CoolVLViewer/docs/pcre/README.Linden

0.  Pre-Checkin Checklist

    Performed from top of repo, default branch, head:

    [ ]  Is tag 'current' at or near head of 'vendor'?

         hg heads
         hg tags

    [ ]  Expected differences between vendor and default?  Very
         little of the original source should be modified.

         hg diff -rcurrent pcre

    [ ]  Are the 'vendor' and 'default' branch source directories
         'pcre' and not 'pcre-<version>'?

    [ ]  Does the pcregrep program (stage/bin/pcregrep) run as expected?
         Are the library bindings (ldd/otool/depends.exe) as expected?


1.  Introduction

    Simple build of pcre library from http://www.pcre.org/.

    With 8.35 repo structure follows standard conventions (see section
    at end).  Repo did not originally have a vendor branch so one was
    synthesized after-the-fact for 7.6.


2.  Modifications

    For Linux and Mac, builds are fairly ordinary with a few options
    explicitly chosen:

    * UTF-8 support
    * Unicode properties support
    * Disabled JIT

    Windows isn't pretty and several approaches are possible.  For
    this, I went with 'cmake' which was nominally already functional
    and was at least capable of doing the right thing with tests and
    generated header files.  CMakeLists.txt was modified to support a
    /Z7 option on static library builds controlled by the
    Linden.Win32.Cache file.  It also generates more status output
    after processing.

    But the model of the cmake writers was for fixed paths and the use
    of GUIs to configure the build.  This doesn't work in a headless
    build environment so a cache preload file,
    pcre/Linden.Win32.Cache, was created with the desired build
    options.  Cmake's cache preload code is, as usual, buggy and
    filenames need to be kept short.  But this seems to get the job
    done and results in working unit tests and installation passes
    with only a little touch-up work needed after the build.

    An additional problem with cmake and our build environment is that
    the resulting INSTALL and RUN_TESTS projects are not part of the
    solution build.  This gives our build_sln invocation of devenv
    errors as it trips over these disabled projects.  'msbuild' will
    work but you do need the .NET dev environment as the command lives
    down in it.  'devenv' invocations still work so that's what we do
    for the INSTALL and RUN_TESTS projects.

    * Should work, doesn't:

         build_sln INSTALL.vcxproj "Debug|Win32"

    * Does work, requires more junk in path:

         msbuild INSTALL.vcxproj /t:Build "/p:Configuration=Debug;Platform=Win32"

    * Ol' reliable:

         devenv PCRE.sln /Build "Debug|Win32" /Project INSTALL.vcxproj

    Another option was a custom Makefile.  I actually had this working
    but decided to go with 'cmake' in the hope of getting project
    files that would do more of the work.  Neither choice is a clear
    winner but cmake it is for now.  Here's some additional information
    for anyone wanting to visit Windows builds:

    ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/Contrib/pcre-winpcre.mak
    (An nmake Makefile that does much of the work but leaves out the
    configuration steps.

    http://www.rexegg.com/pcregrep-pcretest.html
    (Binary build for Windows.  Dig through that and you'll discover
    how they built their packages.)


3.  Source Origin

    8.35:
      http://sourceforge.net/projects/pcre/files/pcre/8.35/pcre-8.35.tar.gz/download


4.  Package Contents

    Common:
    * include/pcre/pcrecpp.h
    * include/pcre/pcre_stringpiece.h
    * include/pcre/pcre_scanner.h
    * include/pcre/pcrecpparg.h
    * include/pcre/pcreposix.h
    * include/pcre/pcre.h

    Windows (debug and release):
    * lib/release/pcre.lib
    * lib/release/pcrecpp.lib
    * lib/release/pcreposix.lib
    * lib/debug/pcrecppd.lib
    * lib/debug/pcred.lib
    * lib/debug/pcreposixd.lib

    Mac OS X (debug and release):
    * lib/release/libpcrecpp.a
    * lib/release/libpcreposix.a
    * lib/release/libpcre.a
    * lib/debug/libpcrecpp.a
    * lib/debug/libpcreposix.a
    * lib/debug/libpcre.a

    Linux (debug and release):
    * lib/release/libpcrecpp.a
    * lib/release/libpcreposix.a
    * lib/release/libpcre.a
    * lib/debug/libpcrecpp.a
    * lib/debug/libpcreposix.a
    * lib/debug/libpcre.a


5.  Consumers/Dependents

    Consumers of static/archive versions of this library should define
    PCRE_STATIC before including any header file.  This prevents some
    declspec ugliness.

    Packages dependent on pcre which will need attention
    (autobuild.xml) after changes.  This is not authoritative, use
    appropriate build tools to find all dependents.

    * colladadom

    * viewer


===================================================================

               Third-Party Library Repo Structure


Introduction

We want to have a way to capture local modifications to a third-party
open-source project, such as libcurl, without needing write access to
their public repository.  We want to be able to carry forward such
modifications to newer versions of the public project.  All this
should be independent of the organizational decision as to whether
it's even desirable to try to submit our local modifications upstream.

