Kuang He
Hi all,
Attached is a patch (against v101.2) to make SAC use a pager specified
by environment variable $PAGER, instead of the current home-grown tool
similar to ``more'', to show the help files. The good thing is that in
this way, people get to use their favorite pager when viewing the help
files. I personally use ``less'', and its biggest advantage over
``more'' is that it can support scrolling backwards using hotkeys like
Up Arrow, PageDown, Ctrl+U, k, etc. Plus, I get to scroll either way
by a few lines or by half a page, i.e. not necessarily by a whole
page.
Known caveat:
- When the user have an unreasonable setting of $PAGER, the error
message of not being able to load the pager might be a bit mysterious
to the users.
$ export PAGER=xxx
$ sac
SAC> h fg
sh: xxx: not found
- I did not check the return value of malloc() in my code, since I was
hoping someone could come up with a uniform way of reporting malloc()
errors so that we could all stick to it. One possible way is to have a
warpper function to malloc(), and use that instead of malloc() all the
time. And yes, I am aware that there are SO MANY malloc()'s in the
code, whose return values have not been checked!
void *
xmalloc(size_t size)
{
void *p;
if ((p = malloc(size)) == NULL)
fprintf(stderr, "Malloc failed: %s", strerror(errno));
return (p);
}
Best regards,
--
Kuang He
Department of Physics
University of Connecticut
Storrs, CT 06269-3046
Tel: +1.860.486.4919
Web: http://www.phys.uconn.edu/~he/
Attached is a patch (against v101.2) to make SAC use a pager specified
by environment variable $PAGER, instead of the current home-grown tool
similar to ``more'', to show the help files. The good thing is that in
this way, people get to use their favorite pager when viewing the help
files. I personally use ``less'', and its biggest advantage over
``more'' is that it can support scrolling backwards using hotkeys like
Up Arrow, PageDown, Ctrl+U, k, etc. Plus, I get to scroll either way
by a few lines or by half a page, i.e. not necessarily by a whole
page.
Known caveat:
- When the user have an unreasonable setting of $PAGER, the error
message of not being able to load the pager might be a bit mysterious
to the users.
$ export PAGER=xxx
$ sac
SAC> h fg
sh: xxx: not found
- I did not check the return value of malloc() in my code, since I was
hoping someone could come up with a uniform way of reporting malloc()
errors so that we could all stick to it. One possible way is to have a
warpper function to malloc(), and use that instead of malloc() all the
time. And yes, I am aware that there are SO MANY malloc()'s in the
code, whose return values have not been checked!
void *
xmalloc(size_t size)
{
void *p;
if ((p = malloc(size)) == NULL)
fprintf(stderr, "Malloc failed: %s", strerror(errno));
return (p);
}
Best regards,
--
Kuang He
Department of Physics
University of Connecticut
Storrs, CT 06269-3046
Tel: +1.860.486.4919
Web: http://www.phys.uconn.edu/~he/
Attachments
-
Kuang He
Looking over the patches I have a couple of suggestions/comments.
Could we use ANSI C syntax rather than K&R for function declarations.
This would allow for the easy creation of header files by copying and
it is a current/standard notation.
Pager use: We need to check the system for the less command on
compilation. In fact we could check for less, then more, then
fallback to the original code in sac. A change like this might allow
for syntax highlighting in the help files. This would mean the
current paging code needs to be made into a seperate function. The
pager value could always be checked for to override the default.
Cheers
Brian
On Sep 14, 2008, at 3:41 AM, "Kuang He" <icrazy<at>gmail.com> wrote:
Hi all,
Attached is a patch (against v101.2) to make SAC use a pager specified
by environment variable $PAGER, instead of the current home-grown tool
similar to ``more'', to show the help files. The good thing is that in
this way, people get to use their favorite pager when viewing the help
files. I personally use ``less'', and its biggest advantage over
``more'' is that it can support scrolling backwards using hotkeys like
Up Arrow, PageDown, Ctrl+U, k, etc. Plus, I get to scroll either way
by a few lines or by half a page, i.e. not necessarily by a whole
page.
Known caveat:
- When the user have an unreasonable setting of $PAGER, the error
message of not being able to load the pager might be a bit mysterious
to the users.
