Can RPM packages be installed into Cygwin?

2012-04-07 Thread De-Jian Zhao

Dear all,

I noticed that there is a command - rpm - under Cygwin 1.7. Does that 
mean RPM packages can be installed into Cygwin?


I tried to install ncbi-blast-2.2.26+-3.i686.rpm (see: 
ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/ ) into 
Cygwin 1.7.13 with the command "install -i 
ncbi-blast-2.2.26+-3.i686.rpm". However, error message appeared as 
below. I tried to search for the missing libs using the setup.exe of 
Cygwin. It seems that some of them were not present, such as libc.so.6, 
libdl.so.2,  libm.so.6,  libnsl.so.1, and  libz.so.1. Where can I get 
these libs? Thanks!


$ rpm -i ncbi-blast-2.2.26+-3.i686.rpm
error: Failed dependencies:
/usr/bin/perl is needed by ncbi-blast-2.2.26+-3
libbz2.so.1 is needed by ncbi-blast-2.2.26+-3
libc.so.6 is needed by ncbi-blast-2.2.26+-3
libc.so.6(GLIBC_2.0) is needed by ncbi-blast-2.2.26+-3
libc.so.6(GLIBC_2.1) is needed by ncbi-blast-2.2.26+-3
libc.so.6(GLIBC_2.1.2) is needed by ncbi-blast-2.2.26+-3
libc.so.6(GLIBC_2.1.3) is needed by ncbi-blast-2.2.26+-3
libc.so.6(GLIBC_2.2) is needed by ncbi-blast-2.2.26+-3
libc.so.6(GLIBC_2.3) is needed by ncbi-blast-2.2.26+-3
libdl.so.2 is needed by ncbi-blast-2.2.26+-3
libdl.so.2(GLIBC_2.0) is needed by ncbi-blast-2.2.26+-3
libdl.so.2(GLIBC_2.1) is needed by ncbi-blast-2.2.26+-3
libgcc_s.so.1 is needed by ncbi-blast-2.2.26+-3
libgcc_s.so.1(GCC_3.0) is needed by ncbi-blast-2.2.26+-3
libgcc_s.so.1(GLIBC_2.0) is needed by ncbi-blast-2.2.26+-3
libm.so.6 is needed by ncbi-blast-2.2.26+-3
libnsl.so.1 is needed by ncbi-blast-2.2.26+-3
libpthread.so.0 is needed by ncbi-blast-2.2.26+-3
libpthread.so.0(GLIBC_2.0) is needed by ncbi-blast-2.2.26+-3
libpthread.so.0(GLIBC_2.1) is needed by ncbi-blast-2.2.26+-3
libpthread.so.0(GLIBC_2.2) is needed by ncbi-blast-2.2.26+-3
libpthread.so.0(GLIBC_2.3.2) is needed by ncbi-blast-2.2.26+-3
librt.so.1 is needed by ncbi-blast-2.2.26+-3
libstdc++.so.6 is needed by ncbi-blast-2.2.26+-3
libstdc++.so.6(CXXABI_1.3) is needed by ncbi-blast-2.2.26+-3
libstdc++.so.6(GLIBCXX_3.4) is needed by ncbi-blast-2.2.26+-3
libstdc++.so.6(GLIBCXX_3.4.5) is needed by ncbi-blast-2.2.26+-3
libz.so.1 is needed by ncbi-blast-2.2.26+-3
perl(Archive::Tar) is needed by ncbi-blast-2.2.26+-3
perl(Digest::MD5) is needed by ncbi-blast-2.2.26+-3
perl(File::Temp) is needed by ncbi-blast-2.2.26+-3
perl(File::stat) is needed by ncbi-blast-2.2.26+-3
perl(Getopt::Long) is needed by ncbi-blast-2.2.26+-3
perl(Net::FTP) is needed by ncbi-blast-2.2.26+-3
perl(Pod::Usage) is needed by ncbi-blast-2.2.26+-3
perl(constant) is needed by ncbi-blast-2.2.26+-3
perl(strict) is needed by ncbi-blast-2.2.26+-3
perl(warnings) is needed by ncbi-blast-2.2.26+-3


--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Can RPM packages be installed into Cygwin?

2012-04-07 Thread De-Jian Zhao

On 2012-4-7 16:12, marco atzeri wrote:

On 4/7/2012 9:40 AM, De-Jian Zhao wrote:

Dear all,

I noticed that there is a command - rpm - under Cygwin 1.7. Does that
mean RPM packages can be installed into Cygwin?

I tried to install ncbi-blast-2.2.26+-3.i686.rpm (see:
ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/ ) into
Cygwin 1.7.13 with the command "install -i
ncbi-blast-2.2.26+-3.i686.rpm". However, error message appeared as
below. I tried to search for the missing libs using the setup.exe of
Cygwin. It seems that some of them were not present, such as libc.so.6,
libdl.so.2, libm.so.6, libnsl.so.1, and libz.so.1. Where can I get these
libs? Thanks!

$ rpm -i ncbi-blast-2.2.26+-3.i686.rpm
error: Failed dependencies:
/usr/bin/perl is needed by ncbi-blast-2.2.26+-3
libbz2.so.1 is needed by ncbi-blast-2.2.26+-3
libc.so.6 is needed by ncbi-blast-2.2.26+-3

{cut]




...isn't it?

from cygwin website (first page):

Cygwin is not:

-a way to run native Linux apps on Windows. You must rebuild your 
application

 from source if you want it to run on Windows.
-a way to magically make native Windows apps aware of UNIX® 
functionality like
 signals, ptys, etc. Again, you need to build your apps from 
source if you want

 to take advantage of Cygwin functionality.

