Brian Drummond wrote:
Can you apply this patch over a clean 0.31
and report if it fixes your issues?
https://sourceforge.net/p/ghdl-updates/code/ci/74c33ea9f3bcaf833c6683114a4d90c8684cfff2/
A quick test of the indexing testcases in ticket #9 shows the abnormal
exception is gone.
Bug 16782 also now runs OK.
While the error messages remain less helpful than the gcc equivalent,
that seems to be a present shortcoming of the mcode version.
If so, I believe we're ready for a Windows release.
I'll put a zip file together after I double check[1] my build and run
it against the full test suite.
This round of (binary/package/port) releases is based
on the 0.31 source package plus the MINIMUM set of
patches required to successfully build each OS package.
<snip>
none of these are added to the 0.31 branch.
Two release build questions relating the above comment to Adam:
Q1) If we're not adding the various OS patches to the 0.31 branch to
create a 0.31.1 tag, then what exactly should I call this thing?
0.31.1 ? 0.31-mcode-windows-patched ?
Referring here to the release naming conventions to use for the
"ghdl -v" version string, the zip file, and the default installation
directory.
Q2) Is it sufficient to include the final patch file in the release zip
file, or should it be checked in somewhere else?
-Brian
[1] test build notes
I manually applied the 74c33e changeset patch file to the 0.31
release for the above quick test.
(The 0.31 tarball source directory I used wasn't versioned, so I
couldn't use hg import.
And the Windows version of GNU patch I'm running was having weird
problems digesting the diff.)
Also note, that 74c33e changeset has libraries.adb edits that don't
apply to 0.31
_______________________________________________
Ghdl-discuss mailing list
[email protected]
https://mail.gna.org/listinfo/ghdl-discuss
