flex-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Paul Hastings <paul.hasti...@gmail.com>
Subject Re: flex compiler drunk again
Date Thu, 16 Jun 2016 07:48:35 GMT
On 6/15/2016 11:37 PM, Alex Harui wrote:
> Why do you say "source"?  Are there no SWCs involved?

sure but its clear its the "core" part of the app--the init is completely 
different, the "about" data, etc. are all from the main part of the app. there's 
no lib SWC involved in that bit. that's compiled straight from the source.

> 1) Compare its timestamp in the filesystem of the good one you grabbed
> during the build against the "final-but-fossil" version. Which timestamp
> is newer?  I would imagine the bad one is newer, which means that one of
> the projects in the workspace is generating it.

yup, 1:22pm a modern version is emitted then at 1:23pm it spits out the 
fossilized version.

> 3) Scan the .project and .actionscriptProperties files in the projects.
> It is possible that one of the other projects has been told to generate
> the same output file.

ding ding ding. in .actionscriptProperties file a mystery branch has popped up 
(on me for not digging deeper in the branches).

<application path="alluvium.mxml"/>
<application path="branches/sammy/alluvium.mxml"/>

not sure why that particular code is in that branch but i suppose this is where 
the compiler is finding its fossils (it looks like the same age as what's being 
output).

if i just delete the reference, will that stop the compiler from compiling it 
(need to figure out how/why that's in my source)?

> 6) You could also try building a custom compiler that generates more
> output to help you track it down.

ha ;-)


thanks for the advice.

Mime
View raw message