554789 - configure chokes on system-sqlite not built with secure_delete though it _is_

Status: RESOLVED FIXED

(Firefox Build System :: General, defect)

Description

[Walter Meinl]

User-Agent:       Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.3a4pre) Gecko/20100323 Minefield/3.7a4pre
Build Identifier: 

checking for sqlite3 >= 3.6.23... yes
checking SQLITE_CFLAGS...  
checking SQLITE_LIBS... -lsqlite3  
checking for SQLITE_SECURE_DELETE support in system SQLite... no
configure: error: System SQLite library is not compiled with SQLITE_SECURE_DELETE.

This is a log from my Sqlite compilation
libtool: compile:  x86_64-pc-linux-gnu-gcc -DSQLITE_SECURE_DELETE -DSQLITE_CHECK_PAGES -DSQLITE_CORE -DNDEBUG -march=native -O2 -pipe -DSQLITE_OS_UNIX=1 -I. -I./src -D_HAVE_SQLITE_CONFIG_H -DNDEBUG -I/usr/include -DSQLITE_THREADSAFE=1 -DSQLITE_THREAD_OVERRIDE_LOCK=-1 -DSQLITE_ENABLE_COLUMN_METADATA -DSQLITE_ENABLE_RTREE -DSQLITE_ENABLE_FTS3 -DSQLITE_ENABLE_FTS3_PARENTHESIS -DUSE_TCL_STUBS=1 

configure checks for "SQLITE_SECURE_DELETE=1" "SQLITE_THREADSAFE=1" and SQLITE_ENABLE_FTS3=1" but the function check won't return !=1" for SQLITE_SECURE_DELETE and SQLITE_ENABLE_FTS3 (only for SQLITE_TRHEADSAFE

Reproducible: Always

Comment 1

[Walter Meinl]

Attachment: remove =1 in SECURE_DELETE and ENABLE_FTS3 checks

who could review this?

Comment 2

[Walter Meinl]

Comment on attachment 434727 [details] [diff] [review]
remove =1 in SECURE_DELETE and ENABLE_FTS3 checks

Mike, I guess debian wants to build with system-sqlite. I'd go for review from a Mozilla peer, but probably getting a review first from a distributor building this option would be helpful

Comment 3

[Mike Hommey [:glandium]]

Comment on attachment 434727 [details] [diff] [review]
remove =1 in SECURE_DELETE and ENABLE_FTS3 checks

This is the right thing to do according to what sqlite3_compileoption_used does.

Comment 4

[(not currently active) Ted Mielczarek]

Shawn: wasn't this problem reported somewhere else?

Comment 5

[Shawn Wilsher :sdwilsh]

(In reply to comment #4)
> Shawn: wasn't this problem reported somewhere else?
bug 551260 comment 15, and I asked for a new bug :)

What I'm confused about is that drh told me doing the =1 should work, so it's possible this is a bug with SQLite?

Comment 6

[D. Richard Hipp]

The "=1" thing in sqlite3_compileoption_used() only works for compile-time options that can take on multiple values, such as SQLITE_TEMP_STORE (which can be set to 0 through 3) or SQLITE_THREADSAFE (which can be set to 0 through 2).  It does not work for SQLITE_SECURE_DELETE or SQLITE_ENABLE_FTS3.  So I think Walter Meinl's patch is correct.

Comment 7

[Shawn Wilsher :sdwilsh]

(In reply to comment #6)
> The "=1" thing in sqlite3_compileoption_used() only works for compile-time
> options that can take on multiple values, such as SQLITE_TEMP_STORE (which can
> be set to 0 through 3) or SQLITE_THREADSAFE (which can be set to 0 through 2). 
> It does not work for SQLITE_SECURE_DELETE or SQLITE_ENABLE_FTS3.  So I think
> Walter Meinl's patch is correct.
Ah, thanks for the clarification!

Comment 8

[Reed Loden [:reed]]

http://hg.mozilla.org/mozilla-central/rev/104d6c3afe51