Closed GoogleCodeExporter closed 9 years ago
I'll note that this also happens when trying: Leiningen -> Reset project
configuration
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 4:28
Can you upload your workspace .metadata/.log file ?
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 4:31
Thank you for the quick response, Laurent!
In skimming the log I'm not sure which is the root problem, but paizo is my
hostname.
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 4:43
Attachments:
If possible, if the issue is reproducible, then maybe you can close your
eclipse, then delete the .log file, start the eclipse again and do as little as
possible to reproduce the issue. This way the log may be easier to check.
Also, have you found a work-around in the mean time ?
Anything special about your setup? http proxies, things like that?
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 4:51
Yes, it is very reproducible. It happens every time now.
This log was made by first deleting the old one
Opening eclipse
Leiningen -> Reset project dependencies
It appears to be pretty much the same log.
No, I haven't found a workaround, which is driving me a little nuts. I'm
checking out other IDEs so I don't lose the day of coding, but I'm very much a
CCW user.
Nothing special about my configuration, and no proxies.
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 4:59
Attachments:
Yes, I have checked the .log, and there's this NullPointerException which is
the root of the problem, and it's happening in a piece of code that is probably
undertested, related to path exclusions in source folders.
Still trying to figure out what could cause that.
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 5:01
Can you open your eclipse installation folder, then the plugins/ folder, then
the ccw.core folder (will have a prefix with the version number), then find and
replace:
- nature.clj in ccw/leiningen/
- jdt.clj in ccw/
with the attached two files of the same name. I have added additional printlns
to get more info on what the culprit fns take as input.
Then if you can post a fresh .log again, that would be nice.
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 5:09
the mentioned files in the above comment
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 5:11
Attachments:
Upps, files are missing.
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 5:13
they're in the following comment :-) so the one 2 comments above this one :-)
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 5:17
I have some plans to someday have part of ccw that is delivered as a git repo,
with an easy way to branch and create test setups for users ! In the mean time,
sorry to have you do this file danse.
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 5:18
nature.clj wasn't working, so i tried moving the println of path into the map
call for source-entries, but it doesnt seem to be printing in this .log
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 5:21
Attachments:
argh, maybe it is outputting it directly to your Eclipse stdout and it's not
redirected to the .log file :-(
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 5:22
Ok, running eclipse from terminal:
path: #<Folder F/dasf/src> , source-folders: (#<Folder F/dasf/src> #<Folder
F/dasf/test> #<Folder F/dasf/dev-resources> #<Folder F/dasf/resources> nil nil)
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 5:25
It didn't look like it was making it to the source-entry println, so I added a
(filter identity) to the threading and this comes out:
path: #<Folder F/dasf/src> , source-folders: (#<Folder F/dasf/src> #<Folder
F/dasf/test> #<Folder F/dasf/dev-resources> #<Folder F/dasf/resources> nil nil)
source-entry: {:path #<Folder F/dasf/src>, :exclusion-patterns (),
:extra-attributes {optional true}}
path: #<Folder F/dasf/test> , source-folders: (#<Folder F/dasf/src> #<Folder
F/dasf/test> #<Folder F/dasf/dev-resources> #<Folder F/dasf/resources> nil nil)
source-entry: {:path #<Folder F/dasf/test>, :exclusion-patterns (),
:extra-attributes {optional true}}
path: #<Folder F/dasf/dev-resources> , source-folders: (#<Folder F/dasf/src>
#<Folder F/dasf/test> #<Folder F/dasf/dev-resources> #<Folder F/dasf/resources>
nil nil)
source-entry: {:path #<Folder F/dasf/dev-resources>, :exclusion-patterns (),
:extra-attributes {optional true}}
path: #<Folder F/dasf/resources> , source-folders: (#<Folder F/dasf/src>
#<Folder F/dasf/test> #<Folder F/dasf/dev-resources> #<Folder F/dasf/resources>
nil nil)
source-entry: {:path #<Folder F/dasf/resources>, :exclusion-patterns (),
:extra-attributes {optional true}}
path: nil , source-folders: (#<Folder F/dasf/src> #<Folder F/dasf/test>
#<Folder F/dasf/dev-resources> #<Folder F/dasf/resources> nil nil)
If that is of any use
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 5:32
Ooo... when I do
source-folders (filter identity (concat (lein-source-folders lein-proj)))
It works!
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 5:37
hourray, you've fixed the bug ! :-)
Will integrate it into the master branch, thanks !
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 5:39
Seems like there are 2 source folders initialized through the leiningen
project.clj map, but that are not present in the project path, thus the 2 nils.
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 5:48
which is odd because it is a brand new project, but I'm up and coding so I'm
not complaining!
Huzzah! Thank you Laurent!
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 5:54
can you just try to print the 2 folders that provoke the 2 nils?
it's just a matter of adding a println trace in the nature.clj file, in the
lein-source-folders function, near line 75.
You add:
(println "lein-raw-source-folders:" (lein-raw-source-folders lein-proj))
and we'll get the list of filesystem paths returned by leiningen, and may
understand where those 2 are coming from.
Maybe something in your ~/.lein/settings.clj or profile.clj
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 6:03
Pardon the switch of project, but that shouldn't have an issue on the error (i
double checked by removing the filter identity fix
lein-raw-source-folders: (/Users/kyle/Documents/brevisdemo_workspace/brevis/src
/Users/kyle/Documents/brevisdemo_workspace/brevis/java
/Users/kyle/Documents/brevisdemo_workspace/brevis/test
/Users/kyle/Documents/brevisdemo_workspace/brevis/dev-resources
/Users/kyle/Documents/brevisdemo_workspace/brevis/resources
/Users/kyle/Documents/brevisdemo_workspace/brevis/target/native
/Library/Java/JavaVirtualMachines/jdk1.7.0_55.jdk/Contents/Home/lib/tools.jar
/Library/Java/JavaVirtualMachines/jdk1.7.0_55.jdk/Contents/Home/lib/sa-jdi.jar)
I've checked and the last 2 jars do exist.
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 6:16
Ahhhh ! There are 2 jars in the end, and they are not located inside the
project's location, thus causing the error.
I don't really understand why they're here. They're not really source folder,
are they?
This looks a little bit like some hack introduce to circumvent leiningen's
strict rules of not having the ability to specify local jars.
What should we do? Drop those 2 jar entries as is done via the filter identity,
or try to get them into the project's java build path?
I'm in favor of dropping them, but I prefer asking you anyway
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 6:32
I think would agree with dropping them. It seems likely that lein-javadoc (v
0.1.1) is putting them there (the other active plugins are midje and localrepo).
Original comment by kephal...@gmail.com
on 28 Nov 2014 at 6:44
Indeed. Consider the fix included with the next release.
I fixed it differently, though: I replaced map with keep in the definition of
lein-source-folders - just to let you know.
Cheers
Original comment by laurent....@gmail.com
on 28 Nov 2014 at 6:48
Original comment by laurent....@gmail.com
on 4 Dec 2014 at 12:06
Original issue reported on code.google.com by
kephal...@gmail.com
on 28 Nov 2014 at 4:22