$ export PAGER=xxx
$ sac
SAC> h fg
sh: xxx: not found
- I did not check the return value of malloc() in my code, since I was
hoping someone could come up with a uniform way of reporting malloc()
errors so that we could all stick to it. One possible way is to have a
warpper function to malloc(), and use that instead of malloc() all the
time. And yes, I am aware that there are SO MANY malloc()'s in the
code, whose return values have not been checked!
void *
xmalloc(size_t size)
{
void *p;
if ((p = malloc(size)) == NULL)
fprintf(stderr, "Malloc failed: %s", strerror(errno));
return (p);
}
Best regards,
--
Kuang He
Department of Physics
University of Connecticut
Storrs, CT 06269-3046
Tel: +1.860.486.4919
Web: http://www.phys.uconn.edu/~he/
<pager-help-file.diff>
_______________________________________________
sac-dev mailing list
sac-dev<at>iris.washington.edu
http://www.iris.washington.edu/mailman/listinfo/sac-dev
-
Dear Brian,
On Sun, Sep 14, 2008 at 12:04 PM, Brian Savage <savage<at>uri.edu> wrote:
Kuang He
Sure, I actually prefer ANSI C syntax myself. I used K&R style
Looking over the patches I have a couple of suggestions/comments.
Could we use ANSI C syntax rather than K&R for function declarations. This
would allow for the easy creation of header files by copying and it is a
current/standard notation.
function declarations just to conform to the existing code base. I'll
use ANSI C style from now on.
Pager use: We need to check the system for the less command on compilation.
This is a good idea, but I'm not sure how to do it, so could you help
In fact we could check for less, then more, then fallback to the original
code in sac.
me with this?
A change like this might allow for syntax highlighting in the help files.
Exactly, and we get to search within help files.
This would mean the current paging code needs to be made into a
I can try isolating the current paging code to a separate function.
seperate function. The pager value could always be checked for to override
the default.
Where do you think I should put it, in src/ucf/wrhelp.c?
Best regards,
--
Kuang He
Department of Physics
University of Connecticut
Storrs, CT 06269-3046
Tel: +1.860.486.4919
Web: http://www.phys.uconn.edu/~he/
-
Kuang He,
ANSI C it is. Unless someone else has any objections.
To add in a check for "less" and or "more" see:
Autoconf: Generic Program Checks
http://www.gnu.org/software/autoconf/manual/autoconf-2.57/
html_chapter/autoconf_5.html#SEC41
Either one of these should work. You can give a list of programs to
check for in the order you would like
Macro: AC_CHECK_PROGS (variable, progs-to-check-for, [value-if-not-
found], [path])
Macro: AC_PATH_PROGS (variable, progs-to-check-for, [value-if-not-
found], [path])
This should go into the configure.in file, then the bootstrap needs
to be re-run.
Isolating the original pager code inside src/sac/wrhelp.c is probably
the best place for it.
Cheers
Brian
On Sep 14, 2008, at 12:32 PM, Kuang He wrote:
Dear Brian,
On Sun, Sep 14, 2008 at 12:04 PM, Brian Savage <savage<at>uri.edu> wrote:
Kuang He
Sure, I actually prefer ANSI C syntax myself. I used K&R style
Looking over the patches I have a couple of suggestions/comments.
Could we use ANSI C syntax rather than K&R for function
declarations. This
would allow for the easy creation of header files by copying and
it is a
current/standard notation.
function declarations just to conform to the existing code base. I'll
use ANSI C style from now on.
Pager use: We need to check the system for the less command on
This is a good idea, but I'm not sure how to do it, so could you help
compilation.
In fact we could check for less, then more, then fallback to the
original
code in sac.
me with this?
A change like this might allow for syntax highlighting in the help
Exactly, and we get to search within help files.
files.
This would mean the current paging code needs to be made into a
I can try isolating the current paging code to a separate function.
seperate function. The pager value could always be checked for to
override
the default.
Where do you think I should put it, in src/ucf/wrhelp.c?
Best regards,
--
Kuang He
Department of Physics
University of Connecticut
Storrs, CT 06269-3046
Tel: +1.860.486.4919
Web: http://www.phys.uconn.edu/~he/
_______________________________________________
sac-dev mailing list
sac-dev<at>iris.washington.edu
http://www.iris.washington.edu/mailman/listinfo/sac-dev
-
Dear Brian,
Thanks for the tip.
