| Age | Commit message (Collapse) | Author |
|---|
|
We introduce a new package structure for Coq:
- `coq-core`: Coq's OCaml tools code and plugins
- `coq-stdlib`: Coq's stdlib [.vo files]
- `coq`: meta-package that pulls `coq-{core,stdlib}`
This has several advantages, in particular it allows to install Coq
without the stdlib which is useful in several scenarios, it also open
the door towards a versioning of the stdlib at the package level.
The main user-visible change is that Coq's ML development files now
live in `$lib/coq-core`, for compatibility in the regular build we
install a symlink and support both setups for a while.
Note that plugin developers and even `coq_makefile` should actually
rely on `ocamlfind` to locate Coq's OCaml libs as to be more robust.
There is a transient state where we actually look for both
`$coqlib/plugins` and `$coqlib/../coq-core/plugins` as to support
the non-ocamlfind plus custom variables.
This will be much improved once #13617 is merged (which requires this
PR first), then, we will introduce a `coq.boot` library so finally
`coqdep`, `coqchk`, etc... can share the same path setup code.
IMHO the plan should work fine.
|
|
Contrarily to the comments, Coq_config.date was not the "release date" but
just another "compile date".
|
|
This an implementation of point 2 of CEP coq/ceps#48
https://github.com/coq/ceps/pull/48
Option -native-compiler of the configure script now impacts the
default value of the option -native-compiler of coqc. The
-native-compiler option of the configure script is added an ondemand
value, which becomes the default, thus preserving the previous default
behavior.
The stdlib is still precompiled when configuring with -native-compiler
yes. It is not precompiled otherwise.
|
|
This ends up in coq_makefile's CAMLFLAGS so we need it to make merlin
work properly.
|
|
We use an indirection, producing the sorted list of subdirectories of
plugins/, so that dune can recognize it hasn't changed and doesn't
rerun configure. Since configure regenerates a timestamp this avoids
recompiling the stdlib.
Fix #12750.
|
|
See CEP#44 for futher details.
|
|
Add headers to a few files which were missing them.
|
|
`plugins` needs to be present to coq_makefile variables are properly
initialized.
|
|
`dev/header.c` is not registered as a dependency, so the configure
step under dune fails in 32bit builds.
Note we don't detect the problem due to dubious code in configure
ignoring stderr messages on process calls.
|
|
|
|
|
|
This is a reduced version of #8503 as to provide a way to build the
reference manual with Dune.
Dune 1.6 supports (experimentally) directories as targets, thus we
introduce a rule that will call `sphinx` to build the manual.
This only provides build, however generation of `.install` rules is
not done, it will be hopefully addressed in #8503.
Note that we set `expire: 1 month` for all the artifacts we build with
Dune. IMHO this makes most sense as not to abuse Gitlab's hosting,
however of course we could consider a different deployment strategy if
wanted.
|
|
|
|
We make `config` into a properly library. This is more uniform and
useful for some clients. This also matches what was done in Dune.
Next step would be to push dependencies on `Coq_config` upwards, only
the actual toplevel binaries should depend on it.
We also remove the stale `camlp5.dbg` and refactor the dbg files a
bit, isolating the bits that are specific to the plugin / lib building
method used by makefile.
|
|
|
|
These are unused and not likely to come back.
|
|
In order to support sending the OPAM prefix to configure via Dune, we
introduced a `COQ_CONFIGURE_PREFIX` variable.
However, this had the pitfall that now the developer had to set it or
else face a hanging build due to configure expecting user input.
While we wait for a larger cleanup in `-prefix`, we introduce a
`no-ask` option in `./configure` that will avoid this problem. If
`-no-ask` is passed to `configure` no interactive question or display
will be shown to the user.
|
|
This is a hack to enable correct OPAM building, the medium-term plan
is to avoid having to set a prefix at configure time but instead using
a set of rules to locate the Coq library.
