Sending state-of-art patch. Did multiple test runs on different machines (without obvious reason, the code didn't change), as always.
On 23.06.2017 at 10:36:43, Bart Schaefer (schaefer@xxxxxxxxxxxxxxxx) wrote:
> Nit-pick: This does not include "the following two paragraphs" in any of
> the licensing comments that refer to them, so technically it doesn't
> pass the license requirements for borrowing from db_gdbm.c.
I've added the two paragraphs
> We might want to update a few of the error strings to make phrasing
> and capitalization more similar to others in the shell, but there's
> only minimal consistency there anyway so probably not urgent.
I've changed messages to start lowercase
Wasn't sure what to do with the NFS bug. I did this:
if (!ht || ht->hsize == 0) {
(void)gdbm_reorganize(dbf);
}
So it reorganizes if clearing hash with =().
Tried compiling with "-ansi" on three machines, it was OK (except for C++ comments), but I've changed the middle-block declarations, to be C89.
As for folding, I attach Emacs screenshot, to show it really helps. Learned the folds from reading NERDTree source (vim plugin), it was huge, but had the folds and they helped.
--
Sebastian Gniazdowski
psprint /at/ zdharma.org
Attachment:
db_gdbm_2_modules_2.diff
Description: Binary data
Attachment:
emacs_folds.png
Description: PNG image
