These are chat archives for gitterHQ/desktop

16th
Apr 2015
Jan Henke
@janhenke
Apr 16 2015 14:27
hi folks, please fix the linux x64 package, the remove script is broken (by returning 1 during the script run) which leaves the package in a broken state
Mike Bartlett
@mydigitalself
Apr 16 2015 14:28
hey jan
Jan Henke
@janhenke
Apr 16 2015 14:28
that is an absolute no go, before publishing please make sure it is at least cleanly removeable again
Mike Bartlett
@mydigitalself
Apr 16 2015 14:28
how much do you know about making debian packages... ?
Jan Henke
@janhenke
Apr 16 2015 14:29
depends what aspect you are asking about, I am trying to learn packing currently
Mike Bartlett
@mydigitalself
Apr 16 2015 14:29
the reason i ask is we're not massive Linux users ourselves, other than command line server stuff
Jan Henke
@janhenke
Apr 16 2015 14:30
hmm then you shouldn't have piblished that package, no package is better than a broken package
Omer Katz
@omerzimp
Apr 16 2015 14:30
@mydigitalself I think that's when fpm can help a lot
Mike Bartlett
@mydigitalself
Apr 16 2015 14:30
we had tested it previously and were able to remove it fine, but perhaps our knowledge of purge wasn't that good
@omerzimp we use fpm for the packaging and i was digging around after I saw @janhenke's tweet and couldn't see how it distringuised between a purge and a remove
Jan Henke
@janhenke
Apr 16 2015 14:31
well bash does not even like the after-{install,removal}.sh files when running them manually
Andy Trevorah
@trevorah
Apr 16 2015 14:31
not even like?
Jan Henke
@janhenke
Apr 16 2015 14:32
@mydigitalself a pruge is a removal of the package+ all configuration files
Mike Bartlett
@mydigitalself
Apr 16 2015 14:32
it has the --after-remove directive but i would assume that --after-purge is more aggressive and didn't see a way to distinguish between the two
Jan Henke
@janhenke
Apr 16 2015 14:32
normal removal keeps the configuration files in case you want to use it again later
Mike Bartlett
@mydigitalself
Apr 16 2015 14:32
yeah i get that
but unless we can figure out how to perform different actions on purge with fpm, then the only option i see is removing all files on a delete as well.
Jan Henke
@janhenke
Apr 16 2015 14:33
as I said your bash files are broken, try running them on the cli
it does not like the array definition
Mike Bartlett
@mydigitalself
Apr 16 2015 14:35
is the after-install one not working either?
Jan Henke
@janhenke
Apr 16 2015 14:35
try running them by hand sh after-removal.sh
ah and by the way, you should not create a new symlink in the first place
if you require a different so file name, you applciation is pretty much broken
Mike Bartlett
@mydigitalself
Apr 16 2015 14:37
we use nw.js and afaik that's specific to nw.js? @malditogeek ?
Jan Henke
@janhenke
Apr 16 2015 14:38
I would suggest you either create a deb file directly, not via a framework or wrapper, or provide a tar ball and leave the installation to the user
the current package is not really helpful to people
Mike Bartlett
@mydigitalself
Apr 16 2015 14:38
anyhoo @janhenke was going to suggest if you fancied helping out with a PR? we're far from experts in this domain and certainly one of the reasons we published this whole repo as open source was for others to get involved and improve it.
Jan Henke
@janhenke
Apr 16 2015 14:39
sorry @mydigitalself I do not have the time to look into that too, my schedule is already overly ful with stuff
also technically speaking that app is not really what I would consider a desktop app, it is more a little broswer hardcoded to run the gitter webpage
I expected a real client in gtk or qt
Jan Henke
@janhenke
Apr 16 2015 14:48
I am just wondering what is the advantage of using this "app"? it currently just seems like having another browser window open and does not provide any advantage over using your normal browser in the first place
Jan Henke
@janhenke
Apr 16 2015 14:53
is the desktop app for windows and mac the same? i.e. just the normal gitter webpage in an own window?
Andy Trevorah
@trevorah
Apr 16 2015 14:54
at the moment, yeah
Jan Henke
@janhenke
Apr 16 2015 14:55
sorry to ask, why do you bother publishing them then?
Andy Trevorah
@trevorah
Apr 16 2015 14:57
having a separation from the browser means that we can do more than what we can do with just a browser window, and also means that we dont have to hope that the user has an up to date browser
Jan Henke
@janhenke
Apr 16 2015 14:58
but that is hardly archived with running a webpage in an own rendering engine, you should write a native app then
Andy Trevorah
@trevorah
Apr 16 2015 14:59
We would if we were a larger team.
Jan Henke
@janhenke
Apr 16 2015 14:59
are you all "web developers"?
Andy Trevorah
@trevorah
Apr 16 2015 14:59
"web developers"?
Jan Henke
@janhenke
Apr 16 2015 15:00
does anybody in your team have a computer science background? studies software engineering, knows C/C++/Java/C#?
Andy Trevorah
@trevorah
Apr 16 2015 15:01
Yeah, pretty much
I have to ask, where exactly are you trying to lead with these questions?
Jan Henke
@janhenke
