Re: Code splittingmysteries
FYI I updated http://code.google.com/p/google-web-toolkit/issues/detail?id=7874 with some of my findings. I haven't found a workaround other than reverting to GWT 2.5.0 and am not sure I can dedicate any more time to this right now so will just revert I guess :( On 12 June 2013 19:34, Ed wrote: > Ed, out of interest, are you using Gin? > > No, but probably would use it if I would start "blank". > When I started with GWT (about 7 years ago), there was no GIN, so I had to > make my own "thinks" like factories, which I use nowadays . This all > resulted in having build my own rich gwt framework/lib, which I am happy > about thought, works well.. > But in general I am careful using external gwt lib's there are many out > there being not maintained anymore, which can result in gwt upgrade > problems.. > I started out with mygwt, contributed some things... but then It became > gxt and became half commercial :( > > -- > You received this message because you are subscribed to the Google Groups > "Google Web Toolkit" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to google-web-toolkit+unsubscr...@googlegroups.com. > To post to this group, send email to google-web-toolkit@googlegroups.com. > Visit this group at > http://groups.google.com/group/google-web-toolkit?hl=en. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
> > Ed, out of interest, are you using Gin? No, but probably would use it if I would start "blank". When I started with GWT (about 7 years ago), there was no GIN, so I had to make my own "thinks" like factories, which I use nowadays . This all resulted in having build my own rich gwt framework/lib, which I am happy about thought, works well.. But in general I am careful using external gwt lib's there are many out there being not maintained anymore, which can result in gwt upgrade problems.. I started out with mygwt, contributed some things... but then It became gxt and became half commercial :( -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
No, it's not you, understanding the details of the SOYC report is hard. I thought it was easy from seeing the Google IO session some years ago, butt... ... It's definitely a step in the right direction but the soyc report still need some optimization steps to be well usable for "bigger" projects (then Hello world) On Wed, Jun 12, 2013 at 5:29 PM, James Horsley wrote: > Yeah. the SOYC report in this case hasn't helped me much either because > I'm not reading it right or it's just not surfacing the right info. The > only 404 I get is regarding the "compiler metrics" link. > > > On 12 June 2013 16:18, Ed Bras wrote: > >> Yep, tried to go back to 2.5, but then I see some other bugs that are >> solved in 2.5.1 ;)... So for me this was no option... >> >> >> On Wed, Jun 12, 2013 at 5:17 PM, Ed Bras wrote: >> >>> Yes, I understand what you mean, I experienced the same... >>> It's hard to hand over a "test case" that fails, to the dev team, such >>> that they can use that to work on :( >>> >>> Yes it would be nice to have better facilities such that we can try to >>> track done the cause our self. Currently the SOYC report isn't helping a >>> lot. >>> To my experience the SOYC report is buggy, as some pages result in a 404 >>> :(.. (I experienced that several times the last years with a correctly >>> created soyc, it's also in some issue if I remember correctly). >>> >>> >>> >>> On Wed, Jun 12, 2013 at 4:45 PM, James Horsley >>> wrote: >>> Yeah. I've tried reproducing the issue with a sample app but thus far haven't been able to; not sure that the sample webapp captures enough of the complexity of the actual webapp we're having the problem with. As I mentioned before can confirm that changing none of the webapp code but switching between 2.5 and 2.5.1 causes everything to end up in the leftover. From the compiler report it's not at all obvious why these classes are in the leftover fragment as they look isolated from what I can tell. On 12 June 2013 13:13, Ed wrote: > Ed, did you ever get this sorted? As I commented on the linked issue >> I'm running into what seems to be a similar problem. > > No. It's hard to track the exact cause. It tried, but had to give up > (a lot of time, and not knowing exactly what happens. > But it's good to see that more people are experiencing the same kind > of problems such that I am not alone ;) > Ray seem to be the code splitting guru. Let's hope he has has some > time to look at this ;) > > If you can isolate the problem in a small program that reproduces this > behavior, please do upload it in the issue. I tried, but failed :( > This will help the gwt dev team a lot to solve the problem (and the > issue will get a higher priority and solved sooner). > > -- > You received this message because you are subscribed to the Google > Groups "Google Web Toolkit" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to google-web-toolkit+unsubscr...@googlegroups.com. > To post to this group, send email to > google-web-toolkit@googlegroups.com. > Visit this group at > http://groups.google.com/group/google-web-toolkit?hl=en. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- You received this message because you are subscribed to a topic in the Google Groups "Google Web Toolkit" group. To unsubscribe from this topic, visit https://groups.google.com/d/topic/google-web-toolkit/_8CdVbBb488/unsubscribe?hl=en . To unsubscribe from this group and all its topics, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out. >>> >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "Google Web Toolkit" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to google-web-toolkit+unsubscr...@googlegroups.com. >> To post to this group, send email to google-web-toolkit@googlegroups.com. >> Visit this group at >> http://groups.google.com/group/google-web-toolkit?hl=en. >> For more options, visit https://groups.google.com/groups/opt_out. >> >> >> > > -- > You received this message because you are subscribed to a topic in the > Google Groups "Google Web Toolkit" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/google-web-toolkit/_8CdVbBb488/unsubscribe?hl=en > . > To unsubscribe from this group and all its topics, send an email to > google-web-toolkit+unsubscr...@googlegroups.com. > To post to this group, send email to google-web-toolkit@googlegroups.com. > V
