2018-05-03 14:34 GMT+02:00 Euan Harris :
> In nsenter from util-linux, the long version of the -n option is
> --net=. BusyBox's version expects --network=, so scripts
> and examples written for util-linux's version cause BusyBox's version
> to exit with the usage message.
>
> Confusingly, until c
2017-10-03 10:23 GMT+02:00 Avi Fishman :
> Tested-by: Avi Fishman
>
I'm afraid it doesn't work like this. :)
You need someone else to mark a patch as tested. I'll try to find some
hardware later that supports block data and test it myself.
--
Best regards,
Bartosz Golaszewski
_
2017-10-02 14:17 GMT+02:00 Avi Fishman :
> When 'i' or 's' modes are selected block array is filled from offset 3
> (blen = 3) but copied to data.block buffer from offset 0 so first 3 bytes
> contains garbage from stack.
> The buffer that is sent is also 3 bytes too long due to those extra 3
> byte
2017-08-28 15:52 GMT+02:00 Bartosz Gołaszewski :
> 2017-08-28 15:48 GMT+02:00 Denys Vlasenko :
>> On Mon, Aug 28, 2017 at 11:49 AM, Bartosz Golaszewski
>> wrote:
>>> Installing headers from upstream i2c-tools makes it impossible to
>>> build the busybox version
2017-08-28 11:49 GMT+02:00 Bartosz Golaszewski :
> Installing headers from upstream i2c-tools makes it impossible to
> build the busybox version due to redefined symbols. This is caused by
> the fact that linux/i2c-dev.h from i2c-tools overwrites the one from
> linux uapi and defines symbols alread
2017-08-28 15:48 GMT+02:00 Denys Vlasenko :
> On Mon, Aug 28, 2017 at 11:49 AM, Bartosz Golaszewski
> wrote:
>> Installing headers from upstream i2c-tools makes it impossible to
>> build the busybox version due to redefined symbols. This is caused by
>> the fact that linux/i2c-dev.h from i2c-tools
2017-04-11 23:58 GMT+02:00 Kaarle Ritvanen :
> Signed-off-by: Kaarle Ritvanen
> ---
> include/libbb.h| 19 +++
> libbb/getopt32.c | 8 +---
> libbb/vfork_daemon_rexec.c | 28 ++--
> runit/sv.c | 7 +--
> shell
Hi Denys,
this patch looks good and fixes an existing issue - is there any
reason why it can't be merged?
Thanks,
Bartosz
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2016-03-17 12:42 GMT+01:00 Bartosz Golaszewski :
> This series adds two namespaces-related applets reimplementing most
> arguments supported by their upstream util-linux counterparts.
>
> The fifth patch is an optional code shrink.
Superseded by v6.
> v5:
> - fixed the define guard in namespace.h
2016-03-15 21:45 GMT+01:00 Mike Frysinger :
> On 15 Mar 2016 12:00, Bartosz Gołaszewski wrote:
>>
>> Actually some time ago I sent a series extending the readahead applet.
>> It was rejected eventually but one of the patches was actually adding
>> a useful macro taken fr
Hi Mike,
thanks for the reviews!
2016-03-17 18:23 GMT+01:00 Mike Frysinger :
> On 17 Mar 2016 15:52, Bartosz Golaszewski wrote:
>> +static void mount_procfs(const char *target)
>> +{
>> + int status;
>> +
>> + status = mount("none", target, NULL, MS_PRIVATE | MS_REC, NULL);
>> + if (s
tside of their .c files, but are not part of any
API and shouldn't be called directly. Is this approach wrong?