Regards
Marco
Thanks, Marco. I was not trying to run a native linux apps, and instead 
tried to rebuild the app from source. The BLAST software provides the 
Build.sh and Check.sh scripts under the dirctory of 
ncbi-blast-2.2.26+-src.tar\ncbi-blast-2.2.26+-src\c++\compilers\cygwin\. 
So I think it is possible to run the software under cygwin. However, the 
compilation stopped at "checking for python", even if python was 
installed. That's to say, I installed the latest python and the 
compilation could find it, but the compilation tried to search for 
python2.3 and 2.4 further, and stopped when searching for 2.4.


What I am concern is that whether the i686.rpm indicates it is prepared 
for cygwin. As you know, when you run "uname -smo" , you can get 
something like this "CYGWIN_NT-5.1 i686 Cygwin". I think the package 
ncbi-blast-2.2.26+-3.i686.rpm is for Cygwin. Is that right? Has anyone 
ever succeeded in installing a RPM package into cygwin using "rpm -i *rpm"?


Regards
Dejian

--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Can RPM packages be installed into Cygwin?

2012-04-09 Thread De-Jian Zhao

On 2012-4-8 14:47, marco atzeri wrote:

On 4/8/2012 5:10 AM, De-Jian Zhao wrote:

On 2012-4-7 16:12, marco atzeri wrote:



Thanks, Marco. I was not trying to run a native linux apps, and instead
tried to rebuild the app from source. The BLAST software provides the
Build.sh and Check.sh scripts under the dirctory of
ncbi-blast-2.2.26+-src.tar\ncbi-blast-2.2.26+-src\c++\compilers\cygwin\.
So I think it is possible to run the software under cygwin. However, the
compilation stopped at "checking for python", even if python was
installed. That's to say, I installed the latest python and the
compilation could find it, but the compilation tried to search for
python2.3 and 2.4 further, and stopped when searching for 2.4.

What I am concern is that whether the i686.rpm indicates it is prepared
for cygwin. As you know, when you run "uname -smo" , you can get
something like this "CYGWIN_NT-5.1 i686 Cygwin". I think the package
ncbi-blast-2.2.26+-3.i686.rpm is for Cygwin. Is that right? Has anyone
ever succeeded in installing a RPM package into cygwin using "rpm -i 
*rpm"?


Regards
Dejian



i686.rpm is not the source package, is the binary for linux on 686 
processor.

On cygwin we do not use rpm for building packages so the source rpm
file will unlikely work.

The source for all platforms is

ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/ncbi-blast-2.2.26+-src.tar.gz 





Regards
Marco

Thank you, Marco. I tried that version. Unfortunately, the compilation 
always stopped at "checking python" even after python was installed. It 
seems to me that it is a tough work to rebuild a software from the 
source code under Cygwin. The dependency is hard to satisfy.


I am curious about the utility rpm under Cygwin environment. When 
running the setup.exe, I noticed that there is a rpm packge. (Search for 
"rpm" at the stage of "Select Packages" when running setup.exe, you can 
find the package under "Utils Default".) Then what is it for?


Regards,
Dejian

--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Can RPM packages be installed into Cygwin?

2012-04-09 Thread De-Jian Zhao

On 2012-4-10 8:30, Buchbinder, Barry (NIH/NIAID) [E] wrote:

De-Jian Zhao sent the following at Monday, April 09, 2012 11:12 AM

On 4/8/2012 5:10 AM, De-Jian Zhao wrote:

On 2012-4-7 16:12, marco atzeri wrote:
Thanks, Marco. I was not trying to run a native linux apps, and instead
tried to rebuild the app from source. The BLAST software provides the
Build.sh and Check.sh scripts under the directory of
ncbi-blast-2.2.26+-src.tar\ncbi-blast-2.2.26+-src\c++\compilers\cygwin\.
So I think it is possible to run the software under cygwin. However, the
compilation stopped at "checking for python", even if python was
installed. That's to say, I installed the latest python and the
compilation could find it, but the compilation tried to search for
python2.3 and 2.4 further, and stopped when searching for 2.4.

What I am concern is that whether the i686.rpm indicates it is prepared
for cygwin. As you know, when you run "uname -smo", you can get
something like this "CYGWIN_NT-5.1 i686 Cygwin". I think the package
ncbi-blast-2.2.26+-3.i686.rpm is for Cygwin. Is that right? Has anyone
ever succeeded in installing a RPM package into cygwin using "rpm -i
*rpm"?


i686.rpm is not the source package, is the binary for linux on 686
processor.
On cygwin we do not use rpm for building packages so the source rpm
file will unlikely work.

The source for all platforms is

ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/ncbi-blast-2.2.26+-src.tar.gz


On 2012-4-8 14:47, marco atzeri wrote: Thank you, Marco. I tried that
version. Unfortunately, the compilation always stopped at "checking
python" even after python was installed. It seems to me that it is a
tough work to rebuild a software from the source code under Cygwin. The
dependency is hard to satisfy.

I am curious about the utility rpm under Cygwin environment. When
running the setup.exe, I noticed that there is a rpm package. (Search for
"rpm" at the stage of "Select Packages" when running setup.exe, you can
find the package under "Utils Default".) Then what is it for?

Looking at the blast directory, they have a source rpm.  Perhaps some source
code is only available as an rpm?

For what it is worth, I tried compiling and it seemed to compile OK when I
used "./configure --without-boost" (recommended by make after a simple
"./configure".  "make check" passed two of three tests.  YMMV.  As I did not
install any dependencies, maybe you could do better.

Also, I noted that ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/
contains some windows binaries.  You might consider using them as native
binaries usually run faster than cygwin binaries.

Also, note that there is a folder ncbi-blast-2.2.26+-src/c++/compilers/cygwin.
I didn't notice it until after I tried compiling.  You might try to figure out
whether the scripts there help you compile blast.