Re: Code splittingmysteries
Ed, out of interest, are you using Gin? On 12 June 2013 16:29, James Horsley wrote: > Yeah. the SOYC report in this case hasn't helped me much either because > I'm not reading it right or it's just not surfacing the right info. The > only 404 I get is regarding the "compiler metrics" link. > > > On 12 June 2013 16:18, Ed Bras wrote: > >> Yep, tried to go back to 2.5, but then I see some other bugs that are >> solved in 2.5.1 ;)... So for me this was no option... >> >> >> On Wed, Jun 12, 2013 at 5:17 PM, Ed Bras wrote: >> >>> Yes, I understand what you mean, I experienced the same... >>> It's hard to hand over a "test case" that fails, to the dev team, such >>> that they can use that to work on :( >>> >>> Yes it would be nice to have better facilities such that we can try to >>> track done the cause our self. Currently the SOYC report isn't helping a >>> lot. >>> To my experience the SOYC report is buggy, as some pages result in a 404 >>> :(.. (I experienced that several times the last years with a correctly >>> created soyc, it's also in some issue if I remember correctly). >>> >>> >>> >>> On Wed, Jun 12, 2013 at 4:45 PM, James Horsley >>> wrote: >>> Yeah. I've tried reproducing the issue with a sample app but thus far haven't been able to; not sure that the sample webapp captures enough of the complexity of the actual webapp we're having the problem with. As I mentioned before can confirm that changing none of the webapp code but switching between 2.5 and 2.5.1 causes everything to end up in the leftover. From the compiler report it's not at all obvious why these classes are in the leftover fragment as they look isolated from what I can tell. On 12 June 2013 13:13, Ed wrote: > Ed, did you ever get this sorted? As I commented on the linked issue >> I'm running into what seems to be a similar problem. > > No. It's hard to track the exact cause. It tried, but had to give up > (a lot of time, and not knowing exactly what happens. > But it's good to see that more people are experiencing the same kind > of problems such that I am not alone ;) > Ray seem to be the code splitting guru. Let's hope he has has some > time to look at this ;) > > If you can isolate the problem in a small program that reproduces this > behavior, please do upload it in the issue. I tried, but failed :( > This will help the gwt dev team a lot to solve the problem (and the > issue will get a higher priority and solved sooner). > > -- > You received this message because you are subscribed to the Google > Groups "Google Web Toolkit" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to google-web-toolkit+unsubscr...@googlegroups.com. > To post to this group, send email to > google-web-toolkit@googlegroups.com. > Visit this group at > http://groups.google.com/group/google-web-toolkit?hl=en. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- You received this message because you are subscribed to a topic in the Google Groups "Google Web Toolkit" group. To unsubscribe from this topic, visit https://groups.google.com/d/topic/google-web-toolkit/_8CdVbBb488/unsubscribe?hl=en . To unsubscribe from this group and all its topics, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out. >>> >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "Google Web Toolkit" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to google-web-toolkit+unsubscr...@googlegroups.com. >> To post to this group, send email to google-web-toolkit@googlegroups.com. >> Visit this group at >> http://groups.google.com/group/google-web-toolkit?hl=en. >> For more options, visit https://groups.google.com/groups/opt_out. >> >> >> > > -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
Yeah. the SOYC report in this case hasn't helped me much either because I'm not reading it right or it's just not surfacing the right info. The only 404 I get is regarding the "compiler metrics" link. On 12 June 2013 16:18, Ed Bras wrote: > Yep, tried to go back to 2.5, but then I see some other bugs that are > solved in 2.5.1 ;)... So for me this was no option... > > > On Wed, Jun 12, 2013 at 5:17 PM, Ed Bras wrote: > >> Yes, I understand what you mean, I experienced the same... >> It's hard to hand over a "test case" that fails, to the dev team, such >> that they can use that to work on :( >> >> Yes it would be nice to have better facilities such that we can try to >> track done the cause our self. Currently the SOYC report isn't helping a >> lot. >> To my experience the SOYC report is buggy, as some pages result in a 404 >> :(.. (I experienced that several times the last years with a correctly >> created soyc, it's also in some issue if I remember correctly). >> >> >> >> On Wed, Jun 12, 2013 at 4:45 PM, James Horsley >> wrote: >> >>> Yeah. I've tried reproducing the issue with a sample app but thus far >>> haven't been able to; not sure that the sample webapp captures enough of >>> the complexity of the actual webapp we're having the problem with. As I >>> mentioned before can confirm that changing none of the webapp code but >>> switching between 2.5 and 2.5.1 causes everything to end up in