How often do systems that run SAC come without ``more''? I guess it is
rare. In this case, we can maybe use something like this
AC_CHECK_PROGS(SAC_PAGER, [$PAGER less more cat])
in configure.in, and determine which pager to use at runtime in this way:
if environment variable $PAGER is defined and is not empty
use that (user's preference has more priority)
else
use PAGER specified by configure script
In the worst scenario users will end up using ``cat'', which means no
paging at all -- this should hardly happen. In this case, users could
still use something like Shift + PageUp/PageDn to scroll back and
forth. The good thing about this approach is that we don't have to
maintain the original pager code any more. What do you think?
One the other hand, for users who install SAC from binary
distributions, how do we decide which pager to use if they don't have
their environment variable $PAGER set? I don't want them end up using
``cat'' or the original pager code, since many of them may already
have ``less'' or ``more'' installed. Maybe after this feature is
merged into the code, we can highly recommend them to set $PAGER in
the manual?
Best regards,
--
Kuang He
Department of Physics
University of Connecticut
Storrs, CT 06269-3046
Tel: +1.860.486.4919
Web: http://www.phys.uconn.edu/~he/
On Sun, Sep 14, 2008 at 1:00 PM, Brian Savage <savage<at>uri.edu> wrote:
Kuang He,
ANSI C it is. Unless someone else has any objections.
To add in a check for "less" and or "more" see:
Autoconf: Generic Program Checks
http://www.gnu.org/software/autoconf/manual/autoconf-2.57/html_chapter/autoconf_5.html#SEC41
Either one of these should work. You can give a list of programs to check
for in the order you would like
Macro: AC_CHECK_PROGS (variable, progs-to-check-for, [value-if-not-found],
[path])
Macro: AC_PATH_PROGS (variable, progs-to-check-for, [value-if-not-found],
[path])
This should go into the configure.in file, then the bootstrap needs to be
re-run.
Isolating the original pager code inside src/sac/wrhelp.c is probably the
best place for it.
-
Kuang He
Looks like less is available on Linux and OSX (the GNU version)
I cannot find a version of less on Solaris, but Solaris does have the
more command.
I would propose doing a check like
AC_CHECK_PROGS(SAC_PAGER, [ less more ])
It would be nice to remove this code as the capability exists
elsewhere outside of SAC.
There is similar code in src/dfm/xlh.c to do paging for the listhdr
command, but it is
coupled to the output of the command. Same for the news command.
Would you suggest
that we change the way these commands operate as well ? This could
be done with a
pipe to the pager. It might be more trouble than it is worth. This
also modifies the current behavior of SAC.
I still think having a fallback to the original SAC paging code is
the safest thing to do.
Cheers
Brian
On Sep 16, 2008, at 3:42 AM , Kuang He wrote:
Dear Brian,
Thanks for the tip.
How often do systems that run SAC come without ``more''? I guess it is
rare. In this case, we can maybe use something like this
AC_CHECK_PROGS(SAC_PAGER, [$PAGER less more cat])
in configure.in, and determine which pager to use at runtime in
this way:
if environment variable $PAGER is defined and is not empty
use that (user's preference has more priority)
else
use PAGER specified by configure script
In the worst scenario users will end up using ``cat'', which means no
paging at all -- this should hardly happen. In this case, users could
still use something like Shift + PageUp/PageDn to scroll back and
forth. The good thing about this approach is that we don't have to
maintain the original pager code any more. What do you think?
One the other hand, for users who install SAC from binary
distributions, how do we decide which pager to use if they don't have
their environment variable $PAGER set? I don't want them end up using
``cat'' or the original pager code, since many of them may already
have ``less'' or ``more'' installed. Maybe after this feature is
merged into the code, we can highly recommend them to set $PAGER in
the manual?
Best regards,
--
Kuang He
Department of Physics
University of Connecticut
Storrs, CT 06269-3046
Tel: +1.860.486.4919
Web: http://www.phys.uconn.edu/~he/
On Sun, Sep 14, 2008 at 1:00 PM, Brian Savage <savage<at>uri.edu> wrote:
Kuang He,
_______________________________________________
ANSI C it is. Unless someone else has any objections.