IMPORTANT: I'm not a programmer, I don't work at NCBI, I have no experience in
bioinformatics and have certainly never used BLAST.  DO NOT assume that I know
anything about this more than what I reported above.  I was intrigued and did
an experiment - that is all.

Good luck.

- Barry
   Disclaimer: Statements made herein are not made on behalf of NIAID.
   Disclaimer: Statements made herein are not made on behalf of NCBI.
   Disclaimer: Statements made herein are not made on behalf of NLM.
   Disclaimer: Statements made herein are not made on behalf of NIH.
   Disclaimer: Statements made herein are not made on behalf of PHS.
   Disclaimer: Statements made herein are not made on behalf of DHHS.



Thank you, Barry. I tried the file c++/compilers/cygwin/build.sh but 
failed. I wrote to the help desk of NCBI Blast+. They have no guidelines 
for Cygwin users and also suggested the windows version. Actually, I am 
using the windows version now. The only shortcoming is that the windows 
binaries do not recognize the linux-like Cygwin path, so the soft links 
to the blast databases become invalid. I have to change to the directory 
of the Blast database every time I want to run blast search. Anyway, it 
works although a bit inconvenient. Thanks again for your help.


--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Can RPM packages be installed into Cygwin?

2012-04-09 Thread De-Jian Zhao
 It seems that the compilation under Cygwin is more complicated than 
expected.

Thank you for your explanation, Warren.

On 2012-4-10 8:49, Warren Young wrote:

On 4/9/2012 9:12 AM, De-Jian Zhao wrote:

the compilation
always stopped at "checking python" even after python was installed.


My wild guess based on no personal experience with BLAST+ is that 
there is some linkage error here.  I'd have to try it myself (not 
interested) or examine the config.log file (still not interested) to 
figure out if this guess is correct.


I'd bet there are more Cygwin-using BLAST+ users on the BLAST+ mailing 
list than on the Cygwin list, so you'd have a better chance of getting 
a solution by sending the config.log file to that list than continuing 
to pursue this problem here.


Instead of trying to start with RPM, I'd recommend figuring out how to 
build BLAST+ to build from source first.  You need that experience to 
help you build binary packages.


I think you'll have an easier time learning to build a native Cygwin 
package than trying to get the .src.rpm file to rebuild on Cygwin:


http://cygwin.com/setup.html

That page gives several methods for building Cygwin packages.  I'd 
recommend going with the Cygport option.  Cygport does fewer things 
than RPM, so building .cygport files is easier than building RPM .spec 
files.  (In Cygwin, much of the rest of what RPM does lives in 
setup.exe and cygcheck.exe.)


I doubt there's enough value in the existing BLAST+ .spec file to make 
it an easier path for you than building a .cygport file from scratch.



Then what is [the rpm package] for?


It is for people who wish to build a secondary packaging system on top 
of Cygwin's native system.


It doesn't magically port Linux RPM packages over for you.  You still 
have to do that work yourself.  And, you have to do that work without 
having the benefit of an ecosystem of existing RPM binary packages, 
upon which that BLAST+ .spec file probably depends.




--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Can RPM packages be installed into Cygwin?

2012-04-18 Thread De-Jian Zhao

On 2012-4-17 13:58, Linda Walsh wrote:

De-Jian Zhao wrote:


Dear all,

I noticed that there is a command - rpm - under Cygwin 1.7. Does that 
mean RPM packages can be installed into Cygwin?


I tried to install ncbi-blast-2.2.26+-3.i686.rpm (see: 
ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/ ) into 
Cygwin 1.7.13 with the command "install -i 
ncbi-blast-2.2.26+-3.i686.rpm". However, error message appeared as 
below. libc.so.6, libdl.so.2,  libm.so.6,  libnsl.so.1, and  
libz.so.1. Where can I get these libs? Thanks!


$ rpm -i ncbi-blast-2.2.26+-3.i686.rpm
error: Failed dependencies:
/usr/bin/perl is needed by ncbi-blast-2.2.26+-3
libbz2.so.1 is needed by ncbi-blast-2.2.26+-3
libc.so.6 is needed by ncbi-blast-2.2.26+-3
libc.so.6(GLIBC_2.1.3) is needed by ncbi-blast-2.2.26+-3
libc.so.6(GLIBC_2.2) is needed by ncbi-blast-2.2.26+-3
libc.so.6(GLIBC_2.3) is needed by ncbi-blast-2.2.26+-3


.
On Tue, Apr 10, 2012 at 02:03:29PM +0800, De-Jian Zhao wrote:

> It seems that the compilation under Cygwin is more complicated than
> expected.
==

This has nothing to do with cygwin.

It is 'rpm' looking for installed dependencies in the rpm database.

Since everything was installed with 'setup.exe' which keeps it's own
'database', rpm thinks nothing is installed.

You need to make sure the versions required by the rpm are installed 
by setup,

then you can build from a ncbi-blast-2.2.26+-3.src.rpm
   ^^^

The i686.rpm package is prebuilt for "some" platform (you would have 
to know
which from the download -- i.e. one built for suse won't necessarily 
run on redhat, or ubuntu...  or other combinations.   Ones that do 
will be statically
built and will be usable on some range of OS-versions of some 
particular OS.
Since rpm is a linux thing, it would be built for the linux OS... so 
double-screw.  Wrong OS, wrong executable format...etc...   look for a 
corresponding .src. rpm, OR as someone else already suggested, look 
for the
'tar' that the is included in the ".src".  -- in fact, if you have the 
.src.,
usually the .tar. file they build from is packed into it (then patches 
and OS
specific stuff is applied), so you can often just take the tar from 
your src.rpm
and move it to cygwin, and run it directly if it has a "configure" 
script.


it might also have something like "bootstrap.sh" (or bootstrap 
something) that will produce a configure script -- that will usually 
allow you to build an executable that will work on your platform (in 
this case cygwin) -- but you
will still have to have _needed_ dependencies installed (like 
development packages from cygwin for some of the above).


