John Dias:
Ugh. Sorry; looks like I forgot to record my merge patches. Looks
like you got most of them; I'll send the last one as soon as my
validate is finished.
To avoid this type of problem (and some others), it is worthwhile to
have a separate validation tree of ghc.
The procedure then is this:
* When you pull, pull into your test tree (let's call it ghc-test).
* From your test tree, pull into your working copy (let's call that
ghc).
* Record in ghc and push to ghc-test.
* Validate in ghc-test (and never edit anything) in there.
* When validate succeeds, push from ghc-test.
This ensures that what you validate is what the rest of the world sees
when you push (provided you keep your ghc-test up-to-date). An
additional benefit of this set up is that ghc-test serves as a sort of
quarantine repo for incoming patches. Just pull to ghc-test and run
validate before pulling into ghc. If ghc-test breaks you can still
work in ghc without having to find out which of the patches you just
pulled caused trouble. (As you just demonstrated, the tree does break
and so isolating your work from the main tree is the only way that you
can pull frequently without risking to be stuck on a regular basis.)
Manuel
On Wed, 26 Nov 2008, Simon Peyton-Jones wrote:
ah ok I see what to do. I'll send a patch shortly
S
| -----Original Message-----
| From: Simon Marlow [mailto:[EMAIL PROTECTED] On Behalf
Of Simon Marlow
| Sent: 26 November 2008 09:24
| To: Simon Peyton-Jones
| Cc: John Dias; [email protected]
| Subject: Re: new codegen patches
|
| Simon Peyton-Jones wrote:
| > Oops. where is mkSECAFBlackHoleInfoTableLabel' defined?
|
| Looks like a merge problem - I removed the SE blackhole variants
recently.
|
| Cheers,
| Simon
|
| > Simon
| >
| > /home/simonmar/fp/bin/x86_64-unknown-linux/ghc ... -c cmm/
CmmLive.hs -o dist-
| stage1/build/CmmLive.o -ohi dist-stage1/build/CmmLive.hi
| >
| > codeGen/StgCmmClosure.hs:767:32:
| > Not in scope: `mkSECAFBlackHoleInfoTableLabel'
| > <<ghc: 150025472 bytes, 4 GCs, 178696/183480 avg/max bytes
residency (2 samples), 62M in
| use, 0.00 INIT (0.00 elapsed), 0.15 MUT (0.24 elapsed), 0.04 GC
(0.05 elapsed) :ghc>>
| >
| > | -----Original Message-----
| > | From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
] On Behalf Of John
| Dias
| > | Sent: 26 November 2008 05:32
| > | To: [email protected]
| > | Subject: new codegen patches
| > |
| > | As you may have noticed, I have finally pushed a large number
of patches
| > | for the new code generator. The new codegen is still off by
default, so
| > | for the most part, this patch should not have any effect on
compiled
| > | programs. I validated the patches, but please let me know if
there are any
| > | problems.
| > |
| > | In case you're wondering about the status of the new code
generator: There
| > | are still bugs that show up with the stage2 build, although
the regression
| > | testsuite and nofib testsuite work with the stage1 build.
Hopefully, I'll
| > | have some more patches soon.
| > |
| > | Thanks,
| > | John
| > |
| > | _______________________________________________
| > | Cvs-ghc mailing list
| > | [email protected]
| > | http://www.haskell.org/mailman/listinfo/cvs-ghc
| >
| > _______________________________________________
| > Cvs-ghc mailing list
| > [email protected]
| > http://www.haskell.org/mailman/listinfo/cvs-ghc
|
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc
