On Jun 20, 2007, at 1:48 PM, Philipp von Weitershausen wrote:
Btw, if I interpret http://svn.zope.org/?rev=76390&view=rev correctly, it sounds like Tres has found a solution to that ZClass bug. We just need to propogate the patch to Zope 3.3/2.10 and trunk/ trunk. Right, Tres?
What I fixed was the unique-to-2.9 breakage due to the fact that 'Implements' objects couldn't be pickled. The 2.9 ZClass tests now fail in exactly the same way as the 2.10 / trunk tests.
Ah. What a relief :)
:) apart from that i think jim fixed the zclasses problem on the last day of the sprint, at least he told me so. could be i misunderstood something, but actually i'm quite sure, since he also told me how he did it and that made a lot of sense after having looked at the problem for a couple hours before... :) ah, here it is: http://svn.zope.org/?rev=76408&view=rev should fix the zclasses problem, if i'm not totally mistaken. and btw, andreas' branch[1] produces no test failures for me (at r76878) anyway, so imho we're good to go! :) cheers, andi [1] svn+ssh://svn.zope.org/repos/main/Zope/branches/Zope211-3.4- integration -- zeidler it consulting - http://zitc.de/ - info@zitc.de friedelstraße 31 - 12047 berlin - telefon +49 30 25563779 pgp key at http://zitc.de/pgp - http://wwwkeys.de.pgp.net/ sprint with us! - http://plone.org/events/sprints/potsdam-sprint-2007
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Andreas Zeidler wrote:
On Jun 20, 2007, at 1:48 PM, Philipp von Weitershausen wrote:
Btw, if I interpret http://svn.zope.org/?rev=76390&view=rev correctly, it sounds like Tres has found a solution to that ZClass bug. We just need to propogate the patch to Zope 3.3/2.10 and trunk/ trunk. Right, Tres? What I fixed was the unique-to-2.9 breakage due to the fact that 'Implements' objects couldn't be pickled. The 2.9 ZClass tests now fail in exactly the same way as the 2.10 / trunk tests. Ah. What a relief :)
:)
apart from that i think jim fixed the zclasses problem on the last day of the sprint, at least he told me so. could be i misunderstood something, but actually i'm quite sure, since he also told me how he did it and that made a lot of sense after having looked at the problem for a couple hours before... :)
ah, here it is: http://svn.zope.org/?rev=76408&view=rev should fix the zclasses problem, if i'm not totally mistaken.
That change needs backporting, I think, to the ZODB 3.6 line (for 2.9) and 3.7 line (for 2.10). I'm pretty sure it landed before Jim branched 3.8, so we should be golden for 2.11.
and btw, andreas' branch[1] produces no test failures for me (at r76878) anyway, so imho we're good to go! :)
Is that running only the ZClass tests? Because none of those were failing when running all tests. Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGesxr+gerLs4ltQ4RAoqLAJ430uLiOm0JcP8LaH4kTIUBsxMdZQCeN+wV YYKA1BOS6Bcni7ZWM57X5Is= =efOv -----END PGP SIGNATURE-----
On Jun 21, 2007, at 9:07 PM, Tres Seaver wrote:
ah, here it is: http://svn.zope.org/?rev=76408&view=rev should fix the zclasses problem, if i'm not totally mistaken.
That change needs backporting, I think, to the ZODB 3.6 line (for 2.9) and 3.7 line (for 2.10).
true, but i think philipp was only asking about 2.11 for the time being. i agree the fix should be backported, though. i'd even volunteer to do it, if nobody else wants to... :)
and btw, andreas' branch[1] produces no test failures for me (at r76878) anyway, so imho we're good to go! :)
Is that running only the ZClass tests? Because none of those were failing when running all tests.
