is a major rewrite of sorts too however, going bootstrap and doing > off-the-main-thread I/O (for the most part). If anybody wants to test that > too... > > With the nightly (to be DownThemAll! 3.0), so far I have been waiting for > e10s to settle (and being a little disgruntled with mozilla in general for > various reasons). But now that e10s, and everything that goes along with it, > seems to be somewhat stable, it's time to finish and release that version I > guess. Hi Nils, =] your timing is great. I'm going through highly popular add-ons and checking for e10s compatibility. e10s made it into the release channel in 48 for a % of users (excluding ones with add-ons in the first release to limit variables). the numbers and performance are looking good - so add-on inclusion is on the horizon. we are testing the most used add-ons and if/when you have a version you think is ready for testing... we can test it across several platforms for install, launch, basic functionality, and performance. Thank you for being patient. We're definitely investing in the add-on experience (for developers and users) and webextensions continue to include more APIs https://wiki.mozilla.org/Firefox/AddOns/Status/current . we've got "office hours" every Tuesday if you have technical questions about e10s/add-ons. people will be available to assist you every Tuesday in the #addons channel at irc.mozilla.org. We're here to help! http://atsay.github.io/e10s_office_hours/

>webextensions continue to include more APIs Good for them, I don't see it that in the future DownThemAll! can be ported to WebExtensions while keeping most features. >Thank you for being patient. I am losing some of my patience with each bad mozilla decision of which there have been plenty lately. >we are testing the most used add-ons and if/when you have a version you think is ready for testing... we can test it across several platforms for install, launch, basic functionality, and performance. Our AMO API signed Nighlies should get some testing. https://code.downthemall.net/nightly/ ("trunk" version) It is almost complete, I'm currently mostly wrangling with handling download failures in case of disk full, and making sure stuff does not break when upgrading to a new version (since we are restartless now) and stuff does not break when upgrading from the old 2.0.x versions. And waiting for the great volunteers of babelzilla to (hopefully) finish their locales. DownThemAll! 3.0 (current "trunk") will be released most likely before 1st of September because the mozilla Featured Add-ons Board notified me that I got till that date (less than a month time) to port to e10s, while a Firefox feature gets 18 weeks at minimum to go through the release train, or several years in case of e10s. One of those bad mozilla decisions I mentioned earlier. e10s is mostly undocumented except for the most common case (accessing a content window) requiring me to read tons of code as low level as the js engine to understand what is going on where and why e.g. why nsContextMenu is allowed to do CPOW in "popupshowing" and I am not without generating a warning... >http://atsay.github.io/e10s_office_hours/ Doesn't really state a time zone btw.

Hi Nils, Let me know if my answers are bothering you - i'm trying to help - but know you've been talking to folks for a while. The timezones for the office hours are shown according to your machines' timezone (so the first ones are folks in Europe). I will put a note to test from the location you gave me Sept 1. i don't know if the new documentation that is coming out - like matching SDK to webextensions https://developer.mozilla.org/en-US/Add-ons/WebExtensions/Comparison_with_the_Add-on_SDK or the wiki site with status updates on upcoming webextensions and add-ons/e10s status help https://wiki.mozilla.org/Firefox/AddOns/Status/current#Add-ons_Webextensions_Status_Report

@shell the fact that the timezones are local to the user is nice, but the page doesnt state that, so there is no way for someone to know that without asking

Also this page is terrifying: http://arewee10syet.com are we really saying that we are a month out, and several huge add-ons are not supported? I use: - DownThemall - uBlock Origin - Stylish - HTTPS Everywhere Between this and the new lamebar: http://superuser.com/q/1111480/is-it-possible-to-get-old-suggestion-system-back-in-firefox-48-0 I am about done. If any of the above add-ons are dropped, I am switching to Chrome. Privacy be damned, I dont care anymore.

