GNU bug report logs -
#24256
new module for grep's DFA matcher
Previous Next
Reported by: Jim Meyering <jim <at> meyering.net>
Date: Wed, 17 Aug 2016 21:05:01 UTC
Severity: wishlist
Tags: patch
Done: Paul Eggert <eggert <at> cs.ucla.edu>
Bug is archived. No further changes may be made.
To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 24256 in the body.
You can then email your comments to 24256 AT debbugs.gnu.org in the normal way.
Toggle the display of automated, internal messages from the tracker.
Report forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Wed, 17 Aug 2016 21:05:01 GMT)
Full text and
rfc822 format available.
Acknowledgement sent
to
Jim Meyering <jim <at> meyering.net>
:
New bug report received and forwarded. Copy sent to
bug-grep <at> gnu.org
.
(Wed, 17 Aug 2016 21:05:02 GMT)
Full text and
rfc822 format available.
Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
Now that two gnulib-using packages are using grep's DFA matcher, I am
going to move that code into gnulib, so there is one clear source of
truth.
GNU sed began using dfa.[ch] very recently.
While gawk has been copying grep's dfa.[ch] for a long time, it does
not use gnulib.
To that end, I've written both the add-to-gnulib patch and the
use-from-grep patches and tested via grep's "make distcheck" and via
this for gnulib:
./gnulib-tool --create-testdir --dir=/tmp/x --with-tests --test dfa
Note that this merely moves grep/src/dfa.[ch] to gnulib/lib/ and also
moves the DFA-specific tests from grep/tests/. It does not add any new
tests. I am in no great hurry to make the switch, and expect to wait
at least for a few more pending patches by Norihiro Tanaka, syncing
the result to gnulib before pulling the trigger.
This is all the result of a couple of hours last night, so it's almost
certainly not perfect. Feedback welcome.
[gnulib-dfa.diff (text/plain, attachment)]
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Wed, 17 Aug 2016 21:15:01 GMT)
Full text and
rfc822 format available.
Message #8 received at 24256 <at> debbugs.gnu.org (full text, mbox):
This all sounds good, thanks. Your attachment was a submodule diff (basically, a
pair of git commit IDs), so not much to review.
I have some dfa.c-related changes pending. I assume it'll be OK for me to commit
them to grep as needed, and that when you make the switch you'll grab the
then-current version from grep. If you need me to get out of the water around
that time, please just let me know.
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Thu, 18 Aug 2016 00:12:01 GMT)
Full text and
rfc822 format available.
Message #11 received at 24256 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
On Wed, Aug 17, 2016 at 2:14 PM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
> This all sounds good, thanks. Your attachment was a submodule diff
> (basically, a pair of git commit IDs), so not much to review.
Oops. Wrong working directory. Thanks.
I've attached the real diff, this time.
> I have some dfa.c-related changes pending. I assume it'll be OK for me to
> commit them to grep as needed, and that when you make the switch you'll grab
> the then-current version from grep. If you need me to get out of the water
> around that time, please just let me know.
As long as there are still soon-to-be-committed changes for grep's
dfa.[ch], I'm happy to wait.
[gnulib-dfa.diff (text/plain, attachment)]
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Fri, 09 Sep 2016 05:28:01 GMT)
Full text and
rfc822 format available.
Message #14 received at 24256 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
On Wed, Aug 17, 2016 at 5:11 PM, Jim Meyering <jim <at> meyering.net> wrote:
> On Wed, Aug 17, 2016 at 2:14 PM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
>> This all sounds good, thanks. Your attachment was a submodule diff
>> (basically, a pair of git commit IDs), so not much to review.
>
> Oops. Wrong working directory. Thanks.
> I've attached the real diff, this time.
>
>> I have some dfa.c-related changes pending. I assume it'll be OK for me to
>> commit them to grep as needed, and that when you make the switch you'll grab
>> the then-current version from grep. If you need me to get out of the water
>> around that time, please just let me know.
>
> As long as there are still soon-to-be-committed changes for grep's
> dfa.[ch], I'm happy to wait.
