Обсуждение: ODBC changes
Hi, I've committed several bug fixes. Psqlodbc is 7.01.0005 now. 1) Parameter handling patch by Hiroki Kataoka. This fixes bugs reported by Ludek Finstrle, Matteo Cavalleri and Johann Zuschlag. 2) Decrease the size of some buffers to avoid stack overflow error. 3) Fix for bug reports by Andrea Aime and Jan Wieck. regards, Hiroshi Inoue
Dave Page wrote: > > Precomiled versions for Win32 will be available shortly... > > Hiroshi, I've attached a fixed copy of win32.mak - the one in CVS seems to > be full of ^M's at the end of every line which makes nmake very unhappy. > Could you commit it to CVS please if you're happy with it. > Hmm there's no difference between my win32.mak and what you sent to me. My win32.mak is 12222bytes big. How many bytes does the one in cvs have ? regards, Hiroshi Inoue
Hiroshi Inoue <Inoue@tpf.co.jp> writes:
> Dave Page wrote:
>> Hiroshi, I've attached a fixed copy of win32.mak - the one in CVS seems to
>> be full of ^M's at the end of every line which makes nmake very unhappy.
>> Could you commit it to CVS please if you're happy with it.
> Hmm there's no difference between my win32.mak and what
> you sent to me. My win32.mak is 12222bytes big.
> How many bytes does the one in cvs have ?
Right now, src/interfaces/odbc/win32.mak in CVS *is* 12222 bytes
and it *is* full of DOS-style newlines (try od -c if you doubt it).
I have no idea whether that's a problem. Given that no one on a
non-Windoze platform has any use for the file, I'd have guessed that
DOS-style newlines are the way to go ... but if we can get rid of the
CRs without breaking things on Windows then I'm for it ...
regards, tom lane
Hiroshi Inoue <Inoue@tpf.co.jp> writes:
> I see my copy is 12,222 bytes on both linux and cygwin.
> I also heard from Eiji Tokuya that he sometimes(not always)
> saw corrupted win32.mak. Unfortunately I don't understand
> where the cause is.
$ cd /home/postgres/pgsql/src/interfaces/odbc
$ rm win32.mak
$ cvs update
cvs server: Updating .
U win32.mak
$ ls -l win32.mak
-rw-r--r-- 1 postgres users 12222 May 2 01:23 win32.mak
$ od -c win32.mak | head
0000000 \r \n # \r \n # F i l e : \t \t \t w
0000020 i n 3 2 . m a k \r \n # \r \n # D
0000040 e s c r i p t i o n : \t \t p s q
0000060 l o d b c M a k e f i l e f
0000100 o r W i n 3 2 . \r \n # \r \n #
0000120 C o n f i g u r a t i o n s : \t
0000140 D e b u g , R e l e a s e ,
0000160 M u l t i b y t e D e b u g ,
0000200 M u l t i b y t e R e l e a s e
0000220 \r \n # B u i l d T y p e s :
I see nothing debatable here: what is on the CVS server is a file
containing CR/LF newlines. If you have something else then it's
been through a transfer process that altered the file.
The question to be discussed is whether that's the way to distribute
the file or not. We can certainly remove the CRs if not.
BTW, by my count a Unix-newline version of this file would be 11736
bytes. A 12708-byte file would be broken in some strange new way
(two LFs per newline, maybe?)
regards, tom lane
Precomiled versions for Win32 will be available shortly... Hiroshi, I've attached a fixed copy of win32.mak - the one in CVS seems to be full of ^M's at the end of every line which makes nmake very unhappy. Could you commit it to CVS please if you're happy with it. Regards, Dave. > -----Original Message----- > From: Hiroshi Inoue [mailto:Inoue@tpf.co.jp] > Sent: 01 May 2001 03:40 > To: pgsql-odbc@postgresql.org > Subject: [ODBC] ODBC changes > > > Hi, > > I've committed several bug fixes. > Psqlodbc is 7.01.0005 now. > > 1) Parameter handling patch by Hiroki Kataoka. > This fixes bugs reported by Ludek Finstrle, > Matteo Cavalleri and Johann Zuschlag. > 2) Decrease the size of some buffers to avoid > stack overflow error. > 3) Fix for bug reports by Andrea Aime and Jan Wieck. > > regards, > Hiroshi Inoue > > ---------------------------(end of > broadcast)--------------------------- > TIP 5: Have you checked our extensive FAQ? > > http://www.postgresql.org/users-lounge/docs/faq.html >
Вложения
> -----Original Message----- > From: Hiroshi Inoue [mailto:Inoue@tpf.co.jp] > Sent: 01 May 2001 10:12 > To: Dave Page > Cc: pgsql-odbc@postgresql.org > Subject: Re: [ODBC] ODBC changes > > > Dave Page wrote: > > > > Precomiled versions for Win32 will be available shortly... > > > > Hiroshi, I've attached a fixed copy of win32.mak - the one > in CVS seems to > > be full of ^M's at the end of every line which makes nmake > very unhappy. > > Could you commit it to CVS please if you're happy with it. > > > > Hmm there's no difference between my win32.mak and what > you sent to me. My win32.mak is 12222bytes big. > How many bytes does the one in cvs have ? A fresh copy from CVS comes in at 12,708 bytes on my system (and yes, I agree the one I sent is 12,222 bytes). Regards, Dave.