We use `(env_var ...)` in a dependency rule, which a feature of Dune 1.2
|
|
[Dune](https://github.com/ocaml/dune) is a compositional declarative
build system for OCaml. It provides automatic generation of
`version.ml`, `.merlin`, `META`, `opam`, API documentation; install
management; easy integration with external libraries, test runners,
and modular builds.
In particular, Dune uniformly handles components regardless whether
they live in, or out-of-tree. This greatly simplifies cases where a
plugin [or CoqIde] is checked out in the current working copy but then
distributed separately [and vice-versa]. Dune can thus be used as a
more flexible `coq_makefile` replacement.
For now we provide experimental support for a Dune build. In order to
build Coq + the standard library with Dune type:
```
$ make -f Makefile.dune world
```
This PR includes a preliminary, developer-only preview of Dune for
Coq. There is still ongoing work, see
https://github.com/coq/coq/issues/8052 for tracking status towards
full support.
## Technical description.
Dune works out of the box with Coq, once we have fixed some modularity
issues. The main remaining challenge was to support `.vo` files.
As Dune doesn't support custom build rules yet, to properly build
`.vo` files we provide a small helper script `tools/coq_dune.ml`. The
script will scan the Coq library directories and generate the
corresponding rules for `.v -> .vo` and `.ml4 -> .ml` builds. The
script uses `coqdep` as to correctly output the dependencies of
`.v` files. `coq_dune` is akin to `coq_makefile` and should be able to
be used to build Coq projects in the future.
Due to this pitfall, the build process has to proceed in three stages:
1) build `coqdep` and `coq_dune`; 2) generate `dune` files for
`theories` and `plugins`; 3) perform a regular build with all
targets are in scope.
## FAQ
### Why Dune?
Coq has a moderately complex build system and it is not a secret that
many developer-hours have been spent fighting with `make`.
In particular, the current `make`-based system does offer poor support
to verify that the current build rules and variables are coherent, and
requires significant manual, error-prone. Many variables must be
passed by hand, duplicated, etc... Additionally, our make system
offers poor integration with now standard OCaml ecosystem tools such
as `opam`, `ocamlfind` or `odoc`. Another critical point is build
compositionality. Coq is rich in 3rd party contributions, and a big
shortcoming of the current make system is that it cannot be used to
build these projects; requiring us to maintain a custom tool,
`coq_makefile`, with the corresponding cost.
In the past, there has been some efforts to migrate Coq to more
specialized build systems, however these stalled due to a variety of
reasons. Dune, is a declarative, OCaml-specific build tool that is on
the path to become the standard build system for the OCaml ecosystem.
Dune seems to be a good fit for Coq well: it is well-supported, fast,
compositional, and designed for large projects.
### Does Dune replace the make-based build system?
The current, make-based build system is unmodified by this PR and kept
as the default option. However, Dune has the potential
### Is this PR complete? What does it provide?
This PR is ready for developer preview and feedback. The build system
is functional, however, more work is necessary in order to make Dune
the default for Coq.
The main TODOs are tracked at https://github.com/coq/coq/issues/8052
This PR allows developers to use most of the features of Dune today:
- Modular organization of the codebase; each component is built only
against declared dependencies so components are checked for
containment more strictly.
- Hygienic builds; Dune places all artifacts under `_build`.
- Automatic generation of `.install` files, simplified OPAM workflow.
- `utop` support, `-opaque` in developer mode, etc...
- `ml4` files are handled using `coqp5`, a native-code customized
camlp5 executable which brings much faster `ml4 -> ml` processing.
### What dependencies does Dune require?
Dune doesn't depend on any 3rd party package other than the OCaml compiler.
### Some Benchs:
```
$ /usr/bin/time make DUNEOPT="-j 1000" -f Makefile.dune states
59.50user 18.81system 0:29.83elapsed 262%CPU (0avgtext+0avgdata 302996maxresident)k
0inputs+646632outputs (0major+4893811minor)pagefaults 0swaps
$ /usr/bin/time sh -c "./configure -local -native-compiler no && make -j states"
88.21user 23.65system 0:32.96elapsed 339%CPU (0avgtext+0avgdata 304992maxresident)k
0inputs+1051680outputs (0major+5300680minor)pagefaults 0swaps
```
|
|
|
|
longer use camlp4.
|
|
... in favor of having Public/Internal sub modules in each and
every module grouping functions according to their intended client.
|
|
|
|
We remove coqmktop in favor of a couple of simple makefile rules using
ocamlfind. In order to do that, we introduce a new top-level file that
calls the coqtop main entry.