the >>> leftover. From the compiler report it's not at all obvious why these >>> classes are in the leftover fragment as they look isolated from what I can >>> tell. >>> >>> >>> On 12 June 2013 13:13, Ed wrote: >>> Ed, did you ever get this sorted? As I commented on the linked issue > I'm running into what seems to be a similar problem. No. It's hard to track the exact cause. It tried, but had to give up (a lot of time, and not knowing exactly what happens. But it's good to see that more people are experiencing the same kind of problems such that I am not alone ;) Ray seem to be the code splitting guru. Let's hope he has has some time to look at this ;) If you can isolate the problem in a small program that reproduces this behavior, please do upload it in the issue. I tried, but failed :( This will help the gwt dev team a lot to solve the problem (and the issue will get a higher priority and solved sooner). -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out. >>> >>> -- >>> You received this message because you are subscribed to a topic in the >>> Google Groups "Google Web Toolkit" group. >>> To unsubscribe from this topic, visit >>> https://groups.google.com/d/topic/google-web-toolkit/_8CdVbBb488/unsubscribe?hl=en >>> . >>> To unsubscribe from this group and all its topics, send an email to >>> google-web-toolkit+unsubscr...@googlegroups.com. >>> To post to this group, send email to google-web-toolkit@googlegroups.com >>> . >>> Visit this group at >>> http://groups.google.com/group/google-web-toolkit?hl=en. >>> For more options, visit https://groups.google.com/groups/opt_out. >>> >>> >>> >> >> > -- > You received this message because you are subscribed to the Google Groups > "Google Web Toolkit" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to google-web-toolkit+unsubscr...@googlegroups.com. > To post to this group, send email to google-web-toolkit@googlegroups.com. > Visit this group at > http://groups.google.com/group/google-web-toolkit?hl=en. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
Yep, tried to go back to 2.5, but then I see some other bugs that are solved in 2.5.1 ;)... So for me this was no option... On Wed, Jun 12, 2013 at 5:17 PM, Ed Bras wrote: > Yes, I understand what you mean, I experienced the same... > It's hard to hand over a "test case" that fails, to the dev team, such > that they can use that to work on :( > > Yes it would be nice to have better facilities such that we can try to > track done the cause our self. Currently the SOYC report isn't helping a > lot. > To my experience the SOYC report is buggy, as some pages result in a 404 > :(.. (I experienced that several times the last years with a correctly > created soyc, it's also in some issue if I remember correctly). > > > > On Wed, Jun 12, 2013 at 4:45 PM, James Horsley wrote: > >> Yeah. I've tried reproducing the issue with a sample app but thus far >> haven't been able to; not sure that the sample webapp captures enough of >> the complexity of the actual webapp we're having the problem with. As I >> mentioned before can confirm that changing none of the webapp code but >> switching between 2.5 and 2.5.1 causes everything to end up in the >> leftover. From the compiler report it's not at all obvious why these >> classes are in the leftover fragment as they look isolated from what I can >> tell. >> >> >> On 12 June 2013 13:13, Ed wrote: >> >>> Ed, did you ever get this sorted? As I commented on the linked issue I'm running into what seems to be a similar problem. >>> >>> No. It's hard to track the exact cause. It tried, but had to give up (a >>> lot of time, and not knowing exactly what happens. >>> But it's good to see that more people are experiencing the same kind of >>> problems such that I am not alone ;) >>> Ray seem to be the code splitting guru. Let's hope he has has some time >>> to look at this ;) >>> >>> If you can isolate the problem in a small program that reproduces this >>> behavior, please do upload it in the issue. I tried, but failed :( >>> This will help the gwt dev team a lot to solve the problem (and the >>> issue will get a higher priority and solved sooner). >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "Google Web Toolkit" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to google-web-toolkit+unsubscr...@googlegroups.com. >>> To post to this group, send email to google-web-toolkit@googlegroups.com >>> . >>> Visit this group at >>> http://groups.google.com/group/google-web-toolkit?hl=en. >>> For more options, visit https://groups.google.com/groups/opt_out. >>> >>> >>> >> >> -- >> You received this message because you are subscribed to a topic in the >> Google Groups "Google Web Toolkit" group. >> To unsubscribe from this topic, visit >> https://groups.google.com/d/topic/google-web-toolkit/_8CdVbBb488/unsubscribe?hl=en >> . >> To unsubscribe from this group and all its topics, send an email to >> google-web-toolkit+unsubscr...@googlegroups.com. >> To post to this group, send email to google-web-toolkit@googlegroups.com. >> Visit this group at >> http://groups.google.com/group/google-web-toolkit?hl=en. >> For more options, visit https://groups.google.com/groups/opt_out. >> >> >> > > -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
