Yes, you are right Sven. When I disable network the Spotter is not freezing. Do we have a way to do the DNS resolution asynchronously? I mean, is it doable in Pharo 7 or it is a big problem. I’m not an expert in the questions that require VM, but it seams that http requests can be executed without blocking the UI thread, is it possible to resolve DNS in the same way?
Uko P.S. why there are TWO checkboxes in Settings for the catalog plugin in Spotter? And they are not synchronized by the way, so you need to turn on both. > On 9 Mar 2017, at 09:41, Sven Van Caekenberghe <[email protected]> wrote: > > Guys, > > Please stop speculating, no it is not Monticello, no it is not HTTP access, > no it is not networking itself, it is DNS resolution, in a very specific > situation: > > https://pharo.fogbugz.com/f/cases/18281/NetNameResolver-class-addressForName-sometimes-hangs-when-there-is-no-network > > Note also that many, many people deploy Pharo in production contexts with > heavy multi threading and networking loads with great success - that would > not be possible if we did not get the basics right. > > Sven > >> On 9 Mar 2017, at 09:31, Yuriy Tymchuk <[email protected]> wrote: >> >> Dimitris, >> >> but AFAIK when Spotter freezes it’s not using Metacello. It does an http >> request to pharo catalog. Isn’t this a problem with http request blocking >> the only thread that we have? >> >> Uko >> >>> On 9 Mar 2017, at 08:23, Dimitris Chloupis <[email protected]> wrote: >>> >>> I do not need to, the problem is obvious as the sun when one uses >>> Monticello which can easily freeze the image on a slow connection. Still >>> it's a Monticello problem, not a Pharo problem. But then that's not the >>> most serious problem with Monticello. >>> >>> On Tue, 7 Mar 2017 at 17:04, [email protected] <[email protected]> wrote: >>> I see you are not teaching to people or running demos in a crowd with Pharo. >>> >>> Otherwise, you would get the point I can tell you 100% >>> >>> Phil >>> >>> On Tue, Mar 7, 2017 at 3:19 PM, Dimitris Chloupis <[email protected]> >>> wrote: >>> Frankly I fail to see the problem, if a tool connects to the internet and >>> does not have some sort of timeout or speed check then is the obligation of >>> the tool to fix that and not of Pharo to offer an airplane mode. Its would >>> make an already complex pharo environment even more complex. Not to exclude >>> the fact that having two mode will create useless confusion for beginners. >>> >>> Offer a new mode for Pharo ? No >>> Offer methods that check internet connectivity and internet speed for pharo >>> tool developers to use ? Yes >>> >>> I feel that Pharo can be simplified down to 1% without the need for a big >>> compromise on features. Maybe it does not matter for us that are used to >>> working with Blender but its a huge deal for beginners and if we want to >>> attract a lot more new users. >>> >>> My dream is that version 8 be dedicated ONLY to simplifying the Pharo code, >>> by removing all the extra fat. No new bug fixes (unless created by the >>> removal of the code), no new features. I know we have bootstrap , but thats >>> a completely different thing. >>> >>> On Tue, Mar 7, 2017 at 3:15 PM [email protected] <[email protected]> >>> wrote: >>> https://pharo.fogbugz.com/f/cases/19818/Support-for-Airplane-mode-in-Pharo >>> for keeping track of this >>> >>> On Tue, Mar 7, 2017 at 1:02 AM, Ben Coman <[email protected]> wrote: >>> On Tue, Mar 7, 2017 at 6:35 AM, [email protected] <[email protected]> >>> wrote: >>>> This is not only when teaching. >>>> >>>> I was at a conference today and showcased some Grafoscopio. >>>> >>>> Spotty wifi and crowded bandwith makes it bad looking. >>>> >>>> We need an "Airplane mode" switch. >>> >>> That is a cool way to describe it. A well know paradigm. >>> Its a setting that probably would change daya to day more than any other, >>> so perhaps it would even be reasonable to have this as a permanent >>> mode button on the