All of those packages may not be needed -- dependencies are often based
on what options you select in configure (configure --help tells you 
all it's options).  Sometimes you don't need all the options the 
'default' rpm builds

with making life simpler.

Compilation, period, other than on the system the package was designed 
for is
never completely trivial, but configure makes it nearly so if it is 
used --
most packages these days have such a script -- rpm packages just have 
pre-built
some pre-configured image for distribution -- may not be optimal for 
your needs

nor is it likely to be optimized for your machine.

Hope this was of some help.
Good luck!

Linda




Thank you, Linda. When I saw the error message after installing some 
packages to satisfy the dependencies, I was wondering why the rpm 
package still complained about the already installed packages. I was not 
aware that cygwin and rpm were not sharing the 'database'. Thank you for 
making it clear to me.I will try the src.rpm when I am free.


DJ

--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Can RPM packages be installed into Cygwin?

2012-04-18 Thread De-Jian Zhao

On 2012-4-18 22:03, Christopher Faylor wrote:

On Wed, Apr 18, 2012 at 09:13:20PM +0800, De-Jian Zhao wrote:

On 2012-4-17 13:58, Linda Walsh wrote:

[snip]

Thank you, Linda.  When I saw the error message after installing some
packages to satisfy the dependencies, I was wondering why the rpm
package still complained about the already installed packages.  I was
not aware that cygwin and rpm were not sharing the 'database'.  Thank
you for making it clear to me.I will try the src.rpm when I am free.

So with all of the back and forth about this, you never realized that
Cygwin's setup.exe doesn't use RPM?  Wow.  Wasn't the fact that there
are no .rpm files in the distribution a big enough clue?

cgf


I never expect setup.exe will use rpm packages. There is a rpm command 
in Cygwin. I just thought it could be used to install a rpm package.


The rpm package complained about dependencies during installation. I did 
think that the dependency could be solved by installing proper packages 
through setup.exe. But after installing some libraries through 
setup.exe, the rpm package still complained the same as before. It 
seemed that the rpm package did not see the installed libraries at all.


--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Can RPM packages be installed into Cygwin?

2012-04-18 Thread De-Jian Zhao

On 2012-4-19 0:58, Linda Walsh wrote:

De-Jian Zhao wrote:



Thank you, Linda. When I saw the error message after installing some 
packages to satisfy the dependencies, I was wondering why the rpm 
package still complained about the already installed packages. I was 
not aware that cygwin and rpm were not sharing the 'database'. Thank 
you for making it clear to me.I will try the src.rpm when I am free.


---
Note:
The src.rpm will likely look in the rpm database for it's
pre-reqs as well.
you will need to look inside it (install the src.rpm, and look
in the /usr/src/packages/specs dir for .spec>) -- and you 
will see lines like: (example from suse samba package):


BuildRequires:  cups-devel e2fsprogs-devel fam-devel libacl-devel 
pam-devel popt-devel python-devel readline-devel

%if 0%{?suse_version}
BuildRequires:  cracklib-devel openldap2-devel
%endif
%if 0%{?suse_version} > 1030
BuildRequires:  gdbm-devel pwdutils
%endif
%if 0%{?suse_version} > 930
BuildRequires:  krb5-devel libiniparser-devel libnscd-devel
%endif

(there's a bunch more .. the BuildRequires lines -- will refer to
DISTRO-specific package names (if that was built for Redhat, will refer
to redhat's package names).

The "if" statements are all about configure the same "spec" file for 
different
conditions... if you wanted to make things complex (I wouldn't), you 
could
build all the packages it needs and put in an if statement that would 
be true

for cygwin (would be alot of work!)

Or. look at the minimum set of packages you will need and make sure they
are installed from cygwin,
and comment out all those build-require lines and -- try it.. cross 
fingers.


But it will then likely be built for whatever DISTRO's specific file 
layout

which may not match cygwin's.

Safest -- after you install src.rpm, look in 
/usr/src/packages/SOURCES/ for
.tar.[bz2,gz] -- that will be the original tar 
package

-- prior to any DISTRO specific changes, that you will want.

Unpack that to a dir and hope it is setup with the configure route -- 
or figure

out what it needs to build (make...blah).

I have taken suse src rpms and direct ported them to cygwin -- but I had
to disable all the pre-requisite checking at build and install time, 
due to
the missing rpm database -- but the program was a *simple one*, and it 
worked.
So it is possible.   The more complex the program, the more likely it 
will
take more work to port...but sometimes you get lucky and the developer 
did

a good job setting up a configure script that will work with cygwin...

It is a pain -- you will be porting the source to cygwin -- but it it 
is worth it to you, you can get it done.  (But you really have to want 
it!) ;-)




Hi, Linda. Thank you for your detailed explanation. The "spec" file 
seems different from your description. There is no explicit 
BuildRequires statement. There is no explicit statement of packages 
required. It is as follows:


Star ncbi-blast.spec===

Name:ncbi-blast
Version: 2.2.26+
Release: 3
Source0: %{name}-%{version}.tgz
Summary: NCBI BLAST finds regions of similarity between biological 
sequences.

Exclusiveos: linux
Group:   NCBI/BLAST
License: Public Domain
BuildArch:   i686 x86_64
BuildRoot:   /var/tmp/%{name}-buildroot
Prefix:  /usr

%description
The NCBI Basic Local Alignment Search Tool (BLAST) finds regions of
local similarity between sequences. The program compares nucleotide or
protein sequences to sequence databases and calculates the statistical
significance of matches. BLAST can be used to infer functional and
evolutionary relationships between sequences as well as help identify
members of gene families.