Yes, I understand what you mean, I experienced the same... It's hard to hand over a "test case" that fails, to the dev team, such that they can use that to work on :( Yes it would be nice to have better facilities such that we can try to track done the cause our self. Currently the SOYC report isn't helping a lot. To my experience the SOYC report is buggy, as some pages result in a 404 :(.. (I experienced that several times the last years with a correctly created soyc, it's also in some issue if I remember correctly). On Wed, Jun 12, 2013 at 4:45 PM, James Horsley wrote: > Yeah. I've tried reproducing the issue with a sample app but thus far > haven't been able to; not sure that the sample webapp captures enough of > the complexity of the actual webapp we're having the problem with. As I > mentioned before can confirm that changing none of the webapp code but > switching between 2.5 and 2.5.1 causes everything to end up in the > leftover. From the compiler report it's not at all obvious why these > classes are in the leftover fragment as they look isolated from what I can > tell. > > > On 12 June 2013 13:13, Ed wrote: > >> Ed, did you ever get this sorted? As I commented on the linked issue I'm >>> running into what seems to be a similar problem. >> >> No. It's hard to track the exact cause. It tried, but had to give up (a >> lot of time, and not knowing exactly what happens. >> But it's good to see that more people are experiencing the same kind of >> problems such that I am not alone ;) >> Ray seem to be the code splitting guru. Let's hope he has has some time >> to look at this ;) >> >> If you can isolate the problem in a small program that reproduces this >> behavior, please do upload it in the issue. I tried, but failed :( >> This will help the gwt dev team a lot to solve the problem (and the issue >> will get a higher priority and solved sooner). >> >> -- >> You received this message because you are subscribed to the Google Groups >> "Google Web Toolkit" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to google-web-toolkit+unsubscr...@googlegroups.com. >> To post to this group, send email to google-web-toolkit@googlegroups.com. >> Visit this group at >> http://groups.google.com/group/google-web-toolkit?hl=en. >> For more options, visit https://groups.google.com/groups/opt_out. >> >> >> > > -- > You received this message because you are subscribed to a topic in the > Google Groups "Google Web Toolkit" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/google-web-toolkit/_8CdVbBb488/unsubscribe?hl=en > . > To unsubscribe from this group and all its topics, send an email to > google-web-toolkit+unsubscr...@googlegroups.com. > To post to this group, send email to google-web-toolkit@googlegroups.com. > Visit this group at > http://groups.google.com/group/google-web-toolkit?hl=en. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
Quite frustrating as I don't feel like I'm making much progress and am just going to have to go back to 2.5.0 but don't really want to lose out on the bugfixes from 2.5.1 :( On 12 June 2013 15:45, James Horsley wrote: > Yeah. I've tried reproducing the issue with a sample app but thus far > haven't been able to; not sure that the sample webapp captures enough of > the complexity of the actual webapp we're having the problem with. As I > mentioned before can confirm that changing none of the webapp code but > switching between 2.5 and 2.5.1 causes everything to end up in the > leftover. From the compiler report it's not at all obvious why these > classes are in the leftover fragment as they look isolated from what I can > tell. > > > On 12 June 2013 13:13, Ed wrote: > >> Ed, did you ever get this sorted? As I commented on the linked issue I'm >>> running into what seems to be a similar problem. >> >> No. It's hard to track the exact cause. It tried, but had to give up (a >> lot of time, and not knowing exactly what happens. >> But it's good to see that more people are experiencing the same kind of >> problems such that I am not alone ;) >> Ray seem to be the code splitting guru. Let's hope he has has some time >> to look at this ;) >> >> If you can isolate the problem in a small program that reproduces this >> behavior, please do upload it in the issue. I tried, but failed :( >> This will help the gwt dev team a lot to solve the problem (and the issue >> will get a higher priority and solved sooner). >> >> -- >> You received this message because you are subscribed to the Google Groups >> "Google Web Toolkit" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to google-web-toolkit+unsubscr...@googlegroups.com. >> To post to this group, send email to google-web-toolkit@googlegroups.com. >> Visit this group at >> http://groups.google.com/group/google-web-toolkit?hl=en. >> For more options, visit https://groups.google.com/groups/opt_out. >> >> >> > > -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
Yeah. I've tried reproducing the issue with a sample app but thus far haven't been able to; not sure that the sample webapp captures enough of the complexity of the actual webapp we're having the problem with. As I mentioned before can confirm that changing none of the webapp code but switching between 2.5 and 2.5.1 causes everything to end up in the leftover. From the compiler report it's not at all obvious why these classes are in the leftover fragment as they look isolated from what I can tell. On 12 June 2013 13:13, Ed wrote: > Ed, did you ever get this sorted? As I commented on the linked issue I'm >> running into what seems to be a similar problem. > > No. It's hard to track the exact cause. It tried, but had to give up (a > lot of time, and not knowing exactly what happens. > But it's good to see that more people are experiencing the same kind of > problems such that I am not alone ;) > Ray seem to be the code splitting guru. Let's hope he has has some time to > look at this ;) > > If you can isolate the problem in a small program that reproduces this > behavior, please do upload it in the issue. I tried, but failed :( > This will help the gwt dev team a lot to solve the problem (and the issue > will get a higher priority and solved sooner). > > -- > You received this message because you are subscribed to the Google Groups > "Google Web Toolkit" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to google-web-toolkit+unsubscr...@googlegroups.com. > To post to this group, send email to google-web-toolkit@googlegroups.com. > Visit this group at > http://groups.google.com/group/google-web-toolkit?hl=en. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