I have sync'd the latest grep/dfa bits into my gnulib
dfa-module-adding patch, and updated the grep-side patch that makes
grep use that new module. Attached for anyone who cares to review.
If you're interested in trying it, you should first "cd" into grep's
gnulib directory and use "git am gnulib*.diff" to apply the first
patch there. Then "cd .. && git ci -m 'commit local gnulib+dfa change'
-a", then apply the grep patch and rerun the usual ./bootstrap &&
./configure ... && make && make test
[gnulib-new-dfa-module.diff (text/plain, attachment)]
[grep-use-gnulib-s-new-dfa-module.diff (text/plain, attachment)]
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sat, 10 Sep 2016 18:01:01 GMT)
Full text and
rfc822 format available.
Message #17 received at 24256 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
On Thu, Sep 8, 2016 at 10:27 PM, Jim Meyering <jim <at> meyering.net> wrote:
> On Wed, Aug 17, 2016 at 5:11 PM, Jim Meyering <jim <at> meyering.net> wrote:
>> On Wed, Aug 17, 2016 at 2:14 PM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
>>> This all sounds good, thanks. Your attachment was a submodule diff
>>> (basically, a pair of git commit IDs), so not much to review.
>>
>> Oops. Wrong working directory. Thanks.
>> I've attached the real diff, this time.
>>
>>> I have some dfa.c-related changes pending. I assume it'll be OK for me to
>>> commit them to grep as needed, and that when you make the switch you'll grab
>>> the then-current version from grep. If you need me to get out of the water
>>> around that time, please just let me know.
>>
>> As long as there are still soon-to-be-committed changes for grep's
>> dfa.[ch], I'm happy to wait.
>
> I have sync'd the latest grep/dfa bits into my gnulib
> dfa-module-adding patch, and updated the grep-side patch that makes
> grep use that new module. Attached for anyone who cares to review.
> If you're interested in trying it, you should first "cd" into grep's
> gnulib directory and use "git am gnulib*.diff" to apply the first
> patch there. Then "cd .. && git ci -m 'commit local gnulib+dfa change'
> -a", then apply the grep patch and rerun the usual ./bootstrap &&
> ./configure ... && make && make test
I found/fixed a few nits in the gnulib diff, and changed a test
dependency from progname to getprogname. I have pushed this to gnulib:
[gnulib-new-dfa-module.diff (text/plain, attachment)]
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sat, 10 Sep 2016 18:16:02 GMT)
Full text and
rfc822 format available.
Message #20 received at 24256 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
On Sat, Sep 10, 2016 at 10:59 AM, Jim Meyering <jim <at> meyering.net> wrote:
> On Thu, Sep 8, 2016 at 10:27 PM, Jim Meyering <jim <at> meyering.net> wrote:
>> On Wed, Aug 17, 2016 at 5:11 PM, Jim Meyering <jim <at> meyering.net> wrote:
>>> On Wed, Aug 17, 2016 at 2:14 PM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
>>>> This all sounds good, thanks. Your attachment was a submodule diff
>>>> (basically, a pair of git commit IDs), so not much to review.
>>>
>>> Oops. Wrong working directory. Thanks.
>>> I've attached the real diff, this time.
>>>
>>>> I have some dfa.c-related changes pending. I assume it'll be OK for me to
>>>> commit them to grep as needed, and that when you make the switch you'll grab
>>>> the then-current version from grep. If you need me to get out of the water
>>>> around that time, please just let me know.
>>>
>>> As long as there are still soon-to-be-committed changes for grep's
>>> dfa.[ch], I'm happy to wait.
>>
>> I have sync'd the latest grep/dfa bits into my gnulib
>> dfa-module-adding patch, and updated the grep-side patch that makes
>> grep use that new module. Attached for anyone who cares to review.