Apr 16 2015 15:04
I am just trying to understand the reason why you are trying to use web technologies in a desktop app, which seems like a bad idea
Andy Trevorah
@trevorah
Apr 16 2015 15:08
Because we have a very small team and a existing webapp, we can get a web based desktop app live sooner than if we created native apps from scratch for windows, linux and osx
Jan Henke
@janhenke
Apr 16 2015 15:08
well imho that is not a desktop app at all
I mean it is fine if the service is web only
Andy Trevorah
@trevorah
Apr 16 2015 15:09
which we are
you cant really chat without an internet connection
Jan Henke
@janhenke
Apr 16 2015 15:09
it just seems disappointing to annouce a desktop app, which is nothing more than what I have in my browser anyway
at least for me it was very disappointing, not having a desktop app at all would have been better in my perception
besides the obvious waste of resources to run a desktop app in javascript
Andy Trevorah
@trevorah
Apr 16 2015 15:20
We have more people that use the desktop app than firefox
I understand that 100% native apps are an advantage
but so is having something that we can opensource and get into anyones hands
Jan Henke
@janhenke
Apr 16 2015 15:24
still thinking that desktop app does not deserve that name and I am pretty disappointed about it (in addition to the time it took to manually clean the system again)
Mike Bartlett
@mydigitalself
Apr 16 2015 15:41
@janhenke so a few points on this:
1) there are additional benefits to having an app beyond "it's just native". for instance our webapp has bindings into system notifications, tray/dock, starts on boot, has a "home" outside of the browser so if I want to get to Gitter I just cmd+tab to Gitter and and it feels like it has a place. the same reason that I, like many other people, run and even pay for MailPlane even though it's effectively just a wrapper around a web application. potentially it can interact with other portions of the system too that we currently don't do.
2) our decision to do this is largely down to resourcing. having to build native apps for Mac, Windows, Linux, iOS, Android, Windows Phone is costly. not only one-off costs, but continual maintenance - add a new feature you have to add it to a 6 different platforms, effectively the same thing, but still you have to implement it 6 different times. the added benefit as well is that you introduce a new feature in the web interface and the next page load it's there, for everybody. no waiting for people to update. you can do bucket/multivariate testing easily on the web too, which is an absolute nightmare on the desktop.
3) this approach is becoming increasingly common as JavaScript and browser rendering continues to improve and we're continuing to move toward the "browser as the operating system" paradigm we've been seeing and is core to Google's strategy and so to, quite frankly rather insultingly draw some false line between web developers and computer scientists is rather disingenuous. we are computer scientists and the running of our entire system from back-end to front-end means we are faced with lots of hard problems all the time. web development today is miles away from the brochure-ware of the 90's that you seem to associate with web development.
Jan Henke
@janhenke
Apr 16 2015 16:06
@mydigitalself the "the browser is the new os" thing I consider as the biggest mistake and the most serious design error of today
javascript will never be a serious competitor to langauges like C++/Java or C#
Mike Bartlett
@mydigitalself
Apr 16 2015 17:21

@janhenke in 2001 would never have believed that somebody could build something like MS Office in a browser, I've been purely using Google Docs for 3 years now.

And on the server-side, it's running at scale like you wouldn't believe:
http://www.nearform.com/nodecrunch/node-js-becoming-go-technology-enterprise/

http://blog.shinetech.com/2013/10/22/performance-comparison-between-node-js-and-java-ee/

DINESH S
@Dineshs91
Apr 16 2015 17:40
@janhenke I agree with @mydigitalself statements. For instance take a look at atom text editor, which is build entirely on web technologies.
@janhenke The main deciding factor, is the time that needs to be invested and what kind of application you are developing.
Jan Henke
@janhenke
Apr 16 2015 19:06
I don't say it isn't possible to build such things in browser technologies, I do strongly think it is a big mistake and a huge waste of resources
of course current hardware is often much more powerful than many people need, but that does not mean it is a good idea to consume more resources than nessesary
and running javascript is always an overhead vs running the same thing in native code
just saying javascript is not the solution to every problem, it is fine for dynamic content in websites (where C is an awful idea), but in other places languages like C are a far better idea than javascript
you are free to have your opinion on that matter, I am fully aware I won't change that opinion, just having to express my opinion though that I consider it a bad idea