(In reply to :shell escalante from comment #14) > Hi Nils, Let me know if my answers are bothering you - i'm trying to help - > but know you've been talking to folks for a while. What is bothering me a little indeed is the constant stream of meaningless "sorry for any inconveniences" / "thanks for your patience" style phrases. Please know that this is not all directed at you personally, but at the amalgamation of messages I get from mozilla representatives, and that I am in fact appreciating that you're trying to help. Pair this with a constant stream of users/developers/"people" writing me to express discontent with mozilla decisions (to put it mildly) because I seem like a good guy to write because of my somewhat public opposition to XPI signatures and a WebExtensions-only world and that I was (and still am to a degree) a loyal mozilla supporter and contributors since before there even was Phoenix, and you probably can see how all this stuff is beginning to eat at me... > I will put a note to test from the location you gave me Sept 1. Actually, mozilla (Scott DeVaney; Editorial Manager AMO) says we have to release Sept 1. the very latest, or we'll be unfeatured. So testing that day will not help find issues prior to our release. To quote: > If you'd like your add-on to continue to be featured, please make any adjustments necessary to make your > content e10s compatible (without shims)—as well as 49 version compatible—and submit a new version of your > add-on by September, 1 2016. After September 1, any featured add-on that is not compatible to e10s or > Firefox 49 will be removed. I'm actually not sure if we can make that date either... There are some delays on babelzilla, the XPI signing security theater not especially helping our localizers do their jobs, e.g. the major es, fr, ja locales are not ready still, and getting this new version with massive changes through code review will be a hurdle in itself. > i don't know if the new documentation that is coming out - like matching SDK > to webextensions > https://developer.mozilla.org/en-US/Add-ons/WebExtensions/ > Comparison_with_the_Add-on_SDK DownThemAll! never was an SDK add-on, so offtopic for us at least. May be helpful to authors of SDK add-ons tho. Still, the "Add-on SDK => WebExtensions" section having a lot of "None" even for some highlevel APIs is not exactly reassuring... > or the wiki site with status updates on > upcoming webextensions and add-ons/e10s status help > https://wiki.mozilla.org/Firefox/AddOns/Status/current#Add- > ons_Webextensions_Status_Report ... and adding this status information in is not either. Looking at the current WebExtensions MDN docs, while keeping the other two pages you mentioned in mind, what is offered is not even remotely close to being able to port the most basic DTA features like doing an actual custom download. E.g. "The downloads API is far to limited to be useful for you" is what the Google Chrome people told me back when they asked me if it was possible to port DownThemAll! to a Chrome extension... And the mozilla implementation is currently only a re-implementation of that API.

DownThemAll! 3.0.1 was reviewed and is available from AMO. If you are still willing to spend some time to test (the e10s stuff in particular), go ahead ;)

Thanks Nils, requesting testing today and will let you know the results next week. for Steven, comment 16 - if it helps at all, all of those add-ons are now multi-process (e10s) compatible. for Stu - comment 19 - i own writing a blog on how we plan to do introduce add-ons to e10s in stages (just took on that role). won't fix the past, but hopefully will help clarify for the next 12 months what the plan is and what drove the decisions.

@shell this page is ****: http://arewee10syet.com ---------- - DownThemAll - Compatible: unknown - Shims: Yes - CPOW: 63631 ---------- - uBlock Origin - compatible: compatible - Shims: Yes - CPOW: 26264 Does this mean it only works with a shim? Site is not clear. ---------- - Stylish - compatible: compatible - Shims: Yes - CPOW: 204 Same shims question, and what the **** is CPOW? ---------- - HTTPS Everywhere - compatible: compatible - Shims: No - CPOW: 0 Now this has Shims = No in red, but is a shim even needed if it is already compatible? and we have CPOW = 0 in green, but still no **** idea what CPOW is, because it is not explained on that page, or even the page that is supposed to explain what it is: http://github.com/andymckay/new-arewee10syet.com Get your **** together.

(In reply to Steven Penny from comment #21) > Does this mean it only works with a shim? Site is not clear. It means that a for a version of uBlock Origin telemetry has shown that shims have been recorded. > Now this has Shims = No in red, but is a shim even needed if it is already > compatible? It means that no shims have been recorded in telemetry CPOW are explained here: https://developer.mozilla.org/en-US/Firefox/Multiprocess_Firefox/Cross_Process_Object_Wrappers The shim data is pulled from telemetry which is shown in the repository here: https://github.com/andymckay/new-arewee10syet.com/blob/master/shim-perf.ipynb I will add some explanations of those to the pages if that helps. > Get your **** together. These kind of comments really don't help.