>> If you're interested in trying it, you should first "cd" into grep's
>> gnulib directory and use "git am gnulib*.diff" to apply the first
>> patch there. Then "cd .. && git ci -m 'commit local gnulib+dfa change'
>> -a", then apply the grep patch and rerun the usual ./bootstrap &&
>> ./configure ... && make && make test
>
> I found/fixed a few nits in the gnulib diff, and changed a test
> dependency from progname to getprogname. I have pushed this to gnulib:
I have just now pushed these two commits to grep's repository, making
it use gnulib's new dfa module. The first just updates to latest
gnulib submodule. The second removes the subsumed files and grep's
references to them.
[grep-use-gnulib-dfa-module.diff (text/plain, attachment)]
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sat, 10 Sep 2016 18:46:02 GMT)
Full text and
rfc822 format available.
Message #23 received at 24256 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
On Sat, Sep 10, 2016 at 11:15 AM, Jim Meyering <jim <at> meyering.net> wrote:
> On Sat, Sep 10, 2016 at 10:59 AM, Jim Meyering <jim <at> meyering.net> wrote:
>> On Thu, Sep 8, 2016 at 10:27 PM, Jim Meyering <jim <at> meyering.net> wrote:
>>> On Wed, Aug 17, 2016 at 5:11 PM, Jim Meyering <jim <at> meyering.net> wrote:
>>>> On Wed, Aug 17, 2016 at 2:14 PM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
>>>>> This all sounds good, thanks. Your attachment was a submodule diff
>>>>> (basically, a pair of git commit IDs), so not much to review.
>>>>
>>>> Oops. Wrong working directory. Thanks.
>>>> I've attached the real diff, this time.
>>>>
>>>>> I have some dfa.c-related changes pending. I assume it'll be OK for me to
>>>>> commit them to grep as needed, and that when you make the switch you'll grab
>>>>> the then-current version from grep. If you need me to get out of the water
>>>>> around that time, please just let me know.
>>>>
>>>> As long as there are still soon-to-be-committed changes for grep's
>>>> dfa.[ch], I'm happy to wait.
>>>
>>> I have sync'd the latest grep/dfa bits into my gnulib
>>> dfa-module-adding patch, and updated the grep-side patch that makes
>>> grep use that new module. Attached for anyone who cares to review.
>>> If you're interested in trying it, you should first "cd" into grep's
>>> gnulib directory and use "git am gnulib*.diff" to apply the first
>>> patch there. Then "cd .. && git ci -m 'commit local gnulib+dfa change'
>>> -a", then apply the grep patch and rerun the usual ./bootstrap &&
>>> ./configure ... && make && make test
>>
>> I found/fixed a few nits in the gnulib diff, and changed a test
>> dependency from progname to getprogname. I have pushed this to gnulib:
>
> I have just now pushed these two commits to grep's repository, making
> it use gnulib's new dfa module. The first just updates to latest
> gnulib submodule. The second removes the subsumed files and grep's
> references to them.
Finally, here is a similar change, but now for gnu sed. For it, I also
had to make some small code changes to adapt to the new
localeinfo.[ch] and to the dfasyntax API change. I'll wait a day or so
before pushing this, just in case:
[sed-use-gnulib-dfa-module.diff (text/plain, attachment)]
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sat, 10 Sep 2016 21:01:01 GMT)
Full text and
rfc822 format available.
Message #26 received at 24256 <at> debbugs.gnu.org (full text, mbox):
Hello Jim and all,
> On Sep 10, 2016, at 14:44, Jim Meyering <jim <at> meyering.net> wrote:
>
> Finally, here is a similar change, but now for gnu sed. For it, I also
> had to make some small code changes to adapt to the new
> localeinfo.[ch] and to the dfasyntax API change. I'll wait a day or so
> before pushing this, just in case:
> <sed-use-gnulib-dfa-module.diff>
for sed,
It seems that pulling the latest gnulib (the first out of the two commits in the attached diff) causes the following link error (at least on mac os x):
[...]
make[2]: Nothing to be done for `all'.
Making all in .
