compilesWithoutError failure is hard to understand without the generated sources. about compile-testing HOT 7 CLOSED

I have a commit nearly ready for that. I'll try to get it out later today.

from compile-testing.

I overdid the current version with attempts to do poor-man's diffs, which worked, but were ugly as hell. I'm just going to return the code that was generated, if we can find it (relative to the expected's filename), else a list of generated files.

from compile-testing.

sounds good to me assuming that the 'list of generated files' will also
contain the file contents.

Thanks!

On Tue, Nov 26, 2013 at 10:37 AM, Christian Edward Gruber <
[email protected]> wrote:

I overdid the current version with attempts to do poor-man's diffs, which
worked, but were ugly as hell. I'm just going to return the code that was
generated, if we can find it (relative to the expected's filename), else a
list of generated files.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/34#issuecomment-29318867
.

from compile-testing.

No, what I was thinking was, if I can find a matching file name, but the contents can't be found, I show what we generated with that file-name. Otherwise, I list the names of the files we generated. You can either decide that you meant those names anwyay, and fix your JavaFileObject's name to suit, or you can see that the file you expected to be generated wasn't generated into that filename and fix. If we do find a file, we show you the contents.

If I show every file generated, then it's a crazy huge wall of text, none of which might be relevant, because there wasn't even an appropriately named file generated.

Does that suit, or do you really want the full wall-o-text? Above a few files, and it's nuts.

from compile-testing.

i think we are talking about two different things, are you talking about
#18?

This is about when there is an unexpected compile error in a generated
file. Currently compile_tester will print the compile errors, but since i
don't have access to the generated source it is difficult to understand.

On Tue, Nov 26, 2013 at 1:02 PM, Christian Edward Gruber <
[email protected]> wrote:

No, what I was thinking was, if I can find a matching file name, but the
contents can't be found, I show what we generated with that file-name.
Otherwise, I list the names of the files we generated. You can either
decide that you meant those names anwyay, and fix your JavaFileObject's
name to suit, or you can see that the file you expected to be generated
wasn't generated into that filename and fix. If we do find a file, we show
you the contents.

If I show every file generated, then it's a crazy huge wall of text, none
of which might be relevant, because there wasn't even an appropriately
named file generated.

Does that suit, or do you really want the full wall-o-text? Above a few
files, and it's nuts.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/34#issuecomment-29333559
.

from compile-testing.

Crap - you're right. I am thinking the ".generatesSources(...)" path. Sorry. My bad. I agree, if there's an error in the file, output of the file is hella useful. I'll whip up something.

from compile-testing.

I believe this is stale. Closing for now, and we can reopen if it's not fully implemented.

from compile-testing.