%prep
%setup -q

%build
./configure
cd c++/*/build
%__make -f Makefile.flat

%install
%__mkdir_p $RPM_BUILD_ROOT/%_bindir
cd c++/*/bin
%__install -m755 blastp blastn blastx tblastn tblastx psiblast rpsblast 
rpstblastn blast_formatter deltablast makembindex segmasker dustmasker 
windowmasker makeblastdb makeprofiledb blastdbcmd blastdb_aliastool 
convert2blastmask blastdbcheck legacy_blast.pl update_blastdb.pl 
$RPM_BUILD_ROOT/%_bindir


%clean
rm -rf $RPM_BUILD_ROOT

%files
%defattr(-,root,root)
%_bindir/*

%changelog
* Mon Jul 21 2008 Christiam Camacho 
- See ChangeLog file

===End ncbi-blast.spec===



There is a file ncbi-blast-2.2.26+.tgz under the directory SOURCES. 
There are two files, or build.sh and check.sh, in the directory 
ncbi-blast-2.2.26+\c++\compilers\cygwin\, which is similar to the 
package ncbi-blast-2.2.26+-src.tar.gz. I was stuck with "checking for 
python" when compiling ncbi-blast-2.2.26+-src.tar.gz.



--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Can RPM packages be installed into Cygwin?

2012-04-18 Thread De-Jian Zhao

On 2012-4-19 1:08, Christopher Faylor wrote:

On Thu, Apr 19, 2012 at 12:05:38AM +0800, De-Jian Zhao wrote:

On 2012-4-18 22:03, Christopher Faylor wrote:

On Wed, Apr 18, 2012 at 09:13:20PM +0800, De-Jian Zhao wrote:

On 2012-4-17 13:58, Linda Walsh wrote:

[snip]

Thank you, Linda.  When I saw the error message after installing some
packages to satisfy the dependencies, I was wondering why the rpm
package still complained about the already installed packages.  I was
not aware that cygwin and rpm were not sharing the 'database'.  Thank
you for making it clear to me.I will try the src.rpm when I am free.

So with all of the back and forth about this, you never realized that
Cygwin's setup.exe doesn't use RPM?  Wow.  Wasn't the fact that there
are no .rpm files in the distribution a big enough clue?

I never expect setup.exe will use rpm packages. There is a rpm command
in Cygwin. I just thought it could be used to install a rpm package.

And, although it is a very old, basically-unmaintained package, it can
be used to install rpm packages.


The rpm package complained about dependencies during installation. I did
think that the dependency could be solved by installing proper packages
through setup.exe.
But after installing some libraries through setup.exe, the rpm package
still complained the same as before.  It seemed that the rpm package
did not see the installed libraries at all.

I don't know if you are getting this or not but, yes, you can install
cygwin packages which would allow you to satisfy the needs for building
an rpm package.  That doesn't mean that rpm would magically know about
it since rpm maintains its own package database which obviously is not
updated by setup.exe.  So, while the programs you need to build an
rpm package may be available, rpm does not know this.  The fact that
we use setup.exe and "cygcheck" to query our packages should be a
huge clue that setup.exe does not know anything about rpm.

The --nodeps option to rpm will silence the dependency warnings but
it won't fix real problems if required packages are missing.

rpm is basically a curiosity on Cygwin.  It really isn't that useful.
If you want to build Cygwin packages, you should be investigating
cygport.




Thank you, Christopher. Is there a way to export the package info or lib 
info of setup.exe to rpm? If they can share information, the problem 
will be easily solved.


Besides, I tried --nodeps option with src.rpm. It requires root 
privilege (see below). However, "sudo rpm" and "su root" both failed to 
work. Any way to fix it? Does Cygwin has a root user?

Administrator@acer /cygdrive/d/Blast/source
$ rpm --nodeps --install ncbi-blast-2.2.26+-3.src.rpm
warning: user coremake does not exist - using root
warning: group cppcore does not exist - using root
warning: user coremake does not exist - using root
warning: group cppcore does not exist - using root

As to cygport, it seems that a *.cygport file is required. How to get 
the file?

$ cygport ncbi-blast-2.2.26+-3.src.rpm install
*** ERROR: ncbi-blast-2.2.26+-3.src.rpm.cygport not found.
Administrator@acer /cygdrive/d/Blast/source
$ cygport cbi-blast-2.2.26+-src.tar.gz install
*** ERROR: cbi-blast-2.2.26+-src.tar.gz.cygport not found.

DJ

--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Can RPM packages be installed into Cygwin?

2012-04-19 Thread De-Jian Zhao

On 2012-4-19 7:09, Buchbinder, Barry (NIH/NIAID) [E] wrote:

De-Jian,

Hello.  It's me again.

I suspect that just writing wrapper scripts would be faster than playing
with RPMs and/or trying to compile blast yourself.

As I understand it, blast programs are all command-line.  I think
that the easiest thing would be to just use the Windows version
supplied by NCBI and write wrapper scripts that will convert cygwin's
POSIX paths to Windows paths.  You might have to experiment to see if
you also need to convert file formats from \n to \r\n line endings.

See
   $ man cygpath
   $ man dos2unix
   $ man unix2dos

Luc Hermitte's cyg-wrapper.sh might be all you need.
   http://hermitte.free.fr/cygwin/#Win32
Example (using vim):
   http://vim.wikia.com/wiki/Running_the_win32-version_of_Vim_from_cygwin
I've never used cyg-wrapper.sh, so this suggestion is "theoretical" and
not based on my personal experience.