CCLD sed/sed
Undefined symbols for architecture x86_64:
"_strverscmp", referenced from:
_compile_program in sed_sed-compile.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make[2]: *** [sed/sed] Error 1
make[1]: *** [all-recursive] Error 1
regards,
- assaf
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sun, 11 Sep 2016 00:56:02 GMT)
Full text and
rfc822 format available.
Message #29 received at 24256 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
On Sat, Sep 10, 2016 at 2:00 PM, Assaf Gordon <assafgordon <at> gmail.com> wrote:
> Hello Jim and all,
>
>> On Sep 10, 2016, at 14:44, Jim Meyering <jim <at> meyering.net> wrote:
>>
>> Finally, here is a similar change, but now for gnu sed. For it, I also
>> had to make some small code changes to adapt to the new
>> localeinfo.[ch] and to the dfasyntax API change. I'll wait a day or so
>> before pushing this, just in case:
>> <sed-use-gnulib-dfa-module.diff>
>
> for sed,
> It seems that pulling the latest gnulib (the first out of the two commits in the attached diff) causes the following link error (at least on mac os x):
>
> [...]
> make[2]: Nothing to be done for `all'.
> Making all in .
> CCLD sed/sed
> Undefined symbols for architecture x86_64:
> "_strverscmp", referenced from:
> _compile_program in sed_sed-compile.o
> ld: symbol(s) not found for architecture x86_64
> clang: error: linker command failed with exit code 1 (use -v to see invocation)
> make[2]: *** [sed/sed] Error 1
> make[1]: *** [all-recursive] Error 1
Thanks for reporting that.
I've pushed the attached to gnulib to address that:
[gnulib-strverscmp-vs-osx.diff (text/plain, attachment)]
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Mon, 12 Sep 2016 03:56:01 GMT)
Full text and
rfc822 format available.
Message #32 received at 24256 <at> debbugs.gnu.org (full text, mbox):
On Sat, Sep 10, 2016 at 5:55 PM, Jim Meyering <jim <at> meyering.net> wrote:
> On Sat, Sep 10, 2016 at 2:00 PM, Assaf Gordon <assafgordon <at> gmail.com> wrote:
>> Hello Jim and all,
>>
>>> On Sep 10, 2016, at 14:44, Jim Meyering <jim <at> meyering.net> wrote:
>>>
>>> Finally, here is a similar change, but now for gnu sed. For it, I also
>>> had to make some small code changes to adapt to the new
>>> localeinfo.[ch] and to the dfasyntax API change. I'll wait a day or so
>>> before pushing this, just in case:
>>> <sed-use-gnulib-dfa-module.diff>
>>
>> for sed,
>> It seems that pulling the latest gnulib (the first out of the two commits in the attached diff) causes the following link error (at least on mac os x):
>>
>> [...]
>> make[2]: Nothing to be done for `all'.
>> Making all in .
>> CCLD sed/sed
>> Undefined symbols for architecture x86_64:
>> "_strverscmp", referenced from:
>> _compile_program in sed_sed-compile.o
>> ld: symbol(s) not found for architecture x86_64
>> clang: error: linker command failed with exit code 1 (use -v to see invocation)
>> make[2]: *** [sed/sed] Error 1
>> make[1]: *** [all-recursive] Error 1
>
> Thanks for reporting that.
> I've pushed the attached to gnulib to address that:
Then, I rebased the sed commits on top of latest gnulib and pushed those, too.
Severity set to 'wishlist' from 'normal'
Request was from
Paul Eggert <eggert <at> cs.ucla.edu>
to
control <at> debbugs.gnu.org
.
(Fri, 16 Sep 2016 08:03:01 GMT)
Full text and
rfc822 format available.
Added tag(s) patch.
Request was from
Paul Eggert <eggert <at> cs.ucla.edu>
to
control <at> debbugs.gnu.org
.
(Fri, 16 Sep 2016 08:03:01 GMT)
Full text and
rfc822 format available.
bug closed, send any further explanations to
24256 <at> debbugs.gnu.org and Jim Meyering <jim <at> meyering.net>
Request was from
Paul Eggert <eggert <at> cs.ucla.edu>
to
control <at> debbugs.gnu.org
.