This is very convenient in order to use other builds systems such as
`ocamlbuild` or `jbuilder`.
An additional consideration is that we must perform a side-effect on
init depending on whether we have an OCaml toplevel available [byte]
or not. We do that by using two different object files, one for the
bytecode version other for the native one, but we may want to review
our choice.
We also perform some smaller cleanups taking profit from ocamlfind.
|
|
The patch has three parts:
- Introduction of a configure flag `-bytecode-compiler (yes|no)`
(due to static initialization this is a configure-time option)
- Installing the hooks that register the VM with the pretyper and the
kernel conditionally on the flag.
- Replacing the normalization function in `Redexpr` by compute if the
VM is disabled.
We also rename `Coq_config.no_native_compiler` to `native_compiler`
and `Flags.native_compiler` to `output_native_objects` [see #4607].
|
|
Julien Narboux confirmed that it was dead code (GeoProof is not to be
confused with GeoCoq).
|
|
This seems a safe choice as of today, but more advanced users would
like to tweak it, or we could even refine it by a configure option if
desired.
|
|
We add a new option to configure `-flambda-opts` to allow passing
custom flags to flambda. Example:
```
./configure -flambda-opts "-O3 -unbox-closures"
```
|
|
E.g. -safe-string is set by configure.ml and made available to
both make (via config/Makefile) and coq_makefile (via
config/coq_config.ml -> lib/envars.ml -> CoqMakefile.in).
|
|
|
|
Plugin-writers can now use:
-bypass-API
parameter with "coq_makefile".
The effect of that is that instead of
-I API
the plugin will be compiled with:
-I config" -I dev -I lib -I kernel -I library -I engine -I pretyping -I interp -I parsing -I proofs -I tactics -I toplevel -I printing -I intf -I grammar -I ide -I stm -I vernac
|
|
This allows to centralize in the configuration file the description of
the 3 possible installation layouts (dispatched over directories
shared by multiple application as in unix, self-contained style like
in windows, local non-installation as with option -local).
Also supporting relocalisation when -prefix or -libdir and co is given.
|
|
This goes towards an approach where a local layout can be seen as an
installed layout.
|
|
They were not used for looking for coqide files in the situation when
the effective installation path happens to be exactly the installation
path proposed by default, while relevant files were however (possibly)
installed in these directories.
|
|
Note: "dependant" does exist, but it is a noun and it means a person that
is somehow financially dependent on someone else.
|
|
in error messages
|
|
|
|
|
|
Nothing is done for camlp4
There is an issue with computing camlbindir
|
|
|
|
NB: Please re-run ./configure after pulling this commit
For launching ocamlc/ocamlopt, coqmktop doesn't use Sys.command anymore,
but rather CUnix.sys_command, which is based on Unix.create_process.
This way, we do not need to bother with the underlying shell
(/bin/sh or cmd.exe) and the way arguments are to be quoted :-).
Btw, many cleanups of coqmktop.
Only difficulty: the -coqrunbyteflags of ./configure is a "meta-option"
that collect as a string several sub-options to be given by coqmktop to ocamlc.