> > Ed, did you ever get this sorted? As I commented on the linked issue I'm > running into what seems to be a similar problem. No. It's hard to track the exact cause. It tried, but had to give up (a lot of time, and not knowing exactly what happens. But it's good to see that more people are experiencing the same kind of problems such that I am not alone ;) Ray seem to be the code splitting guru. Let's hope he has has some time to look at this ;) If you can isolate the problem in a small program that reproduces this behavior, please do upload it in the issue. I tried, but failed :( This will help the gwt dev team a lot to solve the problem (and the issue will get a higher priority and solved sooner). -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
> > Ed, did you ever get this sorted? As I commented on the linked issue I'm > running into what seems to be a similar problem. Not. It's hard to track the exact cause. It tired, but had to give up (a lot of time, and not knowing exactly what happens. But it's good to see that more people are experiencing the same kind of problems such that I am not alone ;) Ray same to be the code splitting guru. Let's hope he has has some time to look at this ;) If you can isolate the problem in a small program that reproduces this behavior, please do upload it in the issue. I tried, but failed :( -- You received this message because you are subscribed to the Google Groups "Google Web Toolkit" group. To unsubscribe from this group and stop receiving emails from it, send an email to google-web-toolkit+unsubscr...@googlegroups.com. To post to this group, send email to google-web-toolkit@googlegroups.com. Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
Re: Code splittingmysteries
Ed, did you ever get this sorted? As I commented on the linked issue I'm
running into what seems to be a similar problem.
On Wednesday, April 10, 2013 8:13:07 AM UTC+1, Ed wrote:
>
> I just downgraded to GWT 2.4 and my left over fragment disappeared :(...
> (initial load was almost the same)
> Strange -> or 2.4. had a code splitting involved bug or 2.5. has one...
>
>
>
> On Mon, Mar 25, 2013 at 4:03 PM, Ed Bras
> > wrote:
>
>> BTW: a corresponding issue in the issue tracker:
>> https://code.google.com/p/google-web-toolkit/issues/detail?id=7874
>> Let's hope more people are experiencing the same problem (and vote the
>> issue) such that the dev team put it on their priority list.
>>
>>
>> On Wed, Mar 20, 2013 at 3:22 PM, Ashton Thomas
>>
>> > wrote:
>>
>>> I would also be interested in any updates to this as I am currently
>>> working on the async aspect or my app. (I wouldn't even know how to
>>> identify anomalies due to a lack of experience here)
>>>
>>> I also remembrer having the same questions with soyc
>>>
>>> -at
>>>
>>>
>>> On Wednesday, March 20, 2013 7:53:11 AM UTC-4, Ed wrote:
After upgrading from 2.5.0 to 2.5.1 all of a sudden my code splitting
result is different:
+ Code fragments are bundles together which weren't before.
+ I have a left over fragment, which was 0 (see attachment).
What changed concerning code splitting? Can't find any in the release
notes.
A few observations:
1) All my components have a singleton facade that contains a method to
create the package protected controller implementation that is called
through split point (GWT.runAsync()).
In this way I have a LoginControllerDefault that is present in the left
over, but this package protected controller ONLY appears in the
LoginFacade.
In the same way I have SubscribeControllerControllerD**efault that
appears only in SubscribeControllerFacade...
What suprised me is that they appear both in the same code split,
namely #30 (see screenshot attachment).
No idea why.
I moved the creation of LoginControllerDefault from the Facade
"directly" (see below for code snippets) in to the runAsync, but that had
no effect.
I also removed the LoginControllerDefault completely which had little
effect on the left-over
When I read the soyc report (automatically build during maven nightly
build), It doesn't really help me. This page is hard to understand, like
this page:
soycReport/compile-report/**leftoverStatus-0.html
Even worse, from this page you can go deeper to see why a class is
in a certain split point but that results in a page not found (404).
That is, the following page isn't present. BTW: the page also resulted
in a 404 in previous GWT releases.
soycReport/compile-report/**
methodDependencies-total-0.**html
Why?
I played with the nr of "fragmentCount", but that has little effect,
see the attached screenshots:
+ fragmentCount = 6: screenshot with 4 splits points.
+ fragmentCount = 8: screenshot with 6 splits points.
+ fragmentCount = 10: screenshot with 7 splits points.
Below a code snippet that shows how I create the controllers in split
point #30 (see screenshot). It concerns the creation of the
LoginControllerDefault and SubscribeInfoControllerDefault**.
Note: I don't find it the most elegant code solution, but before I
passed in a Creator interface that was used in GWT.runAsync(), but that
didn't work (big left over). I posted that in the forum and it was
mentioned that it could be that it is because the code removal is run
after
the split point code, and that is probably why it doesn't work and that
Ray
still have a look in to code split code that need to be optimized.