(Fri, 16 Sep 2016 08:03:01 GMT)
Full text and
rfc822 format available.
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sun, 25 Sep 2016 03:44:01 GMT)
Full text and
rfc822 format available.
Message #41 received at 24256 <at> debbugs.gnu.org (full text, mbox):
After this change, 'make dist' no longer works on the current Ubuntu version
(16.04.1 LTS). The failure is as follows. I'm not quite sure how to debug this
sort of thing, as the syntax checks are a bit twisty. Although this isn't a huge
obstacle to issuing a new grep release (as one can run 'make dist' on the
current Fedora version), it's still a bit troublesome.
PS. I updated to the latest gnulib to fix a 'make check' failure on AIX 7.1.
...
space_tab
0.01 space_tab
tight_scope
make[4]: Entering directory '/home/eggert/src/gnu/grep/src'
dfaalloc
dfacomp
dfaexec
dfafree
dfaisfast
dfamust
dfamustfree
dfasuperset
dfasyntax
case_folded_counterparts
init_localeinfo
the above functions should have static scope
/home/eggert/src/gnu/grep/tight-scope.mk:42: recipe for target '_gl_tight_scope'
failed
make[4]: *** [_gl_tight_scope] Error 1
make[4]: Leaving directory '/home/eggert/src/gnu/grep/src'
maint.mk:1578: recipe for target 'sc_tight_scope' failed
make[3]: *** [sc_tight_scope] Error 1
make[3]: Leaving directory '/home/eggert/src/gnu/grep'
Makefile:1710: recipe for target 'run-syntax-check' failed
make[2]: *** [run-syntax-check] Error 2
make[2]: Leaving directory '/home/eggert/src/gnu/grep'
Makefile:1386: recipe for target 'distdir' failed
make[1]: *** [distdir] Error 2
make[1]: Leaving directory '/home/eggert/src/gnu/grep'
Makefile:1487: recipe for target 'dist' failed
make: *** [dist] Error 2
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sun, 25 Sep 2016 03:46:01 GMT)
Full text and
rfc822 format available.
Message #44 received at 24256 <at> debbugs.gnu.org (full text, mbox):
On Sat, Sep 24, 2016 at 8:43 PM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
> After this change, 'make dist' no longer works on the current Ubuntu version
> (16.04.1 LTS). The failure is as follows. I'm not quite sure how to debug
> this sort of thing, as the syntax checks are a bit twisty. Although this
> isn't a huge obstacle to issuing a new grep release (as one can run 'make
> dist' on the current Fedora version), it's still a bit troublesome.
>
> PS. I updated to the latest gnulib to fix a 'make check' failure on AIX 7.1.
>
> ...
> space_tab
> 0.01 space_tab
> tight_scope
> make[4]: Entering directory '/home/eggert/src/gnu/grep/src'
> dfaalloc
> dfacomp
> dfaexec
> dfafree
> dfaisfast
> dfamust
> dfamustfree
> dfasuperset
> dfasyntax
> case_folded_counterparts
> init_localeinfo
> the above functions should have static scope
> /home/eggert/src/gnu/grep/tight-scope.mk:42: recipe for target
> '_gl_tight_scope' failed
> make[4]: *** [_gl_tight_scope] Error 1
> make[4]: Leaving directory '/home/eggert/src/gnu/grep/src'
> maint.mk:1578: recipe for target 'sc_tight_scope' failed
> make[3]: *** [sc_tight_scope] Error 1
> make[3]: Leaving directory '/home/eggert/src/gnu/grep'
> Makefile:1710: recipe for target 'run-syntax-check' failed
> make[2]: *** [run-syntax-check] Error 2
> make[2]: Leaving directory '/home/eggert/src/gnu/grep'
> Makefile:1386: recipe for target 'distdir' failed
> make[1]: *** [distdir] Error 2
> make[1]: Leaving directory '/home/eggert/src/gnu/grep'
> Makefile:1487: recipe for target 'dist' failed
> make: *** [dist] Error 2
>
Hi Paul,
That is probably because you have a stray dfa.o file laying around.
