Summary: | dev-db/sqlite-3.5.4 with fts3 support | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | D.Raisin <info1> |
Component: | New packages | Assignee: | Arfrever Frehtes Taifersar Arahesis (RETIRED) <arfrever> |
Status: | RESOLVED FIXED | ||
Severity: | blocker | CC: | betelgeuse, gentoo, lama, laurento.frittella, luis, marek, spatz, steven |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 280114 | ||
Attachments: |
Portage overlay for sqlite 3.5.4.
Portage overlay for sqlite 3.5.4., second attempt diff against sqlite-3.5.9.ebuild Diff against sqlite 3.6.16 ebuild to add fts3 support patch for compiling sqlite-3.6.16 with fts3 sqlite-3.6.20-fts3 unified diff |
Description
D.Raisin
2008-01-27 16:25:58 UTC
Created attachment 141897 [details]
Portage overlay for sqlite 3.5.4.
I tested this sucessfully on x86
Please, do NOT attach tarballs to bugs. Plaintext patches is all we want. http://www.gentoo.org/proj/en/devrel/handbook/handbook.xml?part=2&chap=3#doc_chap2 Created attachment 141915 [details, diff]
Portage overlay for sqlite 3.5.4., second attempt
Is this how you want it?
(In reply to comment #3) > Created an attachment (id=141915) [edit] > Portage overlay for sqlite 3.5.4., second attempt > > Is this how you want it? > Best would be diff -u against existing ebuilds in the tree. Hello, How is this progressing? sqlite 3.5.4 is no longer in portage. Can this be implemented for newer versions? Instructions for building: http://www.sqlite.org/cvstrac/wiki?p=FtsOne Since it is beta, either portage can add an unstable USE flag feature or a separate module package can be created. Details about building the separate module should also be in the link. Best, Brent I'm not a gentoo developer, so I don't know how to get this patch put into the portage tree. All I could do now is to submit it. Maybe it still works with newer sqlite versions... Cheers, D.Raisin Created attachment 178555 [details, diff]
diff against sqlite-3.5.9.ebuild
Adding this functionality is much easier than the original attachments make it out to be.
Basically, I added a USE flag and it activates a couple of #defines. With this patch, I successfully built an FTS-enabled sqlite3 (and successfully tested the FTS functionality).
The use.local.desc entry should be something along the lines of "Fulltext search support for SQLite".
Created attachment 199080 [details, diff]
Diff against sqlite 3.6.16 ebuild to add fts3 support
This patches the latest ebuild to provide fts3 support. Several firefox plugins (the Brief RSS reader, for example) require this and fail in some strange and non-functional way if it sqlite is not built with fts3.
Created attachment 200562 [details, diff]
patch for compiling sqlite-3.6.16 with fts3
I only succeeded compiling after applying these changes against source tree of sqlite-3.6.16
Whats the status of this bug? :) I used 3.6.17 sqlite and added to src_configure line: append-cppflags -DSQLITE_ENABLE_FTS3=1 And it works - index and match data nicely. USE Flag or by default on would be nice (well on debian server it works also, no compiling needed) (In reply to comment #10) > And it works - index and match data nicely. USE Flag or by default on would be > nice (well on debian server it works also, no compiling needed) Same here. Please do something, devs. Created attachment 210186 [details, diff]
sqlite-3.6.20-fts3 unified diff
fts3 needs to be enabled by default, we are either all in or all out. Patch is avaliable only other change is the append-cpp flag to actually fully enable the build.
