Apache OpenOffice (AOO) Bugzilla – Issue 110427
native0 integrated a (different from DEV300/cws mysqlconnector) mysqlc into OOO320?
Last modified: 2013-08-07 15:35:05 UTC
In a OOO320_m14 clone: $ ls my* mysqlc: images prj source version.mk mysqlcppconn: cppconn_config.pmk makefile.mk mysql-connector-cpp.patch download mysql_config.pmk prj oh? cws mysqlconnector was only integrated for DEV300. Was it intended that it went over to OOO320 via native0? (Who "fixed" it for rebranding and probably thus just merged the cws). If I look at e.g. mysqlc/version.mk: $ hg log version.mk Änderung: 263405:b562f62892bf Nutzer: obo <obo@openoffice.org> Datum: Tue Mar 09 12:48:44 2010 +0100 Zusammenfassung: #i10000# commit repaired changeset Änderung: 263399:28f77bc32e70 Nutzer: Dirk Voelzke <dv@openoffice.org> Datum: Thu Mar 04 16:00:23 2010 +0100 Zusammenfassung: native0ext:#161688# rebranding extensions ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Even more, those two copies (what went into OOO320 and what went into DEV300 via cws mysqlcppconn) differ; see attached patch. Was that intended? If yes, close this and nevermind, but as noone was told.. And if the changes done in OOO320 will get lost in DEV300 that's bad, too 1.0.1 final vs. 1.0.1 alpha.... And the version in OOO320 introduced again a build bug with internal mysqlcppconn (yes, it's a bug to -lmysqlcppconn in mysqlcppconn's build). Can this be cleaned up? please ASAP?
Ccing hr, obo, fs, md, of again one case of really good looking what goes into the tree...
Created attachment 68567 [details] the differences
the differences: - branding (ok..) - extension identifier changes (bad, as the extension with the arch-dep identifier already is on extensions.s.oo and in use in distribution packages= - supress-if-required (ok...) - loosing the buildfixes in mysqlc/source/makefile.mk and mysqlcppconn/mysql_config.pmk - state change Alpha -> Final. Intended? Shouldn't this better have been in a extra cws if so?
either way it's intended (then the bugs should be fixed, for "next build"), or it's not and it should be reverted from OOO320 (at which time it's "next build" in any case)
oh, and you loosed some other bugfixes, including the "default to tcp" one...
Created attachment 68602 [details] correct diff
fs->rene: you should compare the right things before weeping. The DEV300-version which you used for comparison was never intended to be copied to OOO320. If we would have done so, you would probably have cried, too, since this is indeed an alpha version, which differs from the release extension. Instead, the version from CWS mysqlnative (at a point in SVN where this CWS did not yet see unsuccessful cross-code-line rebase attempts) which was released as 1.0.0 was copied over to the OOO320 code line, with minor adjustments. mysqlc-ppconn.diff contains the differences between the released 1.0.0 extension version, and the version in OOO320.m14. As you can see, this is indeed basically rebranding (plus a change to the extension's OXT file name, whose cause I do not know, but which is certainly uncritical). Resolving issue as INVALID.
closing
> fs->rene: you should compare the right things before weeping. The DEV300- > version which you used for comparison was never intended to be copied to > OOO320. If we would have done so, you would probably have cried, too, since > this is indeed an alpha version, which differs from the release extension. No, I wouldn't. Actually I right now have a extension still based on mysqlnative and had mysqlconnector backported now. > Instead, the version from CWS mysqlnative (at a point in SVN where this CWS > did not yet see unsuccessful cross-code-line rebase attempts) which was > released as 1.0.0 was copied over to the OOO320 code line, with minor > adjustments. Aha. So it was intended. Good that people actually get told. > mysqlc-ppconn.diff contains the differences between the released 1.0.0 > extension version, and the version in OOO320.m14. As you can see, this is > indeed basically rebranding (plus a change to the extension's OXT file name, > whose cause I do not know, but which is certainly uncritical). - EXTENSION_ZIPNAME:=$(EXTENSIONNAME)-$(EXTENSION_VERSION_BASE)-$(MYSQLC_STATUS:s/ /-/:l)-$(RTL_OS:l)-$(RTL_ARCH:l) + EXTENSION_ZIPNAME:=$(EXTENSIONNAM I'd not call this "rebranding" but "changing the extension identifier" Besides that, what you added to OOO320 DOES miss the buildfixes I did in cws mysqlnative. Should I really need patches to apply myself again because you cannot get them over properly? > Resolving issue as INVALID. If you wish, I now now that the merge was intended (of course no one told anyone) but that the extension identifier changes is BAD. And the lost buildfixes too.
(and I needed cws mysqlconnector because of Issue 110068, where the patch didn't apply against the old mysqlnative iirc)
> I'd not call this "rebranding" but "changing the extension identifier" And even if you wanted to get rid of parts, I don't think you wanted to get rif of the arch-stuff there (or do you? I'd be happy) The current change just results in changing from Identifier: com.sun.star.mysql-connector-ooo-linux_x86_64 to Identifier: com.sun.star.mysql-connector-ooo
reopening because of the extension identifier change at least (and we REALLY should get the buildfixes in)
> I'd not call this "rebranding" but "changing the extension identifier" No, the *extension identifier* is the thing mentioned in the description.xml. What was changed here is the name of the resulting .oxt file, which has nothing to do with the extension identifier. In fact, even in CWS mysqlnative, the .oxt files as generated during build were (for at least one platform) renamed before being uploaded to the extension site ... An now, I cannot answer you why this change was made, that's up to Dirk to answer, who (I think) did it. > Besides that, what you added to OOO320 DOES miss the buildfixes I did in > cws mysqlnative. Should I really need patches to apply myself again because > you cannot get them over properly? Hmm - is it possible that you added those fixes after the 1.0 release? We just copied over the 1.0 release version. But since SVN screw up the thing during a rebase attempt, it might well be possible we missed something there. Re-closing as INVALID. Nearly all of the stuff you claimed in the original description is bogus, including the "extension identifier" thing, as explained above. If you need a build issue to be fix, submit a usual build breaker issue, please.
> An now should read "And no"
closing.
> Hmm - is it possible that you added those fixes after the 1.0 release? We just > copied over the 1.0 release version. But since SVN screw up the thing during a > rebase attempt, it might well be possible we missed something there. i did them when I did them. As you can't properly sync extension released with code, yes, I might have committed it after the release- > Re-closing as INVALID. Nearly all of the stuff you claimed in the original > description is bogus, If you add mysqlc to OOO320 silently and loose important stuff, I should be wary, shouldn't I? You should have told people that you did this... > including the "extension identifier" thing, as explained above. True. That might be my error. > If you need a build issue to be fix, submit a usual build breaker issue, please. as you wish. Issue
"If you add mysqlc to OOO320 silently" In a BUGFIX release. 3.2.1..
> If you need a build issue to be fix, submit a usual build breaker issue, please. there you go. Issue 110460