On 2016-02-26 5:52 PM, Fabrice Desré wrote: > On 02/26/2016 02:42 PM, Ehsan Akhgari wrote: > >>> Look at what you need to implement to get the simplest gecko based >>> product that builds with --enable-application=my_great_app and compare >>> it to the same project as an Electron app. >> >> I never said that it is as easy as Electron, did I? :-) Ben explicitly >> said he doesn't care for Electron compat, hence my suggestion. > > But your suggestion is not a reasonable alternative for people looking > at an alternative to Electron, even with no api compat.
Yeah, that's true. You'll have to forgive me but this thread started as a request for the <webview> API, it's sometimes very difficult to reverse engineer what problems people are trying to solve based on the solutions they propose. :-) >>> It's not whether it's doable (especially not whether a gecko engineer >>> can do it) but if it's a reasonable tool for others to use. Gecko is not >>> one unfortunately. >> >> Who are these "others"? If you are talking about a random person who >> wants to create a new desktop app, then yes, that solution is far from >> desirable. If you're talking about a Mozilla engineer wanting to create >> a new product on top of Gecko that is completely separate from Firefox, >> this is the way to do it. If you're implying that in order to satisfy >> the second use case we need to solve the first one, then I'm not sure if >> I agree. > > I say that we should focus more on solving the first use case, since the > second one is ... not really a problem for a Mozilla engineer and > happens rarely. But if your point is that we should not care about 3rd > party devs that want to use gecko, I'm puzzled - that seems like major > shortsightedness coming from a 'platform' team. If I misunderstood, my > apologies. Please note that this is not about what _I_ want as someone on the platform team. I personally think that in hindsight, it was a mistake for us to miss the runtime embedding ship at least twice (once with WebKit and now with Chromium Embedding Framework). For the last few years, our strategy for investing in Gecko has mostly been as a vehicle for investing in the end user facing products that we build, so a *large* portion of the platform team have been working on things intended to support Firefox OS and now Firefox. The cost of building and maintaining a successful product based around Gecko is quite significant, and I'm afraid that we just don't currently have enough investment in Gecko in order to be able to sustain this. Therefore, effectively when someone asks "can we build something like Electron on top of Gecko", they're asking for a whole new set of investments from a team who is working hard to just keep up with the day to day activities already promised to the product teams. In this sense, I agree with Myk. This is not a technical issue, the ask here is to shift what we invest in as an organization, and I personally agree that it would be nice to have that (and have more investment in Gecko.) I just don't have a good answer, since promising to build what's being asked here with our current organization-wide investments is dishonest and will just set people up for failure. But it also pains me to have to recommend competing Web platform implementations for this use case... :( _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