Fortunately, the Subversion folks articulated a process years ago that
addresses this very requirement.  They call it "Vendor Branches."  The
same tactic, suitably adapted, works with Mercurial too.

The essence of the idea is that we capture and tag a particular
snapshot of the open-source project.  We develop our local
modifications to that, and the repository tip incorporates them.  But
when we want to update to a newer version of the public project, we
bring it into the repository in such a way that we can discover the
changes from the original snapshot and the new one -- and then have
Mercurial apply those deltas to the ''combined'' source.

The following material is adapted from
http://svnbook.red-bean.com/en/1.1/ch07s05.html, the Red Bean
Subversion book, but recast for Mercurial.  The Linden source for this
material is an internal wiki.  There may be superceding documentation
on the public wiki when you read this.  We recommend searching there
for updates to conventions below.  And note that each particular
library may implement variations of this scheme.


General Vendor Branch Management Procedure

Managing vendor branches generally works like this.  You create a
named branch ("vendor") to store the vendor source snapshots.  Then
you import the third party code into that branch.  Your modified
branch (named "default") is based on "vendor".  You always make your
local changes to the default branch.  With each new release of the
code you are tracking you bring it into the "vendor" branch and merge
the changes into "default", resolving whatever conflicts occur between
your local changes and the upstream changes.

Perhaps an example will help to clarify this algorithm.  We'll use a
scenario where your development team is creating a calculator program
that links against a third-party complex number arithmetic library,
libcomplex.  We'll construct a repository specifically for our
locally-modified version of that library.  To begin, we must
initialize our repository and create at least one file in our
"default" branch.

 $ hg init ourcomplex
 $ cd ourcomplex
 $ touch README.txt
 $ hg commit README.txt

Now we can create the vendor branch and do the import of the first
vendor drop.  We'll call our vendor branch "vendor", and each
successive code drop will be tagged "current".

 $ hg branch vendor
 $ tar -xjf ../libcomplex-1.0.tar.bz2
 $ mv libcomplex-1.0 libcomplex
 $ hg addremove
 $ hg commit -m "1.0 source drop"
 $ hg tag -r tip current
 $ hg tag -r current 1.0

We now have the current version of the libcomplex source code in
branch "vendor", tagged "current" and in a non-version-specific source
code subdirectory ("libcomplex").  Next, we merge it into the default
branch.  It is in the default branch that we will make our
customizations.

 $ hg update default
 $ hg merge vendor
 $ hg commit -m "initial: 1.0"

We get to work customizing the libcomplex code.  Before we know it,
our modified version of libcomplex is now completely integrated into
our calculator program.

A few weeks later, the developers of libcomplex release a new version
of their library, version 1.1, which contains some features and
functionality that we really want.  We'd like to upgrade to this new
version, but without losing the customizations we made to the existing
version.  What we essentially would like to do is to replace our
current baseline version of libcomplex 1.0 with a copy of libcomplex
1.1, and then have Mercurial re-apply the custom modifications we
previously made to that library to the new version.  But we actually
approach the problem from the other direction, applying the changes
made to libcomplex between versions 1.0 and 1.1 to our modified copy
of it.

To perform this upgrade, we update our repository to our vendor
branch, and update the "current" tag with the new libcomplex 1.1
source code.  We quite literally replace the existing files with the
new files, clearing out the whole tree and exploding the libcomplex
1.1 release tarball in its place.  The goal here is to make the tip of
our vendor branch contain only the libcomplex 1.1 code, and to ensure
that all that code is under version control.  Oh, and we want to do
this with as little version control history disturbance as possible.

 $ hg update vendor
 $ rm -rf *
 $ tar -xjf ../libcomplex-1.1.tar.bz2
 $ mv libcomplex-1.1 libcomplex
 $ hg addremove -s 60
 $ # Additional 'hg add' and 'hg rm' commands if needed
 $ hg commit -m "1.1 source drop"

After unpacking the 1.1 tarball, hg status will show files with local
modifications as well as, perhaps, some unversioned or missing files.
If we did what we were supposed to do, the unversioned files are only
those new files introduced in the 1.1 release of libcomplex.  The
missing files are files that were in 1.0 but not in 1.1.  The 'hg
addremove' command deals with both, and more: the '-s 60' switch
directs Mercurial to compare added files to deleted files, recognizing
any file at least 60% similar as a move/rename.

For simple or stable libraries, the 'hg addremove' command should be
reliable.  For more complicated libraries subject to refactoring or
large gaps of time between updates (e.g. libcurl), it can get a little
lost trying to match files in the old release with files in the new
release.  Pay attention to the output of the command or better still,
do dry runs.  Files erroneously moved can be excluded with the '-X'
option and then dealt with individually with 'hg add' and 'hg rm'
commands after 'hg addremove'.  (The readme file in the curl library
should document a particularly challenging case.)

The 'addremove' process doesn't have to be perfect.  Recreating the
evolution of the upstream source tree isn't universally practical.
But we'd like to capture movement of files in the vendor branch that
are modified in the default branch.  If achieving that becomes too
tedious, then re-implementation of the default branch edit in a new
file is fine.  Just note it here for the next developer.