For instance ./configure -coqrunbyteflags "-dllib -lcoqrun". We need now to
parse all these sub-options. To ease that, I've made the following changes
to the ./configure options:
* The -coqrunbyteflags and its blank-separated argument isn't accepted
anymore, and is replaced by a new option -vmbyteflags which expects
a comma-separated argument. For instance:
./configure -vmbyteflags "-dllib,-lcoqrun"
Btw, on this example, the double-quotes aren't mandatory anymore :-)
* The -coqtoolsbyteflags isn't accepted anymore. To the best of my
knowledge, nobody ever used it directly (it was internally triggered
as a byproduct of the -custom option). The only interest I can think of
for this option was to cancel the default use of ocamlc custom-linking
on Win32 and MacOS. For that, ./configure now provides a -no-custom option.
|
|
- Revised Coqtop.parse_args in a cleaner and lighter style
- Improved error message in case of argument parse failure:
* tell which option is expecting a related argument
* in case of unknown options, warn about them all at once
* do not hide the previous error messages by filling the
screen with usage(). Instead, suggest the use of --help.
- Specialized boolean config field Coq_config.arch_is_win32
- Faster Envars.coqlib, which is back to (unit->string), and
just access Flags.coqlib. Caveat: it must be initialized once
via Envars.set_coqlib
- Avoid keeping an opened channel to the "revision" file
- Direct load of theories/init/prelude.vo, no detour via Loadpath
Beware : ./configure must be runned after this commit
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16726 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
compiler (implemented on Matthieu's request).
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16240 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
Ok, I wasn't aware of the funny behavior of cd in presence of $CDPATH.
But the last "repair" was worse, trying to write into non-existing file
theories/config/coq_config.ml
Things should be better now:
* no more Coq_config.theories_dirs at all, since it was completely unused :-)
* concerning Coq_config.plugins_dirs, we list them without any "cd" into plugins,
hence keeping the "plugins/" part in their paths, and adapt accordingly
the only use (!) of plugins_dirs, in coq_makefile
Please run ./configure again after upgrading to this commit
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15860 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15715 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
* Since MinGW/Msys doesn't provide a cygpath utility, we emulate
it via an ocaml script tools/mingwpath.ml
* Avoid the crazy sed portions for backslash escaping, instead use
a more robust ocaml script tools/escape_string.ml based on
String.escaped
* No more config/Makefile.template + sed on it, but rather a
"cat << EOF > ..." as for config/coq_config.ml
* Normally, support of Cygwin should be preserved, as well as
mingw32 cross-compilation from linux (cf. myocamlbuild.ml)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15348 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
./configure -typerex .
It causes (non-fatal) errors when compiling files without a .mli (the problem
seems to have something to do with the flag -intf-suffix .cmi).
In practice, most typerex functionalities don't work well because typerex
fails its lookup into files compiled with -rectypes.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@15302 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
1/ man dir is now prefix/share/man and not prefix/man git diff!
2/ a data dir option for coqide extra data.
3/ typo
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14821 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14715 85f007b7-540e-0410-9357-904b9bb8a0f7
|
|
I assume that once Coq is installed in non-local mode and run from its
installed path, sources are no longer available. The coqsrc variable
doesn't make any sense, then, and its intended value can always be
inferred from Sys.executable_name. Moving it to Envars.coqroot.
Make coqlib optional. Currently, it is set to None only in -local mode
or with ocamlbuild. When set to None, -local layout is assumed
(binaries in ./bin, library in .). The behaviour should not be changed
when an explicit coqlib has been given to ./configure.
This commit should make it possible to run a Coq compiled with -local
from anywhere (no hard-coded absolute path embedded in the
executables, intermediary step to bug #2565). It WILL BREAK settings
re-using source trees after installation in non-local mode (are there
actual use cases for that?).
Hard-coded absolute paths still remain:
- in the build system, so the need to re-run ./configure after moving
the source tree is still expected for now;
- in coqrunbyteflags, I think we are limited by ocaml itself;
- docdir.
All absolute paths should be removed, ultimately.
As a side-effect, simplify computing of Envars.coqbin. I don't see any
good reason to keep it as a function.
Disclaimers:
- initialization of Sys.executable_name is not consistent across all
architectures; relying so much on it might trigger bugs. I'm pretty
sure something will explode if one adds arbitrary symlinks on top of
that;
- ocamlbuild stuff not tested.
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@14500 85f007b7-540e-0410-9357-904b9bb8a0f7
|