How to solve/analyze these issues?
What are your experiences?
Any suggestions are more then welcome as I am out of ideas and find it
hard to solve this, but it frustrates me ;)
Loading and creating the the Login and SubscribeInfo controller:
-
The code snippet that is called by the LoginControllerBuilder to load
the Login controller:
public void create(final AsyncCallback callback) {
getLazyLoader().**createInCodeSplitGeneral30(**LoginController.class,
callback);
}
The controller loading code (in other class):
public void createInCodeSplitGeneral30(**final Class type,
final AsyncCallback callback, final SimpleMediator mediator) {
GWT.runAsync(new RunAsyncCallback() {
public void onSuccess() {
callback.onSuccess(**createCodeSplitGeneral30(type, mediator));
}
public void onFailure(final Throwable reason) {
callback.onFailure(reason);
}
});
}
>>>
Re: Code splittingmysteries
I just downgraded to GWT 2.4 and my left over fragment disappeared :(...
(initial load was almost the same)
Strange -> or 2.4. had a code splitting involved bug or 2.5. has one...
On Mon, Mar 25, 2013 at 4:03 PM, Ed Bras wrote:
> BTW: a corresponding issue in the issue tracker:
> https://code.google.com/p/google-web-toolkit/issues/detail?id=7874
> Let's hope more people are experiencing the same problem (and vote the
> issue) such that the dev team put it on their priority list.
>
>
> On Wed, Mar 20, 2013 at 3:22 PM, Ashton Thomas wrote:
>
>> I would also be interested in any updates to this as I am currently
>> working on the async aspect or my app. (I wouldn't even know how to
>> identify anomalies due to a lack of experience here)
>>
>> I also remembrer having the same questions with soyc
>>
>> -at
>>
>>
>> On Wednesday, March 20, 2013 7:53:11 AM UTC-4, Ed wrote:
>>>
>>> After upgrading from 2.5.0 to 2.5.1 all of a sudden my code splitting
>>> result is different:
>>> + Code fragments are bundles together which weren't before.
>>> + I have a left over fragment, which was 0 (see attachment).
>>>
>>> What changed concerning code splitting? Can't find any in the release
>>> notes.
>>>
>>> A few observations:
>>> 1) All my components have a singleton facade that contains a method to
>>> create the package protected controller implementation that is called
>>> through split point (GWT.runAsync()).
>>> In this way I have a LoginControllerDefault that is present in the left
>>> over, but this package protected controller ONLY appears in the LoginFacade.
>>> In the same way I have SubscribeControllerControllerD**efault that
>>> appears only in SubscribeControllerFacade...
>>> What suprised me is that they appear both in the same code split, namely
>>> #30 (see screenshot attachment).
>>>
>>> No idea why.
>>> I moved the creation of LoginControllerDefault from the Facade
>>> "directly" (see below for code snippets) in to the runAsync, but that had
>>> no effect.
>>> I also removed the LoginControllerDefault completely which had little
>>> effect on the left-over
>>>
>>> When I read the soyc report (automatically build during maven nightly
>>> build), It doesn't really help me. This page is hard to understand, like
>>> this page:
>>>soycReport/compile-report/**leftoverStatus-0.html
>>>
>>> Even worse, from this page you can go deeper to see why a class is in
>>> a certain split point but that results in a page not found (404).
>>> That is, the following page isn't present. BTW: the page also resulted
>>> in a 404 in previous GWT releases.
>>> soycReport/compile-report/**methodDependencies-total-0.
>>> **html
>>> Why?
>>>
>>> I played with the nr of "fragmentCount", but that has little effect, see
>>> the attached screenshots:
>>> + fragmentCount = 6: screenshot with 4 splits points.
>>> + fragmentCount = 8: screenshot with 6 splits points.
>>> + fragmentCount = 10: screenshot with 7 splits points.
>>>
>>> Below a code snippet that shows how I create the controllers in split
>>> point #30 (see screenshot). It concerns the creation of the
>>> LoginControllerDefault and SubscribeInfoControllerDefault**.
>>> Note: I don't find it the most elegant code solution, but before I
>>> passed in a Creator interface that was used in GWT.runAsync(), but that
>>> didn't work (big left over). I posted that in the forum and it was
>>> mentioned that it could be that it is because the code removal is run after
>>> the split point code, and that is probably why it doesn't work and that Ray
>>> still have a look in to code split code that need to be optimized.
>>>
>>> How to solve/analyze these issues?
>>> What are your experiences?
