On Sun, Feb 3, 2019 at 11:50 AM Laurent <laurent.ors...@gmail.com> wrote:
> I've pushed a repair. It should be integrated with the next racket nightly > release I suppose. > > The fix is to check the compiled version of all scripts in the library, > and if it differs from the current racket version it gets recompiled. The > bonus is that scripts that have not been compiled yet will now be, which > means faster loading time when triggering them. > > The caveat is that dependencies are not recompiled. This means that if a > script depends on a package/collection that hasn't been migrated/updated > yet, an error may be displayed on DrRacket startup (but appart from this > script DrRacket should still work fine). > Can this error be caught and captioned with a better message or, even better, perhaps be integrated somehow into quickscript? (Maybe putting something into the menu or some other place somehow?) > Feedback more than welcome. > > @Robby: I've tested this manually, by installing an old version of racket, > and compiling scripts with the old one while running the new version of > drracket, but I'm not sure how to automate this properly (as in "write a > portable test file"). Have you done something like this before? > Maybe save an old compiled file, putting it into place and then starting up drr the way that other test suites start it up, and checking that it starts up successfully? Robbg > > On Fri, Feb 1, 2019 at 1:37 PM Laurent <laurent.ors...@gmail.com> wrote: > >> Okay then, I'll work on a fix tomorrow. >> >> On Fri, Feb 1, 2019 at 1:23 PM Robby Findler <ro...@eecs.northwestern.edu> >> wrote: >> >>> I can't say concretely what will go wrong I am sorry to say. There are >>> just a number of moving parts and I can't quite see how they will fail to >>> fit together. I am just basing this on my vague, holistic understanding of >>> how things work currently. (There are no places where on set of sources is >>> used with multiple versions in the standard distribution and nontrivial >>> work has gone into making that case work out and maintain that invariant.) >>> >>> In other words, go for it! We have a fallback if this don't work out. :) >>> >>> Robby >>> >>> On Fri, Feb 1, 2019 at 7:01 AM Laurent <laurent.ors...@gmail.com> wrote: >>> >>>> On Fri, Feb 1, 2019 at 11:30 AM Robby Findler < >>>> ro...@eecs.northwestern.edu> wrote: >>>> >>>>> My feeling is that our compilation infrastructure isn't really set up >>>>> to work like that currently. In general things will probably be smoother >>>>> if >>>>> there is only the expectation that a given set of files is used with only >>>>> one version at a time. >>>>> >>>>> It is possible that someone may use multiple versions of DrRacket on >>>>> their computer switching between them or even having them open at the same >>>>> time. >>>>> >>>> >>>> Frankly, for any application, if one has multiple versions open at the >>>> same time and things go wrong, I don't see how one can blame the developers >>>> with a straight face—the one legitimate case is probably when you're one of >>>> the developers yourself. >>>> >>>> If you want to use several versions but not at the same time (which is >>>> also asking for trouble in any case), then recompilation on startup should >>>> work nicely anyway, just requiring a little more time. Copying the same >>>> scripts in different folders will very likely mean divergence of source at >>>> some point and likely frustration for the user. >>>> >>>> Or am I missing something? >>>> >>>> >>>>> >>>>> Robby >>>>> >>>>> On Fri, Feb 1, 2019 at 4:12 AM Laurent <laurent.ors...@gmail.com> >>>>> wrote: >>>>> >>>>>> Thanks for bringing this up. >>>>>> >>>>>> Indeed the compile option from the quickscript menu should work, but >>>>>> it's not ideal. >>>>>> >>>>>> I'm not a big fan of the copy option. (To me that sounds like asking >>>>>> for trouble, but I may be wrong.) >>>>>> >>>>>> A nicer solution may be to simply check if the scripts are compiled >>>>>> with the current version, and if not just silently compile them on >>>>>> DrRacket >>>>>> startup? The user could also be querried if that's better. >>>>>> >>>>>> On Thu, Jan 31, 2019, 23:02 Greg Trzeciak <gtrzec...@gmail.com wrote: >>>>>> >>>>>>> As long as the user-scripts won't be lost (i.e. the scripts will be >>>>>>> copied over) it sounds OK. >>>>>>> >>>>>>> On Thursday, January 31, 2019 at 10:42:08 PM UTC+1, Robby Findler >>>>>>> wrote: >>>>>>>> >>>>>>>> Probably it is best if quickscript (either silently or after >>>>>>>> asking) >>>>>>>> copied scripts from a version-specific directory to a new >>>>>>>> (version-specific) one, much like how the pkg system works. >>>>>>>> >>>>>>>> Robby >>>>>>>> >>>>>>>> On Thu, Jan 31, 2019 at 3:38 PM Greg Trzeciak <gtrz...@gmail.com> >>>>>>>> wrote: >>>>>>>> > >>>>>>>> > I had quickscript already installed when upgrading Racket. >>>>>>>> > >>>>>>>> > On DrRacket startup I got the following error: >>>>>>>> > >>>>>>>> > Error in script file "run-in-interactions.rkt": >>>>>>>> read-compiled-linklet: version mismatch >>>>>>>> > expected: "7.2" >>>>>>>> > found: "6.90.0.24" >>>>>>>> > in: >>>>>>>> C:\Users\Myself\AppData\Roaming\Racket\quickscript\user-scripts\compiled\run-in-interactions_rkt.zo >>>>>>>> >>>>>>>> > >>>>>>>> > After closing the message box DrRacket is opening with no >>>>>>>> problems >>>>>>>> > >>>>>>>> > Deleting compiled folder removes the error on the next start up >>>>>>>> (and so would probably running "Compile Scripts and reload" from >>>>>>>> Script >>>>>>>> menu). >>>>>>>> > But since quickscript is now part of DrRacket - the error may >>>>>>>> reappear in the next upgrades since it looks like user-scripts is not >>>>>>>> recompiled during upgrade. >>>>>>>> > >>>>>>>> > G. >>>>>>>> > >>>>>>>> > -- >>>>>>>> > You received this message because you are subscribed to the >>>>>>>> Google Groups "Racket Users" group. >>>>>>>> > To unsubscribe from this group and stop receiving emails from it, >>>>>>>> send an email to racket-users...@googlegroups.com. >>>>>>>> > For more options, visit https://groups.google.com/d/optout. >>>>>>>> >>>>>>> -- >>>>>>> You received this message because you are subscribed to the Google >>>>>>> Groups "Racket Users" group. >>>>>>> To unsubscribe from this group and stop receiving emails from it, >>>>>>> send an email to racket-users+unsubscr...@googlegroups.com. >>>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>>> >>>>>> -- >>>>>> You received this message because you are subscribed to the Google >>>>>> Groups "Racket Users" group. >>>>>> To unsubscribe from this group and stop receiving emails from it, >>>>>> send an email to racket-users+unsubscr...@googlegroups.com. >>>>>> For more options, visit https://groups.google.com/d/optout. >>>>>> >>>>> -- >>>> You received this message because you are subscribed to the Google >>>> Groups "Racket Users" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to racket-users+unsubscr...@googlegroups.com. >>>> For more options, visit https://groups.google.com/d/optout. >>>> >>> -- > You received this message because you are subscribed to the Google Groups > "Racket Users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to racket-users+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
