My only bit against this is that when we use xpcshell we help improving our product.
The part where we "polyfill the xpcshell api to work like node" is absolutely not necessary, should we want to stay on xpcshell. I never properly understood why we did this. Also the longer term goal has always bean to be able to run the build system inside Firefox (eg WebIDE). That said, WebIDE is simply running a shell command these days. So I don't have a strong opposition, but let's just properly evaluate the pros and the cons. Le 23/11/2015 16:27, Fred Lin a écrit : > Recently we have plenty of discussion around moving to the Next > Generation Architecture. So I guess its the right time to ask some > bold questions. > > I'd like to rise a cliche topic about our corner stone, the build > script, which still use xpcshell as our main build tool. Lots of work > is consumed by polyfilling the xpcshell api to work like node. I feel > we shot our shoes to not use node directly. It slow the pace when our > frontend developers need to take care of api/compatibility issues for > xpcshell, while outside web developer could just pick and use those > tools without brain. As the result, it move us away from outside web > developers. > > Here are my 3 Questions (for different audience): > > 1. is it bother you when you can't use node in the main build procedure? > 2. Is there still any concern that stop us switch from xpcshell to > node? Or just use xpcshell when needed and use node for general build > operation? > 3. is it a significant effort to switch from xpcshell to node? > > -- > Fred > > > _______________________________________________ > dev-fxos mailing list > [email protected] > https://lists.mozilla.org/listinfo/dev-fxos
signature.asc
Description: OpenPGP digital signature
_______________________________________________ dev-fxos mailing list [email protected] https://lists.mozilla.org/listinfo/dev-fxos