Related:
   http://www.ncbi.nlm.nih.gov/books/NBK1762/
   http://www.ncbi.nlm.nih.gov/books/NBK52637/

Good luck,

- Barry
   Disclaimer: Statements made herein are not made on behalf of NIAID.
   Disclaimer: Please see http://cygwin.com/ml/cygwin/2012-04/msg00247.html
   for other disclaimers and statements concerning the limited
   reliability of my advice on this topic.



Hi,Barry. Thank you for your suggestion. Actually I am using the windows 
version of BLAST these days. The only inconvenience is that windows 
version of blast does not recognize the linux-like paths, as mentioned 
in my previous replies.


Now, it is solved after reading your reply. I was inspired by the notion 
of "convert cygwin's POSIX paths to Windows paths". I tried to convert 
the path into windows path using "cygpath -w" and it works smoothly. An 
example command line is as below:


blastall -p blastn -i nt.fas -d $(cygpath -w genome.db) -o test.out -m 8

I have modified the PATH variable so as to the blast programs could be 
automatically identified. genome.db is a soft link to the target 
database under the working directory. blastall can not interpret the 
soft link; but after translating the path using $(cygpath -w genome.db), 
it works perfectly.


Furthermore, there is in fact no need to create the linux-style soft 
link. I added a variable DB to store the windows-style path of the 
database and this makes a more simple solution.


DB="D:\Blast\blast-2.2.13\db\genome\genome.db"
blastall -p blastn -i nt.fas -d $DB -o test.out -m 8

Up to now, the problem is solved. Thanks to all that participated in the 
discussion. I should have make it clear that I just want to use the 
blast program under Cygwin. The efforts were diverted into two 
directions. One was to compile the program from source code. Many people 
contributed their ideas and the discussion deepened my understanding of 
Cygwin. The other was to use the windows version. The problem of 
different styles of paths was settled today. Thanks to all for your help.


Best regards,
DJ


--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Why /usr/bin/*.dll must be executable?

2012-04-20 Thread De-Jian Zhao

Hi,

When I type "cyg" and Tab, many executables starting with "cyg" are 
listed (Display all 262 possibilities? (y or n) y). I find that many of 
them are *.dll libraries under /usr/bin/. This is inconvenient to find 
the real executable applications (*.exe). Since *.dll files are only 
libraries, they are not necessary to have the attribute of "x". Thus, I 
run the command "chmod a-x /usr/bin/*.dll". Unexpectedly, cygwin is 
corrupted. I closed the terminal and failed to restart Cygwin. I started 
my  older version of Cygwin (I did not deleted it after installing a new 
version), and added "x" to the previous *.dll files. The dead Cygwin 
revived.


I am confused why /usr/bin/*.dll should be executable. I thought they 
were only library files. When I tried to run a dll file, bash says 
"cannot execute binary file". Are there some hidden stories?


The error messages are as follows:

Administrator@acer /usr/bin
$ chmod a-x *.dll

Administrator@acer /usr/bin
$ lh
 18 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: retry: Resource temporarily unavailable
2020930 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: retry: Resource temporarily unavailable
4039235 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: retry: Resource temporarily unavailable
8057832 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: retry: Resource temporarily unavailable
16074662 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: Resource temporarily unavailable


DJ

--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Why /usr/bin/*.dll must be executable?

2012-04-20 Thread De-Jian Zhao

On 2012-4-20 21:02, De-Jian Zhao wrote:

Hi,

When I type "cyg" and Tab, many executables starting with "cyg" are 
listed (Display all 262 possibilities? (y or n) y). I find that many 
of them are *.dll libraries under /usr/bin/. This is inconvenient to 
find the real executable applications (*.exe). Since *.dll files are 
only libraries, they are not necessary to have the attribute of "x". 
Thus, I run the command "chmod a-x /usr/bin/*.dll". Unexpectedly, 
cygwin is corrupted. I closed the terminal and failed to restart 
Cygwin. I started my  older version of Cygwin (I did not deleted it 
after installing a new version), and added "x" to the previous *.dll 
files. The dead Cygwin revived.


I am confused why /usr/bin/*.dll should be executable. I thought they 
were only library files. When I tried to run a dll file, bash says 
"cannot execute binary file". Are there some hidden stories?


The error messages are as follows:

Administrator@acer /usr/bin
$ chmod a-x *.dll

Administrator@acer /usr/bin
$ lh
 18 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: retry: Resource temporarily unavailable
2020930 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: retry: Resource temporarily unavailable
4039235 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: retry: Resource temporarily unavailable
8057832 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: retry: Resource temporarily unavailable
16074662 [main] -bash 4512 fork: child -1 - forked process died 
unexpectedly, retry 0, exit code -1073741790, errno 11

-bash: fork: Resource temporarily unavailable


DJ



lh is an alias to 'ls -lh'.

--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Why /usr/bin/*.dll must be executable?

2012-04-20 Thread De-Jian Zhao

On 2012-4-20 21:07, Václav Zeman wrote:

On 20 April 2012 15:02, De-Jian Zhao wrote:

Hi,

When I type "cyg" and Tab, many executables starting with "cyg" are listed
(Display all 262 possibilities? (y or n) y). I find that many of them are
*.dll libraries under /usr/bin/. This is inconvenient to find the real
executable applications (*.exe). Since *.dll files are only libraries, they
are not necessary to have the attribute of "x". Thus, I run the command
"chmod a-x /usr/bin/*.dll". Unexpectedly, cygwin is corrupted. I closed the
terminal and failed to restart Cygwin. I started my  older version of Cygwin
(I did not deleted it after installing a new version), and added "x" to the
previous *.dll files. The dead Cygwin revived.

I am confused why /usr/bin/*.dll should be executable. I thought they were
only library files. When I tried to run a dll file, bash says "cannot
execute binary file". Are there some hidden stories?