i know, but that's only true for the regular 2.9 and 2.10 branches; on both andreas' (for trunk) and my (for 2.10) "zodb 3.8 integration" branches the zclasses tests would fail both when run separately as well as with all other tests. imho that funny effect that they wouldn't fail when running all tests went away with zodb 3.8. afaik it's even the same on theuni's branch for 2.9. but to answer the question: with jim's fix the zclasses tests run, separately and with all others, on the 2.11 and 2.10 branches (i haven't check theuni's). andi -- zeidler it consulting - http://zitc.de/ - info@zitc.de friedelstraße 31 - 12047 berlin - telefon +49 30 25563779 pgp key at http://zitc.de/pgp - http://wwwkeys.de.pgp.net/ sprint with us! - http://plone.org/events/sprints/potsdam-sprint-2007
On 21 Jun 2007, at 23:47 , Andreas Zeidler wrote:
On Jun 21, 2007, at 9:07 PM, Tres Seaver wrote:
ah, here it is: http://svn.zope.org/?rev=76408&view=rev should fix the zclasses problem, if i'm not totally mistaken.
That change needs backporting, I think, to the ZODB 3.6 line (for 2.9) and 3.7 line (for 2.10).
true, but i think philipp was only asking about 2.11 for the time being. i agree the fix should be backported, though. i'd even volunteer to do it, if nobody else wants to... :)
That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10.
On Jun 25, 2007, at 10:05 AM, Philipp von Weitershausen wrote:
true, but i think philipp was only asking about 2.11 for the time being. i agree the fix should be backported, though. i'd even volunteer to do it, if nobody else wants to... :)
That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10.
when are they due? unfortunately i won't be able to do the backporting before thursday, since we've got a project release on wednesday... andi -- zeidler it consulting - http://zitc.de/ - info@zitc.de friedelstraße 31 - 12047 berlin - telefon +49 30 25563779 pgp key at http://zitc.de/pgp - http://wwwkeys.de.pgp.net/ sprint with us! - http://plone.org/events/sprints/potsdam-sprint-2007
--On 25. Juni 2007 10:13:44 +0200 Andreas Zeidler <az@zitc.de> wrote:
On Jun 25, 2007, at 10:05 AM, Philipp von Weitershausen wrote:
true, but i think philipp was only asking about 2.11 for the time being. i agree the fix should be backported, though. i'd even volunteer to do it, if nobody else wants to... :)
That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10.
when are they due? unfortunately i won't be able to do the backporting before thursday, since we've got a project release on wednesday...
I defer the 2.9/2.10 for some days until I know how we should deal with the current state of the Zope 3.2/3.3 branches. Andreas
On Jun 25, 2007, at 12:09 PM, Andreas Jung wrote:
On Jun 25, 2007, at 10:05 AM, Philipp von Weitershausen wrote:
That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10.
when are they due? unfortunately i won't be able to do the backporting before thursday, since we've got a project release on wednesday...
I defer the 2.9/2.10 for some days until I know how we should deal with the current state of the Zope 3.2/3.3 branches.
that sounds like the release aren't likely to happen before friday, which will leave enough time for me to do the backporting after our release... thanks. :) andi -- zeidler it consulting - http://zitc.de/ - info@zitc.de friedelstraße 31 - 12047 berlin - telefon +49 30 25563779 pgp key at http://zitc.de/pgp - http://wwwkeys.de.pgp.net/ sprint with us! - http://plone.org/events/sprints/potsdam-sprint-2007
--On 25. Juni 2007 12:27:48 +0200 Andreas Zeidler <az@zitc.de> wrote:
On Jun 25, 2007, at 12:09 PM, Andreas Jung wrote:
On Jun 25, 2007, at 10:05 AM, Philipp von Weitershausen wrote:
That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10.
when are they due? unfortunately i won't be able to do the backporting before thursday, since we've got a project release on wednesday...
