Hi!
Today I moved osp module into "modules". Now it does not compile any more because it links against rr and auth module, but there is no rr auth in "modules" yet. Should I just change it to include rr and auth from "modules_k" or should I move it back to "modules_k" directory?
regards Klaus
On Jul 02, 2009 at 15:52, Klaus Darilion klaus.mailinglists@pernau.at wrote:
Hi!
Today I moved osp module into "modules". Now it does not compile any more because it links against rr and auth module, but there is no rr auth in "modules" yet. Should I just change it to include rr and auth from "modules_k" or should I move it back to "modules_k" directory?
It's unlikely that the same version would work with both ser rr & auth and k rr & auth (but I might be wrong). To be on the safe side I think you should move it back (to modules_k and modules_s) and make a note somewhere on the wiki (maybe a new page about module merge effort) that merging failed only because dependencies on rr and auth. You could use git revert <commit_hash> to undo your original commit.
Andrei
On Fri, Jul 3, 2009 at 1:47 PM, Andrei Pelinescu-Onciulandrei@iptel.org wrote:
On Jul 02, 2009 at 15:52, Klaus Darilion klaus.mailinglists@pernau.at wrote:
Hi!
Today I moved osp module into "modules". Now it does not compile any more because it links against rr and auth module, but there is no rr auth in "modules" yet. Should I just change it to include rr and auth from "modules_k" or should I move it back to "modules_k" directory?
It's unlikely that the same version would work with both ser rr & auth and k rr & auth (but I might be wrong). To be on the safe side I think you should move it back (to modules_k and modules_s) and make a note somewhere on the wiki (maybe a new page about module merge effort) that merging failed only because dependencies on rr and auth. You could use git revert <commit_hash> to undo your original commit.
This particular module heavily depends on the tm callbacks and thus we should have only one version and that is the version which is works with the tm module from ser. I don't know which version Klaus took, but I would suggest to take a version which works with our current version of tm and add missing features to it from the other implementation (i.e. the rr and auth stuff which originally linked with headers in modules_k).
As far as I can remember there were only minor differences between the two versions, such as that the ser version did not use the rr api, or something like that. But that should easy to fix.
Jan.
I now have reverted the commit and leave it as it was.
regards klaus
Jan Janak schrieb:
On Fri, Jul 3, 2009 at 1:47 PM, Andrei Pelinescu-Onciulandrei@iptel.org wrote:
On Jul 02, 2009 at 15:52, Klaus Darilion klaus.mailinglists@pernau.at wrote:
Hi!
Today I moved osp module into "modules". Now it does not compile any more because it links against rr and auth module, but there is no rr auth in "modules" yet. Should I just change it to include rr and auth from "modules_k" or should I move it back to "modules_k" directory?
It's unlikely that the same version would work with both ser rr & auth and k rr & auth (but I might be wrong). To be on the safe side I think you should move it back (to modules_k and modules_s) and make a note somewhere on the wiki (maybe a new page about module merge effort) that merging failed only because dependencies on rr and auth. You could use git revert <commit_hash> to undo your original commit.
This particular module heavily depends on the tm callbacks and thus we should have only one version and that is the version which is works with the tm module from ser. I don't know which version Klaus took, but I would suggest to take a version which works with our current version of tm and add missing features to it from the other implementation (i.e. the rr and auth stuff which originally linked with headers in modules_k).
As far as I can remember there were only minor differences between the two versions, such as that the ser version did not use the rr api, or something like that. But that should easy to fix.
Jan.
That's a pity, I was hoping that you kick our ... and solve it before we are done discussing it with Andrei :-).
Jan.
On Fri, Jul 3, 2009 at 3:08 PM, Klaus Darilionklaus.mailinglists@pernau.at wrote:
I now have reverted the commit and leave it as it was.
regards klaus
Jan Janak schrieb:
On Fri, Jul 3, 2009 at 1:47 PM, Andrei Pelinescu-Onciulandrei@iptel.org wrote:
On Jul 02, 2009 at 15:52, Klaus Darilion klaus.mailinglists@pernau.at wrote:
Hi!
Today I moved osp module into "modules". Now it does not compile any more because it links against rr and auth module, but there is no rr auth in "modules" yet. Should I just change it to include rr and auth from "modules_k" or should I move it back to "modules_k" directory?
It's unlikely that the same version would work with both ser rr & auth and k rr & auth (but I might be wrong). To be on the safe side I think you should move it back (to modules_k and modules_s) and make a note somewhere on the wiki (maybe a new page about module merge effort) that merging failed only because dependencies on rr and auth. You could use git revert <commit_hash> to undo your original commit.
This particular module heavily depends on the tm callbacks and thus we should have only one version and that is the version which is works with the tm module from ser. I don't know which version Klaus took, but I would suggest to take a version which works with our current version of tm and add missing features to it from the other implementation (i.e. the rr and auth stuff which originally linked with headers in modules_k).
As far as I can remember there were only minor differences between the two versions, such as that the ser version did not use the rr api, or something like that. But that should easy to fix.
Jan.
Andrei Pelinescu-Onciul schrieb:
You could use git revert <commit_hash> to undo your original commit.
Hi Andrei!
This leads me to a git question. If I perform the revert, the git log tells me:
rename {modules => modules_k}/osp/Makefile (100%) rename {modules => modules_k}/osp/README (100%) copy {modules => modules_k}/osp/RELEASE-NOTES.txt (100%) copy {modules => modules_k}/osp/destination.c (100%) copy {modules => modules_k}/osp/destination.h (100%) copy {modules => modules_k}/osp/doc/Makefile (100%) copy {modules => modules_k}/osp/doc/osp.xml (100%) copy {modules => modules_k}/osp/doc/osp_admin.xml (100%) rename {modules => modules_k}/osp/doc/osp_devel.xml (100%) copy {modules => modules_k}/osp/etc/cacert_0.pem (100%) copy {modules => modules_k}/osp/etc/localcert.pem (100%) ... copy {modules => modules_k}/osp/usage.c (100%) copy {modules => modules_k}/osp/usage.h (100%) create mode 100644 modules_s/osp/Makefile create mode 100644 modules_s/osp/README rename {modules => modules_s}/osp/RELEASE-NOTES.txt (75%) rename {modules => modules_s}/osp/destination.c (70%)
e.g. the file modules/osp/RELEASE-NOTES.txt is first copied to modules_k/osp/RELEASE-NOTES.txt and the renamed to modules_s/osp/RELEASE-NOTES.txt.
That is of course wrong, but if I verify the result, it seems to be ok.
Is this just a strange log-message from git or is there something I have to be aware?
thanks Klaus
-
Andrei Pelinescu-Onciul -
Jan Janak -
Klaus Darilion