| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
This fixes #9767 by silently ignoring input lines which are not valid
UTF-8. We hereby assume that all file paths are valid UTF-8.
We also now actually test both python2 and python3 on the CI.
|
|
If you have file1.mlg and file2.ml, with file2 depending on file1,
ocamldep was before generating file1.ml so wouldn't generate
[file2.cmx: file1.cmx] (ocamldep is silent on non-found dependencies).
This has been causing nondeterministic failures in quickchick
recently.
I guess it didn't come up in the past because ml4 files tend to be at
the end of the dependency chain.
|
|
The kernel no longer has to read the configure flag, its value can now
be overriden by a coqtop/coqc argument, and more generally is easier to
set from a toplevel (such as the checker).
We also add a `-bytecode-compiler` flag.
Fixes #4607
|
|
|
|
Fixes #8158
|
|
|
|
There was recently a spurious failure on AppVeyor (I've forgotten which
PR). This commit fixes that particular failure.
|
|
|
|
This fixes #5675 (in yet another way).
The issue was that `$` (end of string regex) was not properly escaped in
`"`s.
This handles the issue that is displayed in
```
cat A.v.timing.diff
After | Code | Before || Change | % Change
---------------------------------------------------------------------------------------------------
0m01.44s | Total | 0m01.56s || -0m00.12s | -7.92%
---------------------------------------------------------------------------------------------------
0m00.609s | Chars 163 - 208 [Definition~foo1~:=~Eval~comp~i...] | 0m00.627s || -0m00.01s | -2.87%
0m00.527s | Chars 069 - 162 [Definition~foo0~:=~Eval~comp~i...] | 0m00.552s || -0m00.02s | -4.52%
0m00.304s | Chars 000 - 026 [Require~Coq.ZArith.BinInt.] | 0m00.379s || -0m00.07s | -19.78%
N/A | Chars 027 - 068 [Declare~Reduction~comp~:=~nati...] | 0m00.006s || -0m00.00s | -100.00%
0m00.s | Chars 027 - 068 [Declare~Reduction~comp~:=~vm_c...] | N/A || +0m00.00s | N/A
--- A.v.timing.diff.desired.processed 2018-03-23 22:22:19.000000000 +0000
+++ A.v.timing.diff.processed 2018-03-23 22:22:19.000000000 +0000
@@ -1,4 +1,4 @@
- N/A | Chars 27 - 68 [Declare~Reduction~comp~:=~nati] | ms || -ms | N/A
+ N/A | Chars 27 - 68 [Declare~Reduction~comp~:=~nati] | ms || -ms | -%
------
------
After | Code | Before || Change | % Change
```
where, because `Declare Reduction` takes 0.006s rather than 0s, the %
change shows up as -100% rather than N/A.
|
|
|
|
As per https://github.com/coq/coq/pull/6756/files#r168028764
|
|
|
|
|
|
|
|
|
|
This fixes #6350 (and even comes with a test case).
Refering to other directories as `-R … ""` is maybe not best practice,
but some people out there do it, so as long as it does not cause too
much trouble, we can continue to support it.
|
|
When none of the numbers get over 100, the width of the table was
different.
See https://github.com/coq/coq/pull/6736#issuecomment-365386802
|
|
|
|
Should help with
https://github.com/coq/coq/issues/5675#issuecomment-353604702
Also replace a tab with spaces
|
|
This should help with #5675, in particular with
https://github.com/coq/coq/issues/5675#issuecomment-349716292
|
|
... in favor of having Public/Internal sub modules in each and
every module grouping functions according to their intended client.
|
|
|
|
|
|
subdirectory.
|
|
This allows us to avoid doing git clean.
|
|
Calling the test a second time after a make clean was failing due to
an existing "src" directory left by the first call.
|
|
These work on precomputed build logs (in this case, from a recent
partial build of fiat-crypto).
They are meant to serve as human-readable sanity checks of output
format.
Separate out the sane bits of template/init.sh from the ones messing
with directory structure (which are fragile and make assumptions about
where the calling script is sourcing it from).
N.B. The test-suite removes all *.log files, so we use *.log.in.
N.B. We set COQLIB in precomputed-time-tests/run.sh, not the Makefile,
because coqc, on Windows, doesn't handle cygwin paths passed via
-coqlib, and `pwd` gives cygwin paths.
N.B. We have .gitattributes to satisfy the linter (as per
https://github.com/coq/coq/pull/6149#issuecomment-346410990)
|
|
This should help debug things like
https://github.com/coq/coq/issues/5675#issuecomment-345324924 if they
ever show up again.
|
|
The two lines that this commit remove are spurious as a `git clean -dfx || true`
is already performed in `test-suite/coq-makefile/template/init.sh`.
While this resolves the accidental dependency on git, I am still unhappy with
this call of `git clean -dfx`.
|
|
This should (hopefully) fix #5675.
|
|
The validity of this test depended on Makefile issueing warnings in English.
We simply force the global language of the shell to be C.
|
|
|
|
|
|
CAMLPKGS is now used to hold extra findlib -packages
The previous solution was to use CAMLFLAGS but since 4.05 an
invocation of `ocamlopt -pack useless.cmxa foo.cmx -o packedfoo.cmx`
fails saying that `useless.cmxa` is not a compilation unit description.
CAMLPKGS is used in all `ocamlopt` invocations but for the one
performing the packing.
|
|
|
|
|
|
Replacing ; with && and enabling bash's pipefail option
|
|
This commit adds timing scripts from
https://github.com/JasonGross/coq-scripts/tree/master/timing into the
tools folder, and integrates them into coq_makefile and Coq's makefile.
The main added makefile targets are:
- the `TIMING` variable - when non-empty, this creates for each built
`.v` file a `.v.timing` variable (or `.v.before-timing` or
`.v.after-timing` for `TIMING=before` and `TIMING=after`, respectively)
- `pretty-timed TGTS=...` - runs `make $(TGTS)` and prints a table of
sorted timings at the end, saving it to `time-of-build-pretty.log`
- `make-pretty-timed-before TGTS=...`, `make-pretty-timed-after
TGTS=...` - runs `make $(TGTS)`, and saves the timing data to the file
`time-of-build-before.log` or `time-of-build-after.log`, respectively
- `print-pretty-timed-diff` - prints a table with the difference between
the logs recorded by `make-pretty-timed-before` and
`make-pretty-timed-after`, saving the table to
`time-of-build-both.log`
- `print-pretty-single-time-diff BEFORE=... AFTER=...` - this prints a
table with the differences between two `.v.timing` files, and saves
the output to `time-of-build-pretty.log`
- `*.v.timing.diff` - this saves the result of
`print-pretty-single-time-diff` for each target to the
`.v.timing.diff` file
- `all.timing.diff` (`world.timing.diff` and `coq.timing.diff` in Coq's
own Makefile) - makes all `*.v.timing.diff` targets
N.B. We need to make `make pretty-timed` fail if `make` fails. To do
this, we need to get around the fact that pipes swallow exit codes.
There are a few solutions in
https://stackoverflow.com/questions/23079651/equivalent-of-pipefail-in-gnu-make;
we choose the temporary file rather than requiring the shell of the
makefile to be bash.
|
|
|
|
|
|
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
|
|
|
|
|
|
|
|
|
|
|
|
|
|
It used to generate only .cmo (the packed one).
While this works if the plugin has no external dependencies,
it does not if it does.
The bug affected only bytecode builds
|