I defer the 2.9/2.10 for some days until I know how we should deal with the current state of the Zope 3.2/3.3 branches.
that sounds like the release aren't likely to happen before friday, which will leave enough time for me to do the backporting after our release... thanks. :)
I think some time early next week (depending on my schedule). Andreas
On 6/25/07, Andreas Jung <lists@zopyx.com> wrote:
I think some time early next week (depending on my schedule).
OK. I also promised to merge #2153, that gives me time to do that (unless somebody protests). http://zope.org/Collectors/Zope/2153 -- Lennart Regebro: Zope and Plone consulting. http://www.colliberty.com/ +33 661 58 14 64
--On 25. Juni 2007 13:00:38 +0200 Lennart Regebro <regebro@gmail.com> wrote:
On 6/25/07, Andreas Jung <lists@zopyx.com> wrote:
I think some time early next week (depending on my schedule).
OK. I also promised to merge #2153, that gives me time to do that (unless somebody protests). http://zope.org/Collectors/Zope/2153
Back to the original subject...any objections to merge this 2.11 integration branch back on the trunk *with* failing ZClasses tests. Andreas
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Andreas Jung wrote:
--On 25. Juni 2007 13:00:38 +0200 Lennart Regebro <regebro@gmail.com> wrote:
On 6/25/07, Andreas Jung <lists@zopyx.com> wrote:
I think some time early next week (depending on my schedule). OK. I also promised to merge #2153, that gives me time to do that (unless somebody protests). http://zope.org/Collectors/Zope/2153
Back to the original subject...any objections to merge this 2.11 integration branch back on the trunk *with* failing ZClasses tests.
Shouldn't the Zope2 trunk be using ZODB 3.8 at this point? And won't we therefore get Jim's fix? Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGf/Oo+gerLs4ltQ4RAuSbAJ4p+Wz9S5QqkDfdwPc4MKrKu1dV3QCfShPT iAGtpbGN/Y1+zc55mk0vilk= =p91d -----END PGP SIGNATURE-----
--On 25. Juni 2007 12:56:08 -0400 Tres Seaver <tseaver@palladion.com> wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Andreas Jung wrote:
--On 25. Juni 2007 13:00:38 +0200 Lennart Regebro <regebro@gmail.com> wrote:
On 6/25/07, Andreas Jung <lists@zopyx.com> wrote:
I think some time early next week (depending on my schedule). OK. I also promised to merge #2153, that gives me time to do that (unless somebody protests). http://zope.org/Collectors/Zope/2153
Back to the original subject...any objections to merge this 2.11 integration branch back on the trunk *with* failing ZClasses tests.
Shouldn't the Zope2 trunk be using ZODB 3.8 at this point? And won't we therefore get Jim's fix?
The 2.11 integration branch also uses ZODB 3.8. So merging this branch would bring the Zope 2 trunk together with Zope 3.4 and ZODB 3.8. Andreas
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Andreas Jung wrote:
--On 25. Juni 2007 12:56:08 -0400 Tres Seaver <tseaver@palladion.com> wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Andreas Jung wrote:
--On 25. Juni 2007 13:00:38 +0200 Lennart Regebro <regebro@gmail.com> wrote:
On 6/25/07, Andreas Jung <lists@zopyx.com> wrote:
I think some time early next week (depending on my schedule). OK. I also promised to merge #2153, that gives me time to do that (unless somebody protests). http://zope.org/Collectors/Zope/2153
Back to the original subject...any objections to merge this 2.11 integration branch back on the trunk *with* failing ZClasses tests. Shouldn't the Zope2 trunk be using ZODB 3.8 at this point? And won't we therefore get Jim's fix?
The 2.11 integration branch also uses ZODB 3.8. So merging this branch would bring the Zope 2 trunk together with Zope 3.4 and ZODB 3.8.