@Andy, that was a quite long comment that answered nothing. Viewing http://arewee10syet.com presents obvious questions to me, but since they are not obvious to you, here they are: 1. If DownThemAll is e10s compatible, if so why doesnt the page reflect that? 2. If Shims = Yes, does that mean said add on requires a shim? 3. If Shims = No, does that mean that a shim is not required, or that a shim is required but has not been written? 4. What, in plain english, is CPOW? Note that dumping a link to a 1500 word page is not an acceptable answer to this question

(In reply to Steven Penny from comment #23) > @Andy, that was a quite long comment that answered nothing. Viewing > http://arewee10syet.com presents obvious questions to me, but since they are > not > obvious to you, here they are: > > 1. If DownThemAll is e10s compatible, if so why doesnt the page reflect that? Not sure, probably caching the arewee10syet build cache or the status of DownThemAll changed since the site was built - I note that the add-on has been updated a few times in the last couple of days. I'll double check in the next run. > 2. If Shims = Yes, does that mean said add on requires a shim? It means that telemetry at some point has recorded a shim. So yes. However I think I should expand the site to include the version info. What's important is if the current version has recorded a shim, older versions are less relevant. > 3. If Shims = No, does that mean that a shim is not required, or that a shim > is > required but has not been written? It means that telemetry has not recorded a shim. The add-on might make use of them, but never triggered one. It probably makes sense to change "No" from red to green? > 4. What, in plain english, is CPOW? Note that dumping a link to a 1500 word > page > is not an acceptable answer to this question The document starts with the intro "Cross Process Object Wrappers (CPOWs), which enable chrome code to synchronously access content in multiprocess Firefox." I'm not sure what more I can say that would be more accurate than that page (chances are it would be less so).

> “CPOW” stands for “Cross-process Object Wrapper”1, and is part of the glue > that has allowed e10s to be enabled on Nightly without requiring a full > re-write of the front-end code. It’s also part of the magic that’s allowing a > good number of our most popular add-ons to continue working (albeit slowly). http://mikeconley.ca/blog/2015/02/17/on-unsafe-cpow-usage-in-firefox-desktop-and-why-is-my-nightly-so-sluggish-with-e10s-enabled This is a shim, in my opinion. So essentially, CPOW is a shim for both Firefox itself and addons for Firefox. If this in fact the case, the next obvious question is this: Why is is e10s being "bolted on" to Firefox in the form of CPOWs, which are essentially wrappers/shims/glue code, instead of being a "full rewrite". This seems like a hacky/lazy way to implement what appears will be a major part of Firefox going forward.

Can you discuss this somewhere else, i.e. not in a bug I am watching, please?

(In reply to Nils Maier [:nmaier] from comment #26) > Can you discuss this somewhere else, i.e. not in a bug I am watching, please? Sounds good, please take it to https://mail.mozilla.org/listinfo/dev-addons or file issues on arewee10syet.com directly https://github.com/andymckay/new-arewee10syet.com/issues. Sorry for the noise.

(In reply to Andy McKay [:andym] from comment #22) > It means that a for a version of uBlock Origin telemetry has shown that > shims have been recorded. I am not aware of uBlock Origin ever needing a shim, I am surprised to reports of shims being used. Given that uBO was first developed on Chromium, there has always be a clear separation between the main and content process, and no non-standard JS objects are messaged between the main and content processes. Is there a way for me to find out more details about what could have caused telemetry to report this?

We've been updating with more data and it now shows changes by version: http://arewee10syet.com/#15, that might give a clue. But please let's take this to another bug or email thread.

Any chance you could open a webext-port-dta meta bug, similar to what you have with adp?

With Firefox 57 only WebExtensions are permitted and are, by default, e10s compatible.

(In reply to Firefox Product Integrity Bug Husbandry Bot (contact :emceeaich) from comment #31) > With Firefox 57 only WebExtensions are permitted and are, by default, e10s > compatible. So.. It doesn't mean there couldn't be other issues with an addon? Bug 467520 is an example.

*cough*cough* bug1226546, comment2

I'm not sure why you needinfo'd me, there wasn't a question? (Fwiw I have nothing to do with the addons team at Mozilla, so I'm likely not the right person to ask anyway)

Oh, sorry. Nonetheless I don't see why this (with even more users of TM+) still hasn't had a whiteboard and its own tracker issue.