background, >>> or the task bar much like MS Windows taskbar status icons. >>> >>> cheers -ben >>> >>>> >>>> Maybe can we have that available in the session from the settings. >>>> >>>> Phil >>>> >>>> On Mon, Mar 6, 2017 at 10:55 PM, Yuriy Tymchuk <[email protected]> >>>> wrote: >>>>> >>>>> The problem is that there are people who give lectures in universities >>>>> with bad internet (internet is bad in most of the universities I tried) >>>>> and >>>>> they done want freezes. On the other hand we have people who sit in their >>>>> offices with fast machines, fast internet and they want features. Until >>>>> now >>>>> we were looking at how to satisfy one of the groups and which one exactly. >>>>> Why not to try finding a solution for both. For example we can have the >>>>> plugging enabled by default and have Spotter measuring the time needed to >>>>> do >>>>> a query. Whenever the time surpasses frustration limit (10 sec? I duno) >>>>> Spotter will ask if the user wants to turn off the plugin, of turns it of >>>>> automatically and tells where to re-enable it. The other way would be to >>>>> have known plugins to appear at the bottom of Spotter, and a user can >>>>> install/enable them with one click. So when you search but there is >>>>> nothing >>>>> you know that there are more plugins that may solve your problem. >>>>> >>>>> Yes it is more complicated to implement in in a smart way rather them just >>>>> set the default boolean value of a setting, but we spend too much time >>>>> anyway convincing the other side that we are right… >>>>> >>>>> Uko >>>>> >>>>>> On 6 Mar 2017, at 21:30, stepharong <[email protected]> wrote: >>>>>> >>>>>> One day I would like to invite you to give a lecture in a not that bad >>>>>> french university >>>>>> with a not that good network. Each time Pharo freezes you pay me 5 Euros >>>>>> only and we can >>>>>> go and pay some really expensive restaurant with your money. >>>>>> Simple on my machine it times out every five minutes. >>>>>> >>>>>> <sorry I could not resist> >>>>>> >>>>>>> >>>>>>>> On 6 Mar 2017, at 19:51, stepharong <[email protected]> wrote: >>>>>>>> >>>>>>>> We should make it available from the catalog :) >>>>>>> >>>>>>> Yes, it most certainly should be. >>>>>>> >>>>>>> If only we could then tell user to type >>>>>>> >>>>>>> Shift-Enter Sublimish Enter >>>>>>> >>>>>>> to install it using Spotter right from the catalog ;-) >>>>>>> >>>>>>> Sorry, I couldn't resist ... <ducking> >>>>>>> >>>>>>>> 2017-03-06 9:29 GMT+01:00 Esteban Lorenzano <[email protected]>: >>>>>>>> >>>>>>>>> On 6 Mar 2017, at 09:22, Pavel Krivanek <[email protected]> >>>>>>>>> wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> 2017-03-06 8:20 GMT+01:00 Esteban Lorenzano <[email protected]>: >>>>>>>>> >>>>>>>>>> On 6 Mar 2017, at 07:57, [email protected] wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> https://ci.inria.fr/pharo/job/Pharo-6.0-Issue-Validator-simple/7294/artifact/report.html >>>>>>>>>> is not integrated for some Monkey reason anyway - still should b/c >>>>>>>>>> if you >>>>>>>>>> put sublimish, you need a decent matching Spotter. >>>>>>>>>> >>>>>>>>>> Also, I see >>>>>>>>>> https://pharo.fogbugz.com/f/cases/19796/Remove-duplicate-code-from-SublimishTheme >>>>>>>>>> on 5/3. If this moves forward, why not the other one? >>>>>>>>>> >>>>>>>>>> BTW Sublimish was an extra thing. Who got that in image? >>>>>>>>> >>>>>>>>> I don’t know. >>>>>>>>> It shouldn’t :) >>>>>>>>> >>>>>>>>> What we should do is to create a page for themes and links to it (and >>>>>>>>> they should be in catalog) :) >>>>>>>>> >>>>>>>>> Well, it was me who proposed the slice and yes, I was aware of the >>>>>>>>> fact that I should not do it because of code freeze (I was not aware >>>>>>>>> of the >>>>>>>>> two themes policy). >>>>>>>> >>>>>>>> it is not a policy, it is just the idea I have :) >>>>>>>> >>>>>>>>> >>>>>>>>> We can remove it again, I will prepare a slice. >>>>>>>> >>>>>>>> problem with it is that is makes