Then are the ZClasses tests failing when run standaone on that branch? E.g.:: $ /path/to/python2.4 test.py -s ZClasses Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGgCoG+gerLs4ltQ4RAnkGAKC0+pU8D/74CUzZQfx4ZcoyYSyO9ACg2F7V jX90087XWXnU6hkBOjyBzgs= =dHsV -----END PGP SIGNATURE-----
On Jun 25, 2007, at 10:48 PM, Tres Seaver wrote:
Andreas Jung wrote:
--On 25. Juni 2007 12:56:08 -0400 Tres Seaver <tseaver@palladion.com> wrote:
Shouldn't the Zope2 trunk be using ZODB 3.8 at this point? And won't we therefore get Jim's fix?
The 2.11 integration branch also uses ZODB 3.8. So merging this branch would bring the Zope 2 trunk together with Zope 3.4 and ZODB 3.8.
Then are the ZClasses tests failing when run standaone on that branch? E.g.::
$ /path/to/python2.4 test.py -s ZClasses
no, they are not. :) andi -- zeidler it consulting - http://zitc.de/ - info@zitc.de friedelstraße 31 - 12047 berlin - telefon +49 30 25563779 pgp key at http://zitc.de/pgp - http://wwwkeys.de.pgp.net/ sprint with us! - http://plone.org/events/sprints/potsdam-sprint-2007
On Jun 25, 2007, at 6:24 PM, Andreas Jung wrote:
Back to the original subject...any objections to merge this 2.11 integration branch back on the trunk *with* failing ZClasses tests.
hmm, maybe you lost me somewhere, but aren't we talking about your svn://svn.zope.org/repos/main/Zope/branches/Zope211-3.4-integration branch? like i said before, the ZClasses test do not fail on that branch, neither alone nor when run with all other tests. well, at least not for me, but maybe i'm missing something here? cheers, andi -- zeidler it consulting - http://zitc.de/ - info@zitc.de friedelstraße 31 - 12047 berlin - telefon +49 30 25563779 pgp key at http://zitc.de/pgp - http://wwwkeys.de.pgp.net/ sprint with us! - http://plone.org/events/sprints/potsdam-sprint-2007
On 25 Jun 2007, at 12:09 , Andreas Jung wrote:
--On 25. Juni 2007 10:13:44 +0200 Andreas Zeidler <az@zitc.de> wrote:
On Jun 25, 2007, at 10:05 AM, Philipp von Weitershausen wrote:
true, but i think philipp was only asking about 2.11 for the time being. i agree the fix should be backported, though. i'd even volunteer to do it, if nobody else wants to... :)
That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10.
when are they due? unfortunately i won't be able to do the backporting before thursday, since we've got a project release on wednesday...
I defer the 2.9/2.10 for some days until I know how we should deal with the current state of the Zope 3.2/3.3 branches.
I realize you're waiting for a "release" of the Zope 3.2 branch to make another Zope 2.9.x release. I meant to respond earlier, apologies for the delay. I have no intentions of making another Zope 3.2 release (unless somebody really wants me to, in which case I'll do it). So for all I care, you can just tag Zope 3.2.3, put that in the external of Zope 2.9 and be done with it. What's the matter with the Zope 3.3 branch then?
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Philipp von Weitershausen wrote:
On 25 Jun 2007, at 12:09 , Andreas Jung wrote:
--On 25. Juni 2007 10:13:44 +0200 Andreas Zeidler <az@zitc.de> wrote:
On Jun 25, 2007, at 10:05 AM, Philipp von Weitershausen wrote:
true, but i think philipp was only asking about 2.11 for the time being. i agree the fix should be backported, though. i'd even volunteer to do it, if nobody else wants to... :) That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10. when are they due? unfortunately i won't be able to do the backporting before thursday, since we've got a project release on wednesday... I defer the 2.9/2.10 for some days until I know how we should deal with the current state of the Zope 3.2/3.3 branches.