Finally, once our current working copy contains only the libcomplex
1.1 code, we commit the changes we made to get it looking that way.

Our current vendor branch now contains the new vendor drop.  We move
the 'current' tag to the new version (in the same way we previously
tagged the version 1.0 vendor drop), and then merge the differences
between the version 1.0 and version 1.1 into our default branch.

 $ hg tag -f -r tip current
 $ hg tag -r current 1.1
 $ hg update default
 $ hg merge vendor
 # resolve all the conflicts between their changes and our changes
 # if you will have conflicts in .hgtags, simply take *all* lines
 ...
 $ hg commit -m "update with 1.1"

Any additional work needed to get the merged library working can
now be done on the default branch.


Revision Tags

We don't currently make use of Mercurial tags in the build and release
process for 3rd-party libraries.  But we would like to establish a
convention to document update and release points.  The tags we would
like to establish are:

 * 'current' Points to a succession of vendor releases checked into
   the 'vendor' branch.  Will almost always be at or close to branch
   head.

 * '<version>' Tag on the 'vendor' branch pointing to a verbatim
   checkin of a 3rd-party's <version> release.  Example:  '7.21.1' for
   a particular version of libcurl we have used.

 * Release-type tags on the default branch aren't as useful given how
   Mercurial handles tags and how autobuild works.


Schematic of a Third-Party Repository

Below is the output of the 'hg glog' command showing a library project
going through an initial 1.0 release and an update from the vendor to
1.1.  Significant revisions in the repository lifecycle are as
follows:

 0  Creation of the repo with an initial file.
 1  1.0 code drop on branch 'vendor'
 4  Merge of 1.0 code onto branch 'default'
 5  Modifications to library we wish to keep over time.  Released.
 6  1.1 code drop on branch 'vendor'
 9  Merge of 1.1 code onto branch 'default'
10  Fixes to merge yielding production 1.1 library.  Released.
 

@  changeset:   10:888229641f6e
|  tag:         tip
|  user:        Monty Brandenberg <[email protected]>
|  date:        Wed Oct 30 13:35:51 2013 -0400
|  summary:     Work to get 1.1 merge working.  Release.
|
o    changeset:   9:925ccdf09f50
|\   parent:      5:83c5775c23dc
| |  parent:      8:977001a08e48
| |  user:        Monty Brandenberg <[email protected]>
| |  date:        Wed Oct 30 13:35:20 2013 -0400
| |  summary:     update with 1.1
| |
| o  changeset:   8:977001a08e48
| |  branch:      vendor
| |  user:        Monty Brandenberg <[email protected]>
| |  date:        Wed Oct 30 13:33:49 2013 -0400
| |  summary:     Added tag 1.1 for changeset 5f6cb89add91
| |
| o  changeset:   7:59bce0f6d12f
| |  branch:      vendor
| |  user:        Monty Brandenberg <[email protected]>
| |  date:        Wed Oct 30 13:33:41 2013 -0400
| |  summary:     Added tag current for changeset 5f6cb89add91
| |
| o  changeset:   6:5f6cb89add91
| |  branch:      vendor
| |  tag:         current
| |  tag:         1.1
| |  parent:      3:8525ad934ecd
| |  user:        Monty Brandenberg <[email protected]>
| |  date:        Wed Oct 30 13:33:29 2013 -0400
| |  summary:     1.1 source drop
| |
o |  changeset:   5:83c5775c23dc
| |  tag:         1.0
| |  user:        Monty Brandenberg <[email protected]>
| |  date:        Wed Oct 30 13:32:31 2013 -0400
| |  summary:     Linden-specific changes to the library.  Release
| |
o |  changeset:   4:bccb736585f4
|\|  parent:      0:400e4516c406
| |  parent:      3:8525ad934ecd
| |  user:        Monty Brandenberg <[email protected]>
| |  date:        Wed Oct 30 13:31:40 2013 -0400
| |  summary:     initial:  1.0
| |
| o  changeset:   3:8525ad934ecd
| |  branch:      vendor
| |  user:        Monty Brandenberg <[email protected]>
| |  date:        Wed Oct 30 13:30:21 2013 -0400
| |  summary:     Added tag 1.0 for changeset 8ac3828d03bb
| |
| o  changeset:   2:7aa1a1cb62d9
| |  branch:      vendor
| |  user:        Monty Brandenberg <[email protected]>
| |  date:        Wed Oct 30 13:30:14 2013 -0400
| |  summary:     Added tag current for changeset 8ac3828d03bb
| |
| o  changeset:   1:8ac3828d03bb
|/   branch:      vendor
|    tag:         1.0
|    user:        Monty Brandenberg <[email protected]>
|    date:        Wed Oct 30 13:30:09 2013 -0400
|    summary:     1.0 source drop
|
o  changeset:   0:400e4516c406
   user:        Monty Brandenberg <[email protected]>
   date:        Wed Oct 30 13:29:16 2013 -0400
   summary:     Created repo with initial readme file