Progress is being made. I submitted a cleanup patch in order to get rid
of three headaches before the 2.5 fork. Three good things have since
happened since:
(1) Bjorn Wesen <[email protected]> informs me that the CRIS symbol bugs
will be fixed in the next CRIS port update. This will get rid of the
single worst problem I was looking at. (Somebody else pointed out that
the present symbol names are invisible to `make dep'.)
(2) Greg Banks and Rik van Riel pointed out (separately) that I was too close
to the namespace problem; with a two-line change to my compiler I don't
need to get numeric-prefix symbols cleaned up yet. I think it's still
a good idea for the longer term, however, and a couple of the people
who own those symbols have already endorsed changing them.
(3) The CONFIG_AIC7XXX_TAGGED_QUEUEING to CONFIG_AIC7XXX_TCQ_ON_BY_DEFAULT
change that was on my buglist seems to have come out in the wash in the
pre8 patch.
All that's left for me to do at this point, therefore, is fix the
PRINTER_READBACK glitch. I will do my analysis tools on the assumption
that the CRIS symbols are going to be fixed.
This file is a patch expressed in two ways. First, procedurally, as a
way to duplicate its effects on any kernel source tree. Secondly, as
an explicit patch against 2.4.3-pre8. It replaces my previous patches
against 2.4.3-pre6 and 2.4.3-pre8.
No actual object code will be changed by this patch.
For the procedural form, you will require the following Python 2.0 script
(note that this script has added anoption since my pre8 patch.):
-------------------------------- SCRIPT BEGINS HERE ---------------------------
#!/usr/bin/env python
#
# symbolreplace -- replace symbols in the current Linux source tree
import sys, os, re, getopt
renamings = {} # Dictionary of translations
dobackup = 1
verbose = 0
#
# This is the only bit of knowledge in this script that knows anything
# about the lexical rules of the symbols we're looking for. It's used as a
# guard expression on either side of each old symbol name -- matches
# must begin either with this guard or with a beginning-of-line, and end
# either with this guard or an end-of-line. The purpose is to prevent false
# matches on FOO in FOOBAR and BARFOO.
#
guard = "[^A-Za-z0-9_]"
def backup(file):
"Create a backup for the given file."
# Compute where the RCS directory for this file should live.
# Create it if necessary.
rcs_dir = os.path.join(os.path.dirname(file), "RCS")
if not os.path.exists(rcs_dir):
os.mkdir(rcs_dir)
# If there is no RCS master corresponding to the file, create one,
rcs_file = os.path.join(rcs_dir, os.path.basename(file) + ",v")
if not os.path.exists(rcs_file):
os.system("ci -u -t- " + file + " >/dev/null 2>&1")
# Check out a working copy if needed.
if not os.access(file, os.W_OK):
os.system("co -l " + file + " >/dev/null 2>&1")
def treevisit(root, visitor):
"Apply the visitor function to every file under the given root path."
def treewalker(hook, dirname, files):
for file in files:
here = os.path.join(dirname, file)
if os.path.isfile(here) and here.find("RCS") == -1:
hook(here)
os.path.walk(root, treewalker, visitor)
def replacer(file):
if verbose:
print file
# Suck the entire file into core as a list of lines.
ifp = open(file, "r")
contents = ifp.read()
ifp.close()
# Now perform the substitution on each line.
substitutions = 0
for pattern in renamings.keys():
new_contents = pattern.sub(r"\1"+renamings[pattern]+r"\2", contents)
if new_contents != contents:
substitutions = substitutions + 1
contents = new_contents
# Spew the results.
if substitutions:
if dobackup:
backup(file)
ofp = open(file, "w")
ofp.write(contents)
ofp.close()
def guardify(old):
"Turn the given symbol into a properly guarded regexp."
return re.compile("(^|" + guard +")" + old + "(" + guard + "|$)")
def read_translations(fp):
"Read translation requests from the given file object, ignoring comments."
while 1:
line = fp.readline()
if not line:
break
elif line[0] == '#' or line[0] == '\n':
continue
else:
(old, new) = line.split()
renamings[guardify(old)] = new
# Main sequence
if __name__ == '__main__':
# Main sequence
# Process options
(options, arguments) = getopt.getopt(sys.argv[1:], "d:ns:v")
topdir = "."
for (switch, val) in options:
if (switch == '-d'): # Location of source tree if not .
topdir = val
elif (switch == '-n'): # Suppress backups
dobackup = 0
elif (switch == '-s'): # Do old:new substitution
(old, new) = val.split(":")
renamings[guardify(old)] = new
elif (switch == '-v'): # Report on files as they are processed
verbose = 1
if not renamings:
read_translations(sys.stdin)
treevisit(topdir, replacer)
# That's all, folks!
-------------------------------- SCRIPT ENDS HERE -----------------------------
Here is the instruction I used to generate the bulk of the patch. A note on
the one manual change follows.
symbolreplace -v -d ~/src/linux -s CONFIG_PRINTER_READBACK:CONFIG_PARPORT_1284
After this, I manually removed the former PRINTER_READBACK entry in the
configuration-help file.
Finally, here is the generated patch:
--- Documentation/Configure.help 2001/03/26 18:15:21 1.1
+++ Documentation/Configure.help 2001/03/26 18:31:57
@@ -3724,12 +3724,6 @@
This code is also available as a module (say M), called
parport_mfc3.o. If in doubt, saying N is the safe plan.
-Support IEEE1284 status readback
-CONFIG_PRINTER_READBACK
- If you have a device on your parallel port that support this
- protocol, this option will allow the device to report its status. It
- is safe to say Y.