I realize you're waiting for a "release" of the Zope 3.2 branch to make another Zope 2.9.x release. I meant to respond earlier, apologies for the delay. I have no intentions of making another Zope 3.2 release (unless somebody really wants me to, in which case I'll do it). So for all I care, you can just tag Zope 3.2.3, put that in the external of Zope 2.9 and be done with it.
Please do the rest of the release managment before making that tag: - Update the changelog to indicate the release date, SVN tag, etc. (on the branch before copying!) - Update any version ID in the configure script. Then make the tag.
What's the matter with the Zope 3.3 branch then?
Don't we need a released 3.3.x? Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGgQzy+gerLs4ltQ4RAtOxAKCW0G62p+AKqM+5ixUvPJABaAaqVgCghg+Y 0U5Ns+Sx9MpFxlpsvsTc5Ng= =+O+G -----END PGP SIGNATURE-----
On 26 Jun 2007, at 14:56 , Tres Seaver wrote:
Philipp von Weitershausen wrote:
On 25 Jun 2007, at 12:09 , Andreas Jung wrote:
--On 25. Juni 2007 10:13:44 +0200 Andreas Zeidler <az@zitc.de> wrote:
On Jun 25, 2007, at 10:05 AM, Philipp von Weitershausen wrote:
true, but i think philipp was only asking about 2.11 for the time being. i agree the fix should be backported, though. i'd even volunteer to do it, if nobody else wants to... :) That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10. when are they due? unfortunately i won't be able to do the backporting before thursday, since we've got a project release on wednesday... I defer the 2.9/2.10 for some days until I know how we should deal with the current state of the Zope 3.2/3.3 branches.
I realize you're waiting for a "release" of the Zope 3.2 branch to make another Zope 2.9.x release. I meant to respond earlier, apologies for the delay. I have no intentions of making another Zope 3.2 release (unless somebody really wants me to, in which case I'll do it). So for all I care, you can just tag Zope 3.2.3, put that in the external of Zope 2.9 and be done with it.
Please do the rest of the release managment before making that tag:
- Update the changelog to indicate the release date, SVN tag, etc. (on the branch before copying!)
- Update any version ID in the configure script.
Then make the tag.
Yeah, thanks for the reminder.
What's the matter with the Zope 3.3 branch then?
Don't we need a released 3.3.x?
Right, right. I meant to say: What's the matter with the latest Zope 3.3.x release? Are there any bugfixes in the Zope 3.3 branch that Zope 2 is waiting for? Otherwise we can just release with the latest Zope 3.3.x...
--On 26. Juni 2007 15:00:23 +0200 Philipp von Weitershausen <philipp@weitershausen.de> wrote:
Don't we need a released 3.3.x?
Right, right. I meant to say: What's the matter with the latest Zope 3.3.x release? Are there any bugfixes in the Zope 3.3 branch that Zope 2 is waiting for? Otherwise we can just release with the latest Zope 3.3.x...
There is basically nothing with the 3.2 and 3.3 branches. I just don't know about all details of the current 3.X development in order to feel comfortable enough to create a Zope 2 release from the current branches without having the OK from the related release masters. I could make (internal) tags on my own but this should be done in general by the responsible release manager (just for release safety) :-) Andreas
On Jun 25, 2007, at 10:05 AM, Philipp von Weitershausen wrote:
On 21 Jun 2007, at 23:47 , Andreas Zeidler wrote:
On Jun 21, 2007, at 9:07 PM, Tres Seaver wrote:
ah, here it is: http://svn.zope.org/?rev=76408&view=rev should fix the zclasses problem, if i'm not totally mistaken.