Without fts3 support built we can not move forward with mozilla packages, We need this handle as soon as possible. (In reply to comment #12) > > fts3 needs to be enabled by default, we are either all in or all out. Patch is > avaliable only other change is the append-cpp flag to actually fully enable the > build. > No it doesn't. There's now a fts3 use flag. Just use EAPI 2 use dependencies. This is not fixed. libtool: compile: x86_64-pc-linux-gnu-gcc -O2 -march=core2 -pipe -DSQLITE_OS_UNIX=1 -I. -I./src -D_HAVE_SQLITE_CONFIG_H -DNDEBUG -DSQLITE_ALLOW_XTHREAD_CONNECT=1 -DSQLITE_THREADSAFE=1 -DSQLITE_THREAD_OVERRIDE_LOCK=-1 -DSQLITE_OMIT_LOAD_EXTENSION=1 -DSQLITE_ENABLE_COLUMN_METADATA -DSQLITE_ENABLE_FTS3=1 -DSQLITE_TEMP_STORE=1 -c ./src/loadext.c -o loadext.o >/dev/null 2>&1 libtool: compile: x86_64-pc-linux-gnu-gcc -O2 -march=core2 -pipe -DSQLITE_OS_UNIX=1 -I. -I./src -D_HAVE_SQLITE_CONFIG_H -DNDEBUG -DSQLITE_ALLOW_XTHREAD_CONNECT=1 -DSQLITE_THREADSAFE=1 -DSQLITE_THREAD_OVERRIDE_LOCK=-1 -DSQLITE_OMIT_LOAD_EXTENSION=1 -DSQLITE_ENABLE_COLUMN_METADATA -DSQLITE_ENABLE_FTS3=1 -DSQLITE_TEMP_STORE=1 -c ./src/main.c -fPIC -DPIC -o .libs/main.o libtool: compile: x86_64-pc-linux-gnu-gcc -O2 -march=core2 -pipe -DSQLITE_OS_UNIX=1 -I. -I./src -D_HAVE_SQLITE_CONFIG_H -DNDEBUG -DSQLITE_ALLOW_XTHREAD_CONNECT=1 -DSQLITE_THREADSAFE=1 -DSQLITE_THREAD_OVERRIDE_LOCK=-1 -DSQLITE_OMIT_LOAD_EXTENSION=1 -DSQLITE_ENABLE_COLUMN_METADATA -DSQLITE_ENABLE_FTS3=1 -DSQLITE_TEMP_STORE=1 -c ./src/malloc.c -fPIC -DPIC -o .libs/malloc.o ./src/main.c:20:19: error: fts3.h: No such file or directory make: *** [main.lo] Error 1 make: *** Waiting for unfinished jobs.... libtool: compile: x86_64-pc-linux-gnu-gcc -O2 -march=core2 -pipe -DSQLITE_OS_UNIX=1 -I. -I./src -D_HAVE_SQLITE_CONFIG_H -DNDEBUG -DSQLITE_ALLOW_XTHREAD_CONNECT=1 -DSQLITE_THREADSAFE=1 -DSQLITE_THREAD_OVERRIDE_LOCK=-1 -DSQLITE_OMIT_LOAD_EXTENSION=1 -DSQLITE_ENABLE_COLUMN_METADATA -DSQLITE_ENABLE_FTS3=1 -DSQLITE_TEMP_STORE=1 -c ./src/malloc.c -o malloc.o >/dev/null 2>&1 The patch is needed for both header and as-needed. (In reply to comment #15) You should install dev-lang/tcl so that amalgamation build is used. I suggest to unconditionally require dev-lang/tcl only in DEPEND (DEPEND/RDEPEND already contains "tcl? ( dev-lang/tcl )"). (In reply to comment #15) > This is not fixed. > This warrants a new bug instead of using the old one so I opened bug 293405. forcing fts3 use flag causes emerge -upND world to fail and produce no output except suggesting that fts3 be added as a use flag. emerge -p world also fails with no output except the error. (In reply to comment #18) > forcing fts3 use flag causes emerge -upND world to fail and > produce no output except suggesting that fts3 be added as > a use flag. emerge -p world also fails with no output > except the error. > I have the same today spinoza ~ # emerge -uDav world These are the packages that would be merged, in order: Calculating dependencies... done! emerge: there are no ebuilds built with USE flags to satisfy ">=dev-db/sqlite-3.6.20-r1[fts3]". !!! One of the following packages is required to complete your request: - dev-db/sqlite-3.6.20-r1 (Change USE: +fts3) (dependency required by "www-client/mozilla-firefox-3.5.6" [ebuild]) (dependency required by "world" [argument]) spinoza ~ # (In reply to comment #18 and comment #19) This isn't any bug in the tree or in package manager. The message tells you to enable "fts3" USE flag for dev-db/sqlite, so you should do it. |