cast all the const buffers before passing them in and rely on
runtime crashes to find bugs rather than compile-time errors.

Thanks Microsoft. :D

The alternative is to explicitly copy/paste the type everywhere
forcing the compiler to do the only thing it could possibly do,
and making it painful to change a type later.

Honestly, in my opinion, these warnings should just be disabled
completely everywhere.

Found by MSVC

Just use the raw value and a comment.

We can't use const protection for the sample buffer because the Win32
API takes non-const buffers.  For standard implicit conversions,
simply disable the warning where it occurs on purpose.  It appears
that MSVC builds use the equivilent of -Wconversion (which isn't even
part of -Wall).

The case-sensitive/insensitive logic was backwards
(-i enabled case-sensitive sorting)

The re-sorting by name was always occurring when the -sort option was used.

Case-insensitive sorting by numeric properties (e.g. size) didn't work right.

Since it is possible for someone to change the add/import date of a file
after it's been added to the filebase, the natural/unsorted order might not
actually be the same as "date ascending", so let the sysop/script-author
really re-sort by date ascending if that's what they're trying to do.

I think this is addresses the issue that Nightfox and Nelgin were describing
in #synchronet.

(as a fall back to the full external program name).

... to determine/print the current external progarm being run.

This fixes issue #716

Make mqtt_putnodedat() use a 1-based node number, like the other putnodedat()
functions.

The bug (misuse of mqtt_putnodedat) was actually in js_system.c, but fixed the
API to match the common expectation.

DD File Lister: New sort option in the config file: PER_DIR_CFG, which has Synchronet sort the file list according to the file directory's configuration

See merge request !399

Eric Oulashin authored 1 year ago and Rob Swindell committed 1 year ago

DD File Lister: New sort option in the config file: PER_DIR_CFG, which has Synchronet sort the file list according to the file directory's configuration

DDLightbarMenu: DrawPartial() width check update (previously it wasn't re-drawing the last character of the item text)

See merge request !398

DDLightbarMenu: DrawPartial() width check update (previously it wasn't re-drawing the last character of the item text)

Hopefully, we'll find out why the random test fails sometimes...

The patch that sets the read timeout every time for getData() was
flawed, because getData() itself will sometimes set the read timeout
for the next call.

The patch was intended to ensure that the write timeout never
clobbers the read timeout, so now we track if getData() explicitly
set the timeout and if it did, we don't reset it.

It appears that if the timing is "just wrong", a large POST can
cause TLS to fail.  This has shown up as an inability to edit
pages in the wiki.

This is not a fix however, but simple a workaround until this is
root-caused.

If DL was passed a parameter equal to or greater than the number of
lines from the current line to the last line in the region, it would
try to move a negative-sized box up, which resulted in a crash.

This fixes that three ways:
1) Do not allow moving boxes with negative X/Y sizes
2) In dellines, clamp the lines to the number that could be deleted
3) Don't move lines up of all rest of the lines are being deleted

May fix the issues seen updating wiki pages, may not, but is a less
invasive patch either way.

The caller of this function must clobber the passed in pointer,
and we're returning NULL meaning we would be overwriting the pointer
to the allocation.

If this *does* work, it means that potentially a huge chunk of our
error checking hasn't actually worked in release builds or builds
that use the FORTIFY option.