> -----Original Message----- > From: Tom Lane [mailto:tgl@sss.pgh.pa.us] > Sent: 02 May 2001 15:26 > To: Dave Page > Cc: Hiroshi Inoue; pgsql-odbc@postgresql.org > Subject: Re: [ODBC] ODBC changes > > > Dave Page <dpage@vale-housing.co.uk> writes: > > I did an update to my copy of the source (using WinCVS), > attempted nmake /f > > win32.mak to build the ODBC driver and got an error. > Editting win32.mak > > using vi under cygwin showed a ^M (LF i believe) at the end > of every line. > > Using Programmers File Editor (a very good windows editor) > shows a blank > > line between every line. Removing the ^M's cured the > problem. I did also get > > a fresh copy from CVS to ensure that it wasn't just my copy that was > > affected. > > After further investigation it seems that win32.mak did > originally have > Unix-style newlines (\n), and they were all changed to DOS > style (\r\n) > in the last commit. I have changed them back, on the assumption that > this conversion was unintentional. Thanks Tom - it was probably my fault in the first place :-( Regards, Dave.
> -----Original Message----- > From: Tom Lane [mailto:tgl@sss.pgh.pa.us] > Sent: 02 May 2001 06:31 > To: Hiroshi Inoue > Cc: Dave Page; pgsql-odbc@postgresql.org > Subject: Re: [ODBC] ODBC changes > > > Hiroshi Inoue <Inoue@tpf.co.jp> writes: > > I see my copy is 12,222 bytes on both linux and cygwin. > > I also heard from Eiji Tokuya that he sometimes(not always) > > saw corrupted win32.mak. Unfortunately I don't understand > > where the cause is. > > $ cd /home/postgres/pgsql/src/interfaces/odbc > $ rm win32.mak > $ cvs update > cvs server: Updating . > U win32.mak > $ ls -l win32.mak > -rw-r--r-- 1 postgres users 12222 May 2 01:23 win32.mak > $ od -c win32.mak | head > 0000000 \r \n # \r \n # F i l e : \t > \t \t w > 0000020 i n 3 2 . m a k \r \n # \r \n > # D > 0000040 e s c r i p t i o n : \t \t > p s q > 0000060 l o d b c M a k e f i l > e f > 0000100 o r W i n 3 2 . \r \n # \r \n # > 0000120 C o n f i g u r a t i o n > s : \t > 0000140 D e b u g , R e l e a s e , > 0000160 M u l t i b y t e D e b u g , > 0000200 M u l t i b y t e R e l e > a s e > 0000220 \r \n # B u i l d T y p > e s : > > I see nothing debatable here: what is on the CVS server is a file > containing CR/LF newlines. If you have something else then it's > been through a transfer process that altered the file. > > The question to be discussed is whether that's the way to distribute > the file or not. We can certainly remove the CRs if not. > > BTW, by my count a Unix-newline version of this file would be 11736 > bytes. A 12708-byte file would be broken in some strange new way > (two LFs per newline, maybe?) I don't really have time at the moment to look into this in any detail, but I did an update to my copy of the source (using WinCVS), attempted nmake /f win32.mak to build the ODBC driver and got an error. Editting win32.mak using vi under cygwin showed a ^M (LF i believe) at the end of every line. Using Programmers File Editor (a very good windows editor) shows a blank line between every line. Removing the ^M's cured the problem. I did also get a fresh copy from CVS to ensure that it wasn't just my copy that was affected. Regards, Dave.
Dave Page <dpage@vale-housing.co.uk> writes:
> I did an update to my copy of the source (using WinCVS), attempted nmake /f
> win32.mak to build the ODBC driver and got an error. Editting win32.mak
> using vi under cygwin showed a ^M (LF i believe) at the end of every line.
> Using Programmers File Editor (a very good windows editor) shows a blank
> line between every line. Removing the ^M's cured the problem. I did also get
> a fresh copy from CVS to ensure that it wasn't just my copy that was
> affected.
After further investigation it seems that win32.mak did originally have
Unix-style newlines (\n), and they were all changed to DOS style (\r\n)
in the last commit. I have changed them back, on the assumption that
this conversion was unintentional.
regards, tom lane
Tom Lane wrote: > > Hiroshi Inoue <Inoue@tpf.co.jp> writes: > > Dave Page wrote: > >> Hiroshi, I've attached a fixed copy of win32.mak - the one in CVS seems to > >> be full of ^M's at the end of every line which makes nmake very unhappy. > >> Could you commit it to CVS please if you're happy with it. > > > Hmm there's no difference between my win32.mak and what > > you sent to me. My win32.mak is 12222bytes big. > > How many bytes does the one in cvs have ? > > Right now, src/interfaces/odbc/win32.mak in CVS *is* 12222 bytes > and it *is* full of DOS-style newlines (try od -c if you doubt it). > I have no idea whether that's a problem. Given that no one on a > non-Windoze platform has any use for the file, I'd have guessed that > DOS-style newlines are the way to go ... but if we can get rid of the > CRs without breaking things on Windows then I'm for it ... > Seems Dave's reply hasn't appeared on the list yet. He says "A fresh copy from CVS comes in at 12,708 bytes on my system (and yes, I agree the one I sent is 12,222 bytes)." I see my copy is 12,222 bytes on both linux and cygwin. I also heard from Eiji Tokuya that he sometimes(not always) saw corrupted win32.mak. Unfortunately I don't understand where the cause is. regards, Hiroshi Inoue