Обсуждение: assignment type mismatch complaints
How concerned are we about assignment type mismatch warnings? I got a bunch in the mb stuff, and some in other places from the UnixWare 7.1.3 compiler. We still pass all regression tests. LER -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 E-Mail: ler@lerctr.org US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749
Can you send over a list of the errors and we will check them out. --------------------------------------------------------------------------- Larry Rosenman wrote: > How concerned are we about assignment type mismatch warnings? > > I got a bunch in the mb stuff, and some in other places from the > UnixWare 7.1.3 compiler. We still pass all regression tests. > > LER > -- > Larry Rosenman http://www.lerctr.org/~ler > Phone: +1 972-414-9812 E-Mail: ler@lerctr.org > US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749 > > > ---------------------------(end of broadcast)--------------------------- > TIP 3: if posting/reading through Usenet, please send an appropriate > subscribe-nomail command to majordomo@postgresql.org so that your > message can get through to the mailing list cleanly > -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania19073
Larry Rosenman <ler@lerctr.org> writes:
> How concerned are we about assignment type mismatch warnings?
They're probably all "char versus unsigned char" complaints?
There are a ton of them on compilers that care about it; most of
'em in the multibyte support. While it would be nice to clean up
all that someday, I can't say that I think it's a really profitable
use of time.
One difficulty is that the obvious fix (add a bunch of casts) is
probably a net degradation of the code. Explicit casts will hide
mismatches that are a lot worse than char signedness, and so
cluttering the code with them makes things more fragile IMHO.
I think an acceptable fix would involve running around and changing
datatype and function declarations; which is much more subtle and
thought-requiring than throwing in a cast wherever the compiler
burps.
regards, tom lane
On Sat, 2002-10-26 at 18:27, Tom Lane wrote: > Larry Rosenman <ler@lerctr.org> writes: > > How concerned are we about assignment type mismatch warnings? > > They're probably all "char versus unsigned char" complaints? Probably. The first few I looked at are PG_GETARG_CSTRING to unsigned char assignments. (I can send the whole list to either you, Tom, or the list). > > There are a ton of them on compilers that care about it; most of > 'em in the multibyte support. While it would be nice to clean up > all that someday, I can't say that I think it's a really profitable > use of time. Ok, I understand that. It seems that there are a bunch, but they are just warnings. > > One difficulty is that the obvious fix (add a bunch of casts) is > probably a net degradation of the code. Explicit casts will hide > mismatches that are a lot worse than char signedness, and so > cluttering the code with them makes things more fragile IMHO. > I think an acceptable fix would involve running around and changing > datatype and function declarations; which is much more subtle and > thought-requiring than throwing in a cast wherever the compiler > burps. Understand, and I don't expect it to happen in a beta test :-). > > regards, tom lane -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 E-Mail: ler@lerctr.org US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749
On Sat, 2002-10-26 at 18:34, Larry Rosenman wrote: > On Sat, 2002-10-26 at 18:27, Tom Lane wrote: > > Larry Rosenman <ler@lerctr.org> writes: > > > How concerned are we about assignment type mismatch warnings? > > > > They're probably all "char versus unsigned char" complaints? > Probably. The first few I looked at are PG_GETARG_CSTRING to unsigned > char assignments. (I can send the whole list to either you, Tom, or the > list). > > > > > There are a ton of them on compilers that care about it; most of > > 'em in the multibyte support. While it would be nice to clean up > > all that someday, I can't say that I think it's a really profitable > > use of time. > Ok, I understand that. It seems that there are a bunch, but they are > just warnings. > > > > One difficulty is that the obvious fix (add a bunch of casts) is > > probably a net degradation of the code. Explicit casts will hide > > mismatches that are a lot worse than char signedness, and so > > cluttering the code with them makes things more fragile IMHO. > > I think an acceptable fix would involve running around and changing > > datatype and function declarations; which is much more subtle and > > thought-requiring than throwing in a cast wherever the compiler > > burps. > Understand, and I don't expect it to happen in a beta test :-). If anyone wants to look at these: ftp://ftp.lerctr.org/pub/pg-dev/gmake.out.txt Thanks, LER > > > > > > regards, tom lane > -- > Larry Rosenman http://www.lerctr.org/~ler > Phone: +1 972-414-9812 E-Mail: ler@lerctr.org > US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749 > > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster > -- Larry Rosenman http://www.lerctr.org/~ler Phone: +1 972-414-9812 E-Mail: ler@lerctr.org US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749