DLLs are executables thus they need the +x bit. This is a Windows thing.



If this is a windows thing, removing the x bit should not affect Cygwin. 
Instead, Cygwin is corrupted after removing the x bit.


--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Why /usr/bin/*.dll must be executable?

2012-04-20 Thread De-Jian Zhao

On 2012-4-21 0:27, Corinna Vinschen wrote:

On Apr 21 00:17, De-Jian Zhao wrote:

On 2012-4-20 21:07, Václav Zeman wrote:

On 20 April 2012 15:02, De-Jian Zhao wrote:

Hi,

When I type "cyg" and Tab, many executables starting with "cyg" are listed
(Display all 262 possibilities? (y or n) y). I find that many of them are
*.dll libraries under /usr/bin/. This is inconvenient to find the real
executable applications (*.exe). Since *.dll files are only libraries, they
are not necessary to have the attribute of "x". Thus, I run the command
"chmod a-x /usr/bin/*.dll". Unexpectedly, cygwin is corrupted. I closed the
terminal and failed to restart Cygwin. I started my  older version of Cygwin
(I did not deleted it after installing a new version), and added "x" to the
previous *.dll files. The dead Cygwin revived.

I am confused why /usr/bin/*.dll should be executable. I thought they were
only library files. When I tried to run a dll file, bash says "cannot
execute binary file". Are there some hidden stories?

DLLs are executables thus they need the +x bit. This is a Windows thing.


If this is a windows thing, removing the x bit should not affect
Cygwin. Instead, Cygwin is corrupted after removing the x bit.

Windows requires the x bit for DLLs to be loadable as executable code
into the address space of a process.  As Václav wrote, it's a Windows
thing.


Can Windows see the rwx bits assigned by Cygwin to the files? I tried 
removing the x bit of an executable file blastall.exe (chmod a-x 
blastall.exe); the file can not be executed under Cygwin, but still can 
be executed under cmd console of Windows. It seems that Windows does not 
honor the rwx bits assigned by Cygwin.


Is there a detailed description of the starting process of Cygwin 
system? Or, how does Cygwin interact with *.dll files?


After starting Cygwin, I ran "chmod a-x /usr/bin/cygperl5_10.dll" and 
perl could not be started any more. This could be cured by ran "chmod 
a+x /usr/bin/cygperl5_10.dll". It seems to me that the Cygwin binary 
executables will communicate with their corresponding *.dll files when 
executed. The *.dll provide the required functions and subroutines and 
that's enough. There is no need to mark *.dll with an x bit.


I have no strong background of computer science. Maybe there is some 
basic knowledge beyond my imagination. Hope you can help disclose it to 
me. Thanks.



--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Why /usr/bin/*.dll must be executable?

2012-04-20 Thread De-Jian Zhao

On 2012-4-21 1:29, Larry Hall (Cygwin) wrote:

On 4/20/2012 1:25 PM, Nellis, Kenneth wrote:

From: Corinna Vinschen

Windows requires the x bit for DLLs to be loadable as executable code
into the address space of a process.  As Václav wrote, it's a Windows
thing.


So, I had to play with "chmod -x cyg*.dll" and, of course, it
totally hosed Cygwin. I recovered by reinstalling from scratch
only after realizing that ATTRIB did not have an option to re-
establish X--we're talking XP here.  So, as this is a Windows
thing, wondering how, through Windows, I could restore execute
access to the DLLs.


What's wrong with "chmod +x cyg*.dll"?



It doesn't work any more. You cannot restart a new terminal, either. You 
will be trapped here.



--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Why /usr/bin/*.dll must be executable?

2012-04-20 Thread De-Jian Zhao

On 2012-4-21 1:40, Larry Hall (Cygwin) wrote:

On 4/20/2012 1:32 PM, De-Jian Zhao wrote:

On 2012-4-21 0:27, Corinna Vinschen wrote:

On Apr 21 00:17, De-Jian Zhao wrote:

On 2012-4-20 21:07, Václav Zeman wrote:

On 20 April 2012 15:02, De-Jian Zhao wrote:

Hi,

When I type "cyg" and Tab, many executables starting with "cyg" 
are listed
(Display all 262 possibilities? (y or n) y). I find that many of 
them are
*.dll libraries under /usr/bin/. This is inconvenient to find the 
real
executable applications (*.exe). Since *.dll files are only 
libraries,

they
are not necessary to have the attribute of "x". Thus, I run the 
command
"chmod a-x /usr/bin/*.dll". Unexpectedly, cygwin is corrupted. I 
closed

the
terminal and failed to restart Cygwin. I started my older version of
Cygwin
(I did not deleted it after installing a new version), and added 
"x" to

the
previous *.dll files. The dead Cygwin revived.

I am confused why /usr/bin/*.dll should be executable. I thought 
they were
only library files. When I tried to run a dll file, bash says 
"cannot

execute binary file". Are there some hidden stories?
DLLs are executables thus they need the +x bit. This is a Windows 
thing.



If this is a windows thing, removing the x bit should not affect
Cygwin. Instead, Cygwin is corrupted after removing the x bit.

Windows requires the x bit for DLLs to be loadable as executable code
into the address space of a process. As Václav wrote, it's a Windows
thing.


Can Windows see the rwx bits assigned by Cygwin to the files? I tried
removing the x bit of an executable file blastall.exe (chmod a-x
blastall.exe); the file can not be executed under Cygwin, but still 
can be
executed under cmd console of Windows. It seems that Windows does not 
honor

the rwx bits assigned by Cygwin.

Is there a detailed description of the starting process of Cygwin 
system?

Or, how does Cygwin interact with *.dll files?