ripples. >>>>>>>> If community want it (is nice), we can include it, but I would wait >>>>>>>> for P7, yes. >>>>>>>> >>>>>>>> https://pharo.fogbugz.com/f/cases/19801/Remove-SublimishTheme >>>>>>>> At least it was an advertisement for this theme that moved it forward >>>>>>>> and the integration of it helped to improve the dark theme too. And >>>>>>>> showed >>>>>>>> again how much we need the concept of palettes used strictly in the >>>>>>>> system. >>>>>>>> >>>>>>>> -- Pavel >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> Esteban >>>>>>>> >>>>>>>>> >>>>>>>>> -- Pavel >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> Esteban >>>>>>>>> >>>>>>>>>> >>>>>>>>>> Phil >>>>>>>>>> >>>>>>>>>> On Mon, Mar 6, 2017 at 7:43 AM, Esteban Lorenzano >>>>>>>>>> <[email protected]> wrote: >>>>>>>>>> >>>>>>>>>>> On 6 Mar 2017, at 07:17, Esteban Lorenzano <[email protected]> >>>>>>>>>>> wrote: >>>>>>>>>>> >>>>>>>>>>> Hi, >>>>>>>>>>> >>>>>>>>>>> First, let me be clear: I *like* sublimish theme (not for my use, >>>>>>>>>>> but I like it), and I like (a lot) the windows list on spotter, and >>>>>>>>>>> I like >>>>>>>>>>> in general all this new additions. >>>>>>>>>> >>>>>>>>>> By the way, even if I like it, I would not have integrated >>>>>>>>>> SublimishTheme. For same reason I didn’t integrated DarkMetalTheme: >>>>>>>>>> My idea >>>>>>>>>> is to have just two themes *in image*: one white, one dark, and >>>>>>>>>> allow people >>>>>>>>>> to install other themes from catalog. >>>>>>>>>> >>>>>>>>>> Why? Because each class inside image becomes our responsibility and >>>>>>>>>> then a maintainability issue. With themes in particular already >>>>>>>>>> happened >>>>>>>>>> that existing themes became abandonware inside the image and we >>>>>>>>>> needed to >>>>>>>>>> clean up, not without problems. >>>>>>>>>> >>>>>>>>>> cheers, >>>>>>>>>> Esteban >>>>>>>>>> >>>>>>>>>> ps: which means also yes: that issue got integrated because I failed >>>>>>>>>> to control the flow of issues... >>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> And I know, most of those additions are usability enhancements so a >>>>>>>>>>> good thing. >>>>>>>>>>> >>>>>>>>>>> But can you guys stop pushing new features into a (in theory) >>>>>>>>>>> frozen version? >>>>>>>>>>> >>>>>>>>>>> Pharo 7 will arrive soon… *if* we can fix all things missing. If I >>>>>>>>>>> can suggest an orientation, it would be good, for those that use >>>>>>>>>>> linux and >>>>>>>>>>> mac (windows is not ready), to test and help fix 64bits version. >>>>>>>>>>> >>>>>>>>>>> There are still plenty of bugs to fix there, like for example why >>>>>>>>>>> Spotter does not shows search line. >>>>>>>>>>> >>>>>>>>>>> Last week I added support to Athens and SDL2 in 64bits but that >>>>>>>>>>> meant: >>>>>>>>>>> >>>>>>>>>>> - some important changes in UFFI >>>>>>>>>>> - discover something that seems is failing on callbacks for Cairo >>>>>>>>>>> (no idea if it is a callbacks problem or an image problem that >>>>>>>>>>> causes a >>>>>>>>>>> callback problem). >>>>>>>>>>> >>>>>>>>>>> I would say all this and the fact we are working on 64bits in >>>>>>>>>>> general deserves a lot of testing. >>>>>>>>>>> >>>>>>>>>>> I will add links in Pharo page to download 64bits (I’m at the train >>>>>>>>>>> now), but in the mean time you can download all from here: >>>>>>>>>>> >>>>>>>>>>> http://files.pharo.org/get-files/60/pharo-64.zip >>>>>>>>>>> http://files.pharo.org/get-files/60/pharo64-linux-latest.zip >>>>>>>>>>> (linux) >>>>>>>>>>> http://files.pharo.org/get-files/60/pharo64-mac-latest.zip (mac) >>>>>>>>>>> http://files.pharo.org/get-files/60/sources.zip (in case you need >>>>>>>>>>> them) >>>>>>>>>>> >>>>>>>>>>> cheers, >>>>>>>>>>> Esteban >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> -- >>>>>>>> Using Opera's mail client: http://www.opera.com/mail/ >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> Using Opera's mail client: http://www.opera.com/mail/ >>>>>> >>>>> >>>>> >>>> >>> >>> >>> >> > >