To add in a check for "less" and or "more" see:
Autoconf: Generic Program Checks
http://www.gnu.org/software/autoconf/manual/autoconf-2.57/
html_chapter/autoconf_5.html#SEC41
Either one of these should work. You can give a list of programs
to check
for in the order you would like
Macro: AC_CHECK_PROGS (variable, progs-to-check-for, [value-if-not-
found],
[path])
Macro: AC_PATH_PROGS (variable, progs-to-check-for, [value-if-not-
found],
[path])
This should go into the configure.in file, then the bootstrap
needs to be
re-run.
Isolating the original pager code inside src/sac/wrhelp.c is
probably the
best place for it.
sac-dev mailing list
sac-dev<at>iris.washington.edu
http://www.iris.washington.edu/mailman/listinfo/sac-dev
-
Dear Brian,
On Tue, Sep 16, 2008 at 12:13 PM, Brian Savage <savage<at>uri.edu> wrote:
There is similar code in src/dfm/xlh.c to do paging for the listhdr command,
I agree with you. For the time being, maybe we can just leave the
but it is
coupled to the output of the command. Same for the news command. Would you
suggest
that we change the way these commands operate as well ? This could be done
with a
pipe to the pager. It might be more trouble than it is worth. This also
modifies the current behavior of SAC.
paging part of the listhdr command alone.
Looks like less is available on Linux and OSX (the GNU version)
There are some problems with checking for available pager programs
I cannot find a version of less on Solaris, but Solaris does have the more
command.
I would propose doing a check like
AC_CHECK_PROGS(SAC_PAGER, [ less more ])
It would be nice to remove this code as the capability exists elsewhere
outside of SAC.
I still think having a fallback to the original SAC paging code is the
safest thing to do.
during the ./configure stage:
(1) For example, if the user does not have ``less'' during the
./configure stage, SAC will use ``more'' to do the paging. After SAC
is compiled and installed, if the users installs ``less'' but does not
set the environment variable $PAGER, SAC will still be using ``more''
all the time. Yes, it is the user's responsibility to set $PAGER
correctly, but who knows.
(2) If developer A compiles a binary SAC package on a system that has
``less'' installed and distributes it to user B who does not have
``less'' on her system, SAC will still try to use ``less'' as its
pager, which will result in a failure.
To solve this problem, I propose that we check for pager programs at
runtime, following the practice of ``man''. So when SAC starts, it'll
run through this decision process:
if the environment variable $PAGER is set to a non-empty value {
use it
} else {
SAC will look for executable files at these locations one by one
/usr/bin/less
/bin/less
/usr/bin/more
/bin/more
and will use the first one available.
If none of the above pages are available, SAC will fall back to its
original paging code.
}
Maybe we could put the above in the initialization part of SAC. How
does that sound?
Best regards,
--
Kuang He
Department of Physics
University of Connecticut
Storrs, CT 06269-3046
Tel: +1.860.486.4919
Web: http://www.phys.uconn.edu/~he/
-
On Tue, Sep 16, 2008 at 5:22 PM, Kuang He <icrazy<at>gmail.com> wrote:
To solve this problem, I propose that we check for pager programs at
I've implemented the above algorithm, and the patch (against v101.2)
runtime, following the practice of ``man''. So when SAC starts, it'll
run through this decision process:
if the environment variable $PAGER is set to a non-empty value {
use it
} else {
SAC will look for executable files at these locations one by one
/usr/bin/less
/bin/less
/usr/bin/more
/bin/more
and will use the first one available.
If none of the above pages are available, SAC will fall back to its
original paging code.
}
is attached.
If you have applied the previous patch included in:
http://www.iris.washington.edu/pipermail/sac-dev/2008-September/000100.html
You need to reverse that patch first using ``patch -R'', before
applying this patch.
In this patch:
- The original paging code has been isolated to internal_pager(), in
file src/ucf/wrhelp.c
- SAC will try to determine which pager to use and whether to use the
internal pager or not during the initialization stage
- When the user has a broken setting of $PAGER, the error message is clearer:
$ export PAGER=xxxx
$ sac
SAC> h
sh: xxxx: not found
Error loading pager xxxx!
- A global variable pager has been defined in src/top/initsac.c, and I
don't know whether there is a better place for that.
Best regards,
--
Kuang He
Department of Physics
University of Connecticut
Storrs, CT 06269-3046
Tel: +1.860.486.4919
Web: http://www.phys.uconn.edu/~he/
Attachments
-
-
-
-
-
-
Brian Savage