-
IEEE1284 transfer modes
CONFIG_PARPORT_1284
If you have a printer that supports status readback or device ID, or
--- Documentation/video4linux/CQcam.txt 2001/03/26 18:15:28 1.1
+++ Documentation/video4linux/CQcam.txt 2001/03/26 18:15:28
@@ -43,7 +43,7 @@
CONFIG_PRINTER M for lp.o, parport.o parport_pc.o modules
CONFIG_PNP_PARPORT M for autoprobe.o IEEE1284 readback module
- CONFIG_PRINTER_READBACK M for parport_probe.o IEEE1284 readback module
+ CONFIG_PARPORT_1284 M for parport_probe.o IEEE1284 readback module
CONFIG_VIDEO_DEV M for videodev.o video4linux module
CONFIG_VIDEO_CQCAM M for c-qcam.o Color Quickcam module
@@ -127,7 +127,7 @@
The c-qcam is IEEE1284 compatible, so if you are using the proc file
system (CONFIG_PROC_FS), the parallel printer support
-(CONFIG_PRINTER), the IEEE 1284 system,(CONFIG_PRINTER_READBACK), you
+(CONFIG_PRINTER), the IEEE 1284 system,(CONFIG_PARPORT_1284), you
should be able to read some identification from your quickcam with
modprobe -v parport
--- arch/m68k/config.in 2001/03/26 18:14:52 1.1
+++ arch/m68k/config.in 2001/03/26 18:14:52
@@ -140,7 +140,7 @@
fi
dep_tristate ' Parallel printer support' CONFIG_PRINTER $CONFIG_PARPORT
if [ "$CONFIG_PRINTER" != "n" ]; then
- bool ' Support IEEE1284 status readback' CONFIG_PRINTER_READBACK
+ bool ' Support IEEE1284 status readback' CONFIG_PARPORT_1284
fi
fi
End of patch.
--
<a href="http://www.tuxedo.org/~esr/">Eric S. Raymond</a>
A ``decay in the social contract'' is detectable; there is a growing
feeling, particularly among middle-income taxpayers, that they are not
getting back, from society and government, their money's worth for
taxes paid. The tendency is for taxpayers to try to take more control
of their finances ..
-- IRS Strategic Plan, (May 1984)
Wow, your script was longer than your patch :)
Patch looks ok to me...
--
Jeff Garzik | May you have warm words on a cold evening,
Building 1024 | a full moon on a dark night,
MandrakeSoft | and a smooth road all the way to your door.
Jeff Garzik <[email protected]>:
> Wow, your script was longer than your patch :)
But independent of the version/patchlevel, which was the point of shipping it.
If CML2 is adopted and I become the config system maimtainer,
symbolreplace is one of a number of small tools I'll drop into the scripts
directory to address the namespace-management problem. At over 1800
symbols, power tools are needed. Even just garbage-collecting unused
symbols isn't trivial.
--
<a href="http://www.tuxedo.org/~esr/">Eric S. Raymond</a>
"The power to tax involves the power to destroy;...the power to
destroy may defeat and render useless the power to create...."
-- Chief Justice John Marshall, 1819.
Eric S. Raymond wrote:
> If CML2 is adopted and I become the config system maimtainer,
^
Typo-ROTFL!
--
There is / one art || John Cowan <[email protected]>
no more / no less || http://www.reutershealth.com
to do / all things || http://www.ccil.org/~cowan
with art- / lessness \\ -- Piet Hein
John Cowan <[email protected]>:
> > If CML2 is adopted and I become the config system maimtainer,
> ^
> Typo-ROTFL!
Of such errors are linguistic innovations made. I wonder if this one
will propagate enough that I have to put it in the Jargon File? :-)
--
<a href="http://www.tuxedo.org/~esr/">Eric S. Raymond</a>
Our society won't be truly free until "None of the Above" is always an option.
> if [ "$CONFIG_PRINTER" != "n" ]; then
>- bool ' Support IEEE1284 status readback' CONFIG_PRINTER_READBACK
>+ bool ' Support IEEE1284 status readback' CONFIG_PARPORT_1284
> fi
This isn't really right. Although it's true that CONFIG_PARPORT_1284 enables
the stuff that used to be controlled by CONFIG_PRINTER_READBACK, the two
aren't synonymous. The m68k port ought to just use drivers/parport/Config.in
like everybody else.
p.
Philip Blundell <[email protected]>:
> > if [ "$CONFIG_PRINTER" != "n" ]; then
> >- bool ' Support IEEE1284 status readback' CONFIG_PRINTER_READBACK
> >+ bool ' Support IEEE1284 status readback' CONFIG_PARPORT_1284
> > fi
>
> This isn't really right. Although it's true that
> CONFIG_PARPORT_1284 enables the stuff that used to be controlled by
> CONFIG_PRINTER_READBACK, the two aren't synonymous. The m68k port
> ought to just use drivers/parport/Config.in like everybody else.
Yes, I figured this out. The CML2 rules file section for m68k already uses
the normal, architecture-independent parallel-port subtree (some of the
individual options in it are suppressed on certain platforms).
At this point, the only place CONFIG_PRINTER_READBACK actually occurs
seems to be in doc files. So selecting it is a no-op, but one that
obviously means the user wanted the IEEE1284 support. Even given the
two aren't equivalent, I don't see a better way to handle this than the
above. Let's be kind to our 2.4.x users.
--
<a href="http://www.tuxedo.org/~esr/">Eric S. Raymond</a>
The right of self-defense is the first law of nature: in most
governments it has been the study of rulers to confine this right
within the narrowest limits possible. Wherever standing armies
are kept up, and when the right of the people to keep and bear
arms is, under any color or pretext whatsoever, prohibited,
liberty, if not already annihilated, is on the brink of
destruction."
-- Henry St. George Tucker (in Blackstone's Commentaries)