Fred, Which classes in flex-falcon are interacting with FDB? I’m trying to understand whether we can bundle FDB from the Flex SDK into FlexJS.
Git log for dcc404689eb20d5a037bbf67f4cbb85adec92237 shows that you changed FaultActions for FLEX-34062. I am able to copy 3 jars from the Flex SDK and run fdb, but others were getting some sort of mismatch, like some older version of DebugCLI.java was calling directly to FaultActions like it used to. There seemed to be a correlation between the 4.0 IntelliJ jar resulting in the FaultActions exception and the newer IntelliJ jar not having that issue. Anyway, if it is now working for Kevin, then we don’t have to spend more time on it. Regarding the file location in FDB, I wasn’t sure you had time to work on it so I didn’t mention it. But if you have time that would be great. IMO, it doesn’t require a JIRA issue to track it if you have the fix ready. Thanks, -Alex On 5/15/15, 8:03 AM, "Frédéric THOMAS" <webdoubl...@hotmail.com> wrote: >Alex, > >Btw, Just seen my commit message before to fix the worker but after my >merge of the donated code, FLEX-34291: > >I kept the legacy code to evaluate expressions, the new code based on >Falcon classes will be created in the Falcon repository. >I re-applied previous fixes we had of FDB > >So, it seems there were some falcon dependencies. > >Frédéric THOMAS > >> From: webdoubl...@hotmail.com >> To: dev@flex.apache.org >> Subject: RE: [FlexJS] IntelliJ Integration >> Date: Fri, 15 May 2015 15:53:01 +0100 >> >> >> >> >> > I didn’t think there was a different debugger code base for Falcon. >> >> That's a while, but IIRC and I can be wrong, it seems to me I had to >>remap some classes onto the asc or compiler, can't exactly remember, I >>should look at the git history and compare to be sure. >> >> > I think Piotr and Kevin reported an exception that indicates that >>some of >> > the FDB classes may be monkey-patched in some of the IntelliJ jars and >> > when you combine those with the most recent FDB classes from Apache >>Flex >> > things did not work because as part of the workers change, the APIs >>for >> > FaultAction changed. >> >> AFAIK, it is not the reason why, othewise, how could it work with the >>same FDB of the from the Flex SDK ? >> IMO, he was missing some dependencies that have been filled overlaying >>the SDK. >> >> After a quick look at FaultAction, this class has been reworked by >>Adobe because I integrated it as is and didn't touch it. >> >> > IMO, we would want FlexJS releases to bundle FDB from a recent Flex >>SDK, >> > and getting IntelliJ to work with that configuration would be >>“better”. >> >> Sure even though it could make my life easier if in a eventual IntelliJ >>plugin, I add a FlexJS facet to a FlexSDK than if I create a full >>plugin, even if this one is based on Randori, it still remains a huge >>task. >> >> >> > When you do an overlay, does code hinting also show Flex SDK classes >>as >> > well as the FlexJS classes? >> >> Yes, if you import a mx or spark button, you will see also its >>definition, code hinting, etc... >> >> > Yes, I saw that, and thanks for doing it. One other thing that I’ve >> > noticed in FDB since the workers change is that I can’t just do “list >>49” >> > or “break 49” and have FDB apply that to the “current file”. Since >>the >> > workers change I have to do “list myfile.as:49” or “break >>myfile.as:49”. >> > The “current file” used to be the last file being listed or the file >>we >> > stopped at for a breakpoint. Does the workers feature prevent having >>a >> > “current file” concept? >> >> Hey, why didn't you raise a bug on that before ?? :-) >> >> I had a look at it and yes, in this case I forgot to look for the file >>location. >> >> What do you want to do, raise a bug for the record or you prefer I >>commit the fix directly ? >> >> Thanks, >> Frédéric THOMAS >> >> > From: aha...@adobe.com >> > To: dev@flex.apache.org >> > Subject: Re: [FlexJS] IntelliJ Integration >> > Date: Fri, 15 May 2015 13:04:01 +0000 >> > >> > >> > >> > On 5/15/15, 5:31 AM, "Frédéric THOMAS" <webdoubl...@hotmail.com> >>wrote: >> > >> > >Hi Alex, >> > > >> > >It seems I've been wrong here, I didn't get well the context and I >>was >> > >missing data, I thougt FDB in Falcon was the donated version, it >>seems to >> > >be wrong and FDB is only a copy of the one in the Flex SDK, right ? >> > > >> > >If I'm correct, why don't we use the donated one which was based on >> > >Falcon IIRC ? >> > >> > I didn’t think there was a different debugger code base for Falcon. >> > >> > > >> > >The FDB bug you mentioned was when we haven't yet overlaid the Flex >>SDK >> > >on top of the FlexJS one, or the opposite way, correct ? >> > >If so, doesn't it mean, there were some dependencies missing instead >>? >> > > >> > >When FlexJS is used as a lib on top of a old SDK, in IntelliJ, I can >>set >> > >the debugger I want to use, I tried with the one in FlexJS and the >>one in >> > >my old 4.10, they both work well. >> > >> > I think Piotr and Kevin reported an exception that indicates that >>some of >> > the FDB classes may be monkey-patched in some of the IntelliJ jars and >> > when you combine those with the most recent FDB classes from Apache >>Flex >> > things did not work because as part of the workers change, the APIs >>for >> > FaultAction changed. >> > >> > IMO, we would want FlexJS releases to bundle FDB from a recent Flex >>SDK, >> > and getting IntelliJ to work with that configuration would be >>“better”. >> > When you do an overlay, does code hinting also show Flex SDK classes >>as >> > well as the FlexJS classes? >> > >> > > >> > >Anyway, I fixed the remaining open bug in FDB relative to worker last >> > >night. >> > >> > Yes, I saw that, and thanks for doing it. One other thing that I’ve >> > noticed in FDB since the workers change is that I can’t just do “list >>49” >> > or “break 49” and have FDB apply that to the “current file”. Since >>the >> > workers change I have to do “list myfile.as:49” or “break >>myfile.as:49”. >> > The “current file” used to be the last file being listed or the file >>we >> > stopped at for a breakpoint. Does the workers feature prevent having >>a >> > “current file” concept? >> > >> > Thanks, >> > -Alex >> > >> >> >