After starting Cygwin, I ran "chmod a-x /usr/bin/cygperl5_10.dll" and 
perl

could not be started any more. This could be cured by ran "chmod a+x
/usr/bin/cygperl5_10.dll". It seems to me that the Cygwin binary 
executables

will communicate with their corresponding *.dll files when executed. The
*.dll provide the required functions and subroutines and that's enough.
There is no need to mark *.dll with an x bit.

I have no strong background of computer science. Maybe there is some 
basic
knowledge beyond my imagination. Hope you can help disclose it to me. 
Thanks.


Cygwin DLLs are no different than native Windows DLLs when it comes to
loading them.  The Windows loader is responsible for loading them.  So
we're stuck with whatever restrictions the Windows loader imposes on
Windows apps.



Yes. But does Windows loader can see the x bit assigned by Cygwin, a 
quite different OS?


At least, *.exe does not respect or even is blind to the -x operation of 
Cygwin, as abovementioned.



--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Why /usr/bin/*.dll must be executable?

2012-04-20 Thread De-Jian Zhao

On 2012-4-21 1:57, Larry Hall (Cygwin) wrote:

On 4/20/2012 1:47 PM, De-Jian Zhao wrote:

On 2012-4-21 1:40, Larry Hall (Cygwin) wrote:

On 4/20/2012 1:32 PM, De-Jian Zhao wrote:

On 2012-4-21 0:27, Corinna Vinschen wrote:

On Apr 21 00:17, De-Jian Zhao wrote:

On 2012-4-20 21:07, Václav Zeman wrote:

On 20 April 2012 15:02, De-Jian Zhao wrote:

Hi,

When I type "cyg" and Tab, many executables starting with "cyg" 
are

listed
(Display all 262 possibilities? (y or n) y). I find that many 
of them

are
*.dll libraries under /usr/bin/. This is inconvenient to find 
the real
executable applications (*.exe). Since *.dll files are only 
libraries,

they
are not necessary to have the attribute of "x". Thus, I run the 
command
"chmod a-x /usr/bin/*.dll". Unexpectedly, cygwin is corrupted. 
I closed

the
terminal and failed to restart Cygwin. I started my older 
version of

Cygwin
(I did not deleted it after installing a new version), and 
added "x" to

the
previous *.dll files. The dead Cygwin revived.

I am confused why /usr/bin/*.dll should be executable. I 
thought they

were
only library files. When I tried to run a dll file, bash says 
"cannot

execute binary file". Are there some hidden stories?
DLLs are executables thus they need the +x bit. This is a 
Windows thing.



If this is a windows thing, removing the x bit should not affect
Cygwin. Instead, Cygwin is corrupted after removing the x bit.

Windows requires the x bit for DLLs to be loadable as executable code
into the address space of a process. As Václav wrote, it's a Windows
thing.


Can Windows see the rwx bits assigned by Cygwin to the files? I tried
removing the x bit of an executable file blastall.exe (chmod a-x
blastall.exe); the file can not be executed under Cygwin, but still 
can be
executed under cmd console of Windows. It seems that Windows does 
not honor

the rwx bits assigned by Cygwin.

Is there a detailed description of the starting process of Cygwin 
system?

Or, how does Cygwin interact with *.dll files?

After starting Cygwin, I ran "chmod a-x /usr/bin/cygperl5_10.dll" 
and perl

could not be started any more. This could be cured by ran "chmod a+x
/usr/bin/cygperl5_10.dll". It seems to me that the Cygwin binary 
executables
will communicate with their corresponding *.dll files when 
executed. The
*.dll provide the required functions and subroutines and that's 
enough.

There is no need to mark *.dll with an x bit.

I have no strong background of computer science. Maybe there is 
some basic

knowledge beyond my imagination. Hope you can help disclose it to me.
Thanks.


Cygwin DLLs are no different than native Windows DLLs when it comes to
loading them. The Windows loader is responsible for loading them. So
we're stuck with whatever restrictions the Windows loader imposes on
Windows apps.



Yes. But does Windows loader can see the x bit assigned by Cygwin, a 
quite

different OS?


Cygwin is not an O/S.  The basis for Cygwin is the emulation _DLL_, 
cygwin1.dll.  Cygwin executables and DLLs are Windows binaries, 
subject to whatever limitations the Windows loader has.  That's it.  
If you're not happy about that, you should complain to MS about it.



Then, what limitations does the Windows loader have?
As far as I know, there is no x bits for windows, right? How does the 
"chmod a-x" affect the file behavior under Windows if this is a Windows 
thing? I am completely confused. When I ran "chmod a-x *.dll", I assumed 
that these dll files were not executed directly under Cygwin or Windows; 
they were just some files storing the required functions or something; 
removing the x bits should not affect the behavior of Cygwin and 
Windows. If have run "chmod a-x" on the Windows system dll files, that 
will lead to the corruption of Windows, right? Thank God. I didn't do that.




--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple



Re: Why /usr/bin/*.dll must be executable?

2012-04-20 Thread De-Jian Zhao

On 2012-4-21 3:06, Jon TURNEY wrote:

On 20/04/2012 14:02, De-Jian Zhao wrote:

When I type "cyg" and Tab, many executables starting with "cyg" are listed
(Display all 262 possibilities? (y or n) y). I find that many of them are
*.dll libraries under /usr/bin/. This is inconvenient to find the real
executable applications (*.exe).

'export EXECIGNORE=*.dll' is a cygwin extension to bash to tell it you don't
want to see those pesky dlls.

See /usr/share/doc/Cygwin/bash.README


Cool! That's just what I need! Thanks, Jon.


--
Problem reports:   http://cygwin.com/problems.html
FAQ:   http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info:  http://cygwin.com/ml/#unsubscribe-simple