>>> Any suggestions are more then welcome as I am out of ideas and find it
>>> hard to solve this, but it frustrates me ;)
>>>
>>>
>>> Loading and creating the the Login and SubscribeInfo controller:
>>> -
>>> The code snippet that is called by the LoginControllerBuilder to load
>>> the Login controller:
>>> public void create(final AsyncCallback callback) {
>>> getLazyLoader().**createInCodeSplitGeneral30(**LoginController.class,
>>> callback);
>>> }
>>>
>>>
>>> The controller loading code (in other class):
>>> public void createInCodeSplitGeneral30(**final Class type, final
>>> AsyncCallback callback, final SimpleMediator mediator) {
>>> GWT.runAsync(new RunAsyncCallback() {
>>> public void onSuccess() {
>>> callback.onSuccess(**createCodeSplitGeneral30(type, mediator));
>>> }
>>>
>>> public void onFailure(final Throwable reason) {
>>> callback.onFailure(reason);
>>> }
>>> });
>>> }
>>>
>>> @SuppressWarnings("unchecked")
>>> private T createCodeSplitGeneral30(final Class type, final
>>> SimpleMediator mediator) {
>>> if (isType(type, LoginController.class)) {
>>> return (T) LoginFacade.getInstance().**createController();
>>> }
>>> else if (isType(type, SubscribeController.class)) {
>>> return (T) SubscribeFacade.getInstance().**createControl
Re: Code splittingmysteries
BTW: a corresponding issue in the issue tracker:
https://code.google.com/p/google-web-toolkit/issues/detail?id=7874
Let's hope more people are experiencing the same problem (and vote the
issue) such that the dev team put it on their priority list.
On Wed, Mar 20, 2013 at 3:22 PM, Ashton Thomas wrote:
> I would also be interested in any updates to this as I am currently
> working on the async aspect or my app. (I wouldn't even know how to
> identify anomalies due to a lack of experience here)
>
> I also remembrer having the same questions with soyc
>
> -at
>
>
> On Wednesday, March 20, 2013 7:53:11 AM UTC-4, Ed wrote:
>>
>> After upgrading from 2.5.0 to 2.5.1 all of a sudden my code splitting
>> result is different:
>> + Code fragments are bundles together which weren't before.
>> + I have a left over fragment, which was 0 (see attachment).
>>
>> What changed concerning code splitting? Can't find any in the release
>> notes.
>>
>> A few observations:
>> 1) All my components have a singleton facade that contains a method to
>> create the package protected controller implementation that is called
>> through split point (GWT.runAsync()).
>> In this way I have a LoginControllerDefault that is present in the left
>> over, but this package protected controller ONLY appears in the LoginFacade.
>> In the same way I have SubscribeControllerControllerD**efault that
>> appears only in SubscribeControllerFacade...
>> What suprised me is that they appear both in the same code split, namely
>> #30 (see screenshot attachment).
>>
>> No idea why.
>> I moved the creation of LoginControllerDefault from the Facade
>> "directly" (see below for code snippets) in to the runAsync, but that had
>> no effect.
>> I also removed the LoginControllerDefault completely which had little
>> effect on the left-over
>>
>> When I read the soyc report (automatically build during maven nightly
>> build), It doesn't really help me. This page is hard to understand, like
>> this page:
>>soycReport/compile-report/**leftoverStatus-0.html
>>
>> Even worse, from this page you can go deeper to see why a class is in
>> a certain split point but that results in a page not found (404).
>> That is, the following page isn't present. BTW: the page also resulted in
>> a 404 in previous GWT releases.
>> soycReport/compile-report/**methodDependencies-total-0.*
>> *html
>> Why?
>>
>> I played with the nr of "fragmentCount", but that has little effect, see
>> the attached screenshots:
>> + fragmentCount = 6: screenshot with 4 splits points.
>> + fragmentCount = 8: screenshot with 6 splits points.
>> + fragmentCount = 10: screenshot with 7 splits points.
>>
>> Below a code snippet that shows how I create the controllers in split
>> point #30 (see screenshot). It concerns the creation of the
>> LoginControllerDefault and SubscribeInfoControllerDefault**.
>> Note: I don't find it the most elegant code solution, but before I passed
>> in a Creator interface that was used in GWT.runAsync(), but that didn't
>> work (big left over). I posted that in the forum and it was mentioned that
>> it could be that it is because the code removal is run after the split
>> point code, and that is probably why it doesn't work and that Ray still
>> have a look in to code split code that need to be optimized.
>>
>> How to solve/analyze these issues?
>> What are your experiences?