That change needs backporting, I think, to the ZODB 3.6 line (for 2.9) and 3.7 line (for 2.10).
true, but i think philipp was only asking about 2.11 for the time being. i agree the fix should be backported, though. i'd even volunteer to do it, if nobody else wants to... :)
That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10.
i did the backporting of jim's fix[1] to the 3.6 and 3.7 branches of zodb this morning and created two branches[2,3] of zope 2.9 and 2.10 with no further changes except the updated externals (pointing at the zodb branches, of course). all tests are passing, that is for the zodb branches as well as for both zope branches, and especially for zclasses, run individually and together with all other tests. so i reckon the fix could go into the next releases of 2.9 and 2.10, provided they are using new releases of zodb 3.6 and 3.7 in turn. plenty of work for andreas, i guess... :) ah, one note: just make sure you don't check out the below branches [2,3] under their original name. a regex bug in zope.configuration will give you failures caused by wrong include paths when the path of the checkout contains the string "tests"... cheers, andi [1] http://svn.zope.org/?rev=76408&view=rev [2] svn://svn.zope.org/repos/main/Zope/branches/witsch-2.9-with- passing-zclasses-tests [3] svn://svn.zope.org/repos/main/Zope/branches/witsch-2.10-with- passing-zclasses-tests -- zeidler it consulting - http://zitc.de/ - info@zitc.de friedelstraße 31 - 12047 berlin - telefon +49 30 25563779 pgp key at http://zitc.de/pgp - http://wwwkeys.de.pgp.net/ sprint with us! - http://plone.org/events/sprints/potsdam-sprint-2007
On 28 Jun 2007, at 17:03 , Andreas Zeidler wrote:
On Jun 25, 2007, at 10:05 AM, Philipp von Weitershausen wrote:
On 21 Jun 2007, at 23:47 , Andreas Zeidler wrote:
On Jun 21, 2007, at 9:07 PM, Tres Seaver wrote:
ah, here it is: http://svn.zope.org/?rev=76408&view=rev should fix the zclasses problem, if i'm not totally mistaken.
That change needs backporting, I think, to the ZODB 3.6 line (for 2.9) and 3.7 line (for 2.10).
true, but i think philipp was only asking about 2.11 for the time being. i agree the fix should be backported, though. i'd even volunteer to do it, if nobody else wants to... :)
That would be great, especially since we're on the verge of new maintenance releases for 2.9 and 2.10.
i did the backporting of jim's fix[1] to the 3.6 and 3.7 branches of zodb this morning and created two branches[2,3] of zope 2.9 and 2.10 with no further changes except the updated externals (pointing at the zodb branches, of course). all tests are passing, that is for the zodb branches as well as for both zope branches, and especially for zclasses, run individually and together with all other tests.
so i reckon the fix could go into the next releases of 2.9 and 2.10, provided they are using new releases of zodb 3.6 and 3.7 in turn. plenty of work for andreas, i guess... :)
If I know Andreas he won't do ZODB releases... So, apart from updating version.txt, CHANGES.txt, etc. and making a tag, I don't think there would be much else to a ZODB release... *wink* :)
On Jun 29, 2007, at 11:16 PM, Philipp von Weitershausen wrote:
If I know Andreas he won't do ZODB releases... So, apart from updating version.txt, CHANGES.txt, etc. and making a tag, I don't think there would be much else to a ZODB release... *wink* :)
hmm, i'm not sure if i should do any zodb releases, if that's what you're hinting at. i mean, i haven't been involved in zodb development at all and i've got way too little knowledge about zodb anyway -- especially of what's been going on in those branches recently... so imho it doesn't seem appropriate. as for andreas, you're right. i just saw that he had created the 3.6.2 tag and was therefore assuming he's been doing zodb releases as well, but in fact this was just one out of two tags he's done so far. my mistake... :) cheers, andi -- zeidler it consulting - http://zitc.de/ - info@zitc.de friedelstraße 31 - 12047 berlin - telefon +49 30 25563779 pgp key at http://zitc.de/pgp - http://wwwkeys.de.pgp.net/ sprint with us! - http://plone.org/events/sprints/potsdam-sprint-2007
participants (5)
-
Andreas Jung -
Andreas Zeidler -
Lennart Regebro -
Philipp von Weitershausen -
Tres Seaver