Checkout a branch named b<version-number>
. Then,
$ opam pin add ocamloscope .
The OPAM file is at opam/opam
.
$ mkdir out
$ oco scrape
The command scrapes the installed OCamlFind packages and save the result under out/
directory. All the packages must be compiled with -bin-annot
compiler option and the build files must be kept. Here is an example of such a setting:
$ export OCAMLPARAM=_,_,bin-annot=1
$ export OPAMKEEPBUILDDIR=1
Compilers and packages should be installed using OPAM for easier scraping. If the compiler or some packages are installed by hand, not by OPAM, you have to specify the source directory of these softwares explicitly with --src-dir
option. For example:
$ oco scrape --src-dir $HOME/build/ocaml-4.03.0 --src-dir $HOME/mysrc/mypackage
Scraping may fail if source files or .cm*
files of scraping modules are not found.
oco scrape
creates the following files:
<package>.sig
: Extracted type information from*.cmi
files.<package>.hump
: Program abstraction from*.cmt
and*.cmti
files.<package>.dat
: The final data file of the package, which links<package>.sig
,<package>.hump
and*.hump
files which the package relys on.
You can send <package>.dat
files to OCamlOScope server maintainer to add the package to the server db.
After scraping, you have to manually link these *.dat
files into one:
$ oco link
The scraped data are linked together into one file: out/all.all
.
After linking, you can start a search session:
$ oco search
...
? <input your query>