>> Any suggestions are more then welcome as I am out of ideas and find it
>> hard to solve this, but it frustrates me ;)
>>
>>
>> Loading and creating the the Login and SubscribeInfo controller:
>> -
>> The code snippet that is called by the LoginControllerBuilder to load the
>> Login controller:
>> public void create(final AsyncCallback callback) {
>> getLazyLoader().**createInCodeSplitGeneral30(**LoginController.class,
>> callback);
>> }
>>
>>
>> The controller loading code (in other class):
>> public void createInCodeSplitGeneral30(**final Class type, final
>> AsyncCallback callback, final SimpleMediator mediator) {
>> GWT.runAsync(new RunAsyncCallback() {
>> public void onSuccess() {
>> callback.onSuccess(**createCodeSplitGeneral30(type, mediator));
>> }
>>
>> public void onFailure(final Throwable reason) {
>> callback.onFailure(reason);
>> }
>> });
>> }
>>
>> @SuppressWarnings("unchecked")
>> private T createCodeSplitGeneral30(final Class type, final
>> SimpleMediator mediator) {
>> if (isType(type, LoginController.class)) {
>> return (T) LoginFacade.getInstance().**createController();
>> }
>> else if (isType(type, SubscribeController.class)) {
>> return (T) SubscribeFacade.getInstance().**createController();
>> }
>> else {
>> throw GenExpClass.**createExceptionVarMsg(**getClass(), ERR_ID + "Gn30",
>> type);
>> }
>> }
>>
>>
>> --
> You received this message because you are subscribed to a topic in the
> Google Groups "Google Web Toolkit" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/google-web-toolkit/_8CdVbBb488/unsubsc
Re: Code splittingmysteries
I would also be interested in any updates to this as I am currently working
on the async aspect or my app. (I wouldn't even know how to
identify anomalies due to a lack of experience here)
I also remembrer having the same questions with soyc
-at
On Wednesday, March 20, 2013 7:53:11 AM UTC-4, Ed wrote:
>
> After upgrading from 2.5.0 to 2.5.1 all of a sudden my code splitting
> result is different:
> + Code fragments are bundles together which weren't before.
> + I have a left over fragment, which was 0 (see attachment).
>
> What changed concerning code splitting? Can't find any in the release
> notes.
>
> A few observations:
> 1) All my components have a singleton facade that contains a method to
> create the package protected controller implementation that is called
> through split point (GWT.runAsync()).
> In this way I have a LoginControllerDefault that is present in the left
> over, but this package protected controller ONLY appears in the LoginFacade.
> In the same way I have SubscribeControllerControllerDefault that appears
> only in SubscribeControllerFacade...
> What suprised me is that they appear both in the same code split, namely
> #30 (see screenshot attachment).
>
> No idea why.
> I moved the creation of LoginControllerDefault from the Facade
> "directly" (see below for code snippets) in to the runAsync, but that had
> no effect.
> I also removed the LoginControllerDefault completely which had little
> effect on the left-over
>
> When I read the soyc report (automatically build during maven nightly
> build), It doesn't really help me. This page is hard to understand, like
> this page:
>soycReport/compile-report/leftoverStatus-0.html
>
> Even worse, from this page you can go deeper to see why a class is in a
> certain split point but that results in a page not found (404).
> That is, the following page isn't present. BTW: the page also resulted in
> a 404 in previous GWT releases.
> soycReport/compile-report/methodDependencies-total-0.html
> Why?
>
> I played with the nr of "fragmentCount", but that has little effect, see
> the attached screenshots:
> + fragmentCount = 6: screenshot with 4 splits points.
> + fragmentCount = 8: screenshot with 6 splits points.
> + fragmentCount = 10: screenshot with 7 splits points.
>
> Below a code snippet that shows how I create the controllers in split
> point #30 (see screenshot). It concerns the creation of the
> LoginControllerDefault and SubscribeInfoControllerDefault.
> Note: I don't find it the most elegant code solution, but before I passed
> in a Creator interface that was used in GWT.runAsync(), but that didn't
> work (big left over). I posted that in the forum and it was mentioned that
> it could be that it is because the code removal is run after the split
> point code, and that is probably why it doesn't work and that Ray still
> have a look in to code split code that need to be optimized.
>
> How to solve/analyze these issues?
> What are your experiences?
> Any suggestions are more then welcome as I am out of ideas and find it
> hard to solve this, but it frustrates me ;)
>
>
> Loading and creating the the Login and SubscribeInfo controller:
> -
> The code snippet that is called by the LoginControllerBuilder to load the
> Login controller:
> public void create(final AsyncCallback callback) {
> getLazyLoader().createInCodeSplitGeneral30(LoginController.class,
> callback);
> }
>
>
> The controller loading code (in other class):
> public void createInCodeSplitGeneral30(final Class type, final
> AsyncCallback callback, final SimpleMediator mediator) {
> GWT.runAsync(new RunAsyncCallback() {
> public void onSuccess() {
> callback.onSuccess(createCodeSplitGeneral30(type, mediator));
> }
>
> public void onFailure(final Throwable reason) {
> callback.onFailure(reason);
> }
> });
> }
>
> @SuppressWarnings("unchecked")
> private T createCodeSplitGeneral30(final Class type, final
> SimpleMediator mediator) {
> if (isType(type, LoginController.class)) {
> return (T) LoginFacade.getInstance().createController();
> }
> else if (isType(type, SubscribeController.class)) {
> return (T) SubscribeFacade.getInstance().createController();
> }
> else {
> throw GenExpClass.createExceptionVarMsg(getClass(), ERR_ID + "Gn30", type);
> }
> }
>
>
>
--
You received this message because you are subscribed to the Google Groups
"Google Web Toolkit" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to google-web-toolkit+unsubscr...@googlegroups.com.
To post to this group, send email to google-web-toolkit@googlegroups.com.
Visit this group at http://groups.google.com/group/google-web-toolkit?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