Since that corresponds to a now-VC-removed dfa.c file, "make clean"
fails to remove the .o file. Just remove it manually and the test
should then pass.
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sun, 25 Sep 2016 03:47:01 GMT)
Full text and
rfc822 format available.
Message #47 received at 24256 <at> debbugs.gnu.org (full text, mbox):
On Sat, Sep 24, 2016 at 8:43 PM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
...
> PS. I updated to the latest gnulib to fix a 'make check' failure on AIX 7.1.
Thank you!
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sun, 25 Sep 2016 03:54:01 GMT)
Full text and
rfc822 format available.
Message #50 received at 24256 <at> debbugs.gnu.org (full text, mbox):
Jim Meyering wrote:
> That is probably because you have a stray dfa.o file laying around.
> Since that corresponds to a now-VC-removed dfa.c file, "make clean"
> fails to remove the .o file. Just remove it manually and the test
> should then pass.
Right you are, thanks. Running 'make clean' fixes the problem.
Though it exposes a new problem. 'make clean; make dist' fails when checking
tight scope, with the following symptoms. Perhaps a missing dependency?
space_tab
0.01 space_tab
tight_scope
make[4]: Entering directory '/home/eggert/src/gnu/grep/src'
CC grep.o
In file included from system.h:26:0,
from grep.c:29:
../lib/binary-io.h: In function 'set_binary_mode':
../lib/binary-io.h:57:10: error: 'O_BINARY' undeclared (first use in this function)
return O_BINARY;
^
../lib/binary-io.h:57:10: note: each undeclared identifier is reported only once
for each function it appears in
In file included from grep.c:29:0:
system.h: At top level:
system.h:27:24: fatal error: configmake.h: No such file or directory
compilation terminated.
Makefile:1344: recipe for target 'grep.o' failed
make[4]: *** [grep.o] Error 1
make[4]: Leaving directory '/home/eggert/src/gnu/grep/src'
maint.mk:1578: recipe for target 'sc_tight_scope' failed
make[3]: *** [sc_tight_scope] Error 1
make[3]: Leaving directory '/home/eggert/src/gnu/grep'
Makefile:1710: recipe for target 'run-syntax-check' failed
make[2]: *** [run-syntax-check] Error 2
make[2]: Leaving directory '/home/eggert/src/gnu/grep'
Makefile:1386: recipe for target 'distdir' failed
make[1]: *** [distdir] Error 2
make[1]: Leaving directory '/home/eggert/src/gnu/grep'
Makefile:1487: recipe for target 'dist' failed
make: *** [dist] Error 2
550-day $
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sun, 25 Sep 2016 04:11:02 GMT)
Full text and
rfc822 format available.
Message #53 received at 24256 <at> debbugs.gnu.org (full text, mbox):
On Sat, Sep 24, 2016 at 8:53 PM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
> Jim Meyering wrote:
>>
>> That is probably because you have a stray dfa.o file laying around.
>> Since that corresponds to a now-VC-removed dfa.c file, "make clean"
>> fails to remove the .o file. Just remove it manually and the test
>> should then pass.
>
>
> Right you are, thanks. Running 'make clean' fixes the problem.
>
> Though it exposes a new problem. 'make clean; make dist' fails when checking
> tight scope, with the following symptoms. Perhaps a missing dependency?
>
> space_tab
> 0.01 space_tab
> tight_scope
> make[4]: Entering directory '/home/eggert/src/gnu/grep/src'
> CC grep.o
> In file included from system.h:26:0,
> from grep.c:29:
> ../lib/binary-io.h: In function 'set_binary_mode':
> ../lib/binary-io.h:57:10: error: 'O_BINARY' undeclared (first use in this
> function)
> return O_BINARY;
> ^
> ../lib/binary-io.h:57:10: note: each undeclared identifier is reported only
> once for each function it appears in
> In file included from grep.c:29:0:
> system.h: At top level:
> system.h:27:24: fatal error: configmake.h: No such file or directory
> compilation terminated.
> Makefile:1344: recipe for target 'grep.o' failed
> make[4]: *** [grep.o] Error 1
> make[4]: Leaving directory '/home/eggert/src/gnu/grep/src'
> maint.mk:1578: recipe for target 'sc_tight_scope' failed
> make[3]: *** [sc_tight_scope] Error 1
> make[3]: Leaving directory '/home/eggert/src/gnu/grep'
> Makefile:1710: recipe for target 'run-syntax-check' failed
> make[2]: *** [run-syntax-check] Error 2
> make[2]: Leaving directory '/home/eggert/src/gnu/grep'
> Makefile:1386: recipe for target 'distdir' failed
> make[1]: *** [distdir] Error 2
> make[1]: Leaving directory '/home/eggert/src/gnu/grep'
> Makefile:1487: recipe for target 'dist' failed
> make: *** [dist] Error 2
> 550-day $
Sounds like it must be something fundamental. I've just confirmed that
at commit v2.25-100-gcd8082e `./bootstrap && ./configure && make &&
make distcheck` succeeds on both Fedora 24 and Debian unstable.
Did you rerun ./bootstrap?
Note that I find myself regularly running `(cd gnulib && git co master
&& git pull)` to update each package's gnulib/ subdir to the latest.
Do that before the bootstrap.
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sun, 25 Sep 2016 04:23:02 GMT)
Full text and
rfc822 format available.
Message #56 received at 24256 <at> debbugs.gnu.org (full text, mbox):
Jim Meyering wrote:
> Sounds like it must be something fundamental. I've just confirmed that
> at commit v2.25-100-gcd8082e `./bootstrap && ./configure && make &&
> make distcheck` succeeds on both Fedora 24 and Debian unstable.
>
> Did you rerun ./bootstrap?
Yes.
I think the difference is that I'm not running plain 'make' first. I jump
straight to 'make dist'. That is, from a fresh copy of grep (same commit as
yours), I run:
./bootstrap && ./configure && make dist
and this gets the tight_scope failure I mentioned earlier, on Ubuntu 16.04.1 LTS
anyway.
Information forwarded
to
bug-grep <at> gnu.org
:
bug#24256
; Package
grep
.
(Sun, 25 Sep 2016 04:44:01 GMT)
Full text and
rfc822 format available.
Message #59 received at 24256 <at> debbugs.gnu.org (full text, mbox):
[Message part 1 (text/plain, inline)]
On Sat, Sep 24, 2016 at 9:22 PM, Paul Eggert <eggert <at> cs.ucla.edu> wrote:
> Jim Meyering wrote:
>>
>> Sounds like it must be something fundamental. I've just confirmed that
>> at commit v2.25-100-gcd8082e `./bootstrap && ./configure && make &&
>> make distcheck` succeeds on both Fedora 24 and Debian unstable.
>>
>> Did you rerun ./bootstrap?
>
>
> Yes.
>
> I think the difference is that I'm not running plain 'make' first. I jump
> straight to 'make dist'. That is, from a fresh copy of grep (same commit as
> yours), I run:
>
> ./bootstrap && ./configure && make dist
>
> and this gets the tight_scope failure I mentioned earlier, on Ubuntu 16.04.1
> LTS anyway.
Oh, thanks. That was indeed a missing dependency.
This patch fixes it for me:
[build--avoid-missing-dep-failure.diff (application/octet-stream, attachment)]
bug archived.
Request was from
Debbugs Internal Request <help-debbugs <at> gnu.org>
to
internal_control <at> debbugs.gnu.org
.
(Sun, 23 Oct 2016 11:24:04 GMT)
Full text and
rfc822 format available.
This bug report was last modified 7 years and 158 days ago.
Previous Next
GNU bug tracking system
Copyright (C) 1999 Darren O. Benham,
1997,2003 nCipher Corporation Ltd,
1994-97 Ian Jackson.