2016-03-14 22:24 GMT+01:00 Mike Frysinger :
> On 14 Mar 2016 16:15, Rich Felker wrote:
>> On Mon, Mar 14, 2016 at 02:27:56PM -0400, Mike Frysinger wrote:
>> >
&g
2016-03-14 15:27 GMT+01:00 Mike Frysinger :
> On 14 Mar 2016 11:07, Bartosz Golaszewski wrote:
>> +#ifndef __BB_NAMESPACE_H
>> +#define __BB_NAMESPACE_H
>
> use a naming style like other busybox headers
>
>> +/*
>> + * Longest possible path to a procfs file used in namespace utils. Must be
>> + * a
2016-03-14 15:11 GMT+01:00 Mike Frysinger :
> On 14 Mar 2016 11:08, Bartosz Gołaszewski wrote:
>> 2016-03-11 18:24 GMT+01:00 Mike Frysinger :
>> > On 11 Mar 2016 18:10, Bartosz Golaszewski wrote:
>> >> +int unshare_main(int argc UNUSED_PARAM, char **argv)
>
2016-03-11 18:24 GMT+01:00 Mike Frysinger :
> On 11 Mar 2016 18:10, Bartosz Golaszewski wrote:
>> +int unshare_main(int argc UNUSED_PARAM, char **argv)
>> +{
>> + static struct namespace ns_list[] = {
>
> const ?
>
Nope, we're using char *path to store the mount points.
>> + {
>>
2016-03-11 14:07 GMT+01:00 Bartosz Golaszewski :
> This series adds two namespaces-related applets reimplementing most
> arguments supported by their upstream util-linux counterparts.
>
> The fourth patch is an optional code shrink.
>
> Bartosz Golaszewski (4):
> libbb: add BIT() macro
> unshar
2015-12-04 14:25 GMT+01:00 Larissa Braz :
> Hi Walter,
>
> Thanks for answer me. However, my goal isn't to run bb. Actually, I wanna
> to automatically analyse bb code, specifically the git commit history of
> certain files what makes compile the whole project time consuming. For this
> reason, I'
2015-11-16 11:06 GMT+01:00 Laurent Bercot :
>
> The technical impact of Denys's change was actually extremely small,
> which is the main reason why it was not discussed (much) on the list.
Has it been discussed at all? I can't find anything regarding it and
I'm pretty curious as to what caused th
2015-11-16 21:23 GMT+01:00 Baruch Siach :
> Hi Aaro,
>
> On Mon, Nov 16, 2015 at 10:14:05PM +0200, Aaro Koskinen wrote:
>> On Mon, Nov 16, 2015 at 06:20:51AM +0200, Baruch Siach wrote:
>> > On Mon, Nov 16, 2015 at 02:10:53AM +0200, Aaro Koskinen wrote:
>> > > @@ -144,6 +147,9 @@ int nandwrite_main(
2015-10-24 4:26 GMT+02:00 Denys Vlasenko :
> On Fri, Oct 23, 2015 at 3:14 PM, Bartosz Gołaszewski
> wrote:
>>>> Like
>>>> for example being able to redirect the events to stdout instead of
>>>> only being able to pass them as arguments to PROG? This is
2015-10-22 20:31 GMT+02:00 Denys Vlasenko :
> On Wed, Oct 21, 2015 at 11:59 AM, Bartosz Gołaszewski
> wrote:
>> 2015-10-20 14:02 GMT+02:00 Denys Vlasenko :
>>>> Can this software be downloaded and built by e.g. me?
>>>> I would like to see a bigger p
2015-10-22 16:46 GMT+02:00 Denys Vlasenko :
> On Wed, Oct 21, 2015 at 11:56 AM, Bartosz Gołaszewski
> wrote:
>> 2015-10-20 13:17 GMT+02:00 Denys Vlasenko :
>>>
>>> inotifyd syntax is "inotifyd PROG FILE1[:MASK]...",
>>> and PROG is run w
2015-10-20 14:02 GMT+02:00 Denys Vlasenko :
> Hi Bartosz,
>
> On Fri, Oct 16, 2015 at 8:32 PM, Denys Vlasenko
> wrote:
>> On Tue, Sep 15, 2015 at 12:19 PM, Bartosz Gołaszewski
>> wrote:
>>> Hi Denys,
>>>
>>> after giving it a thought I wou
2015-10-20 13:17 GMT+02:00 Denys Vlasenko :
>
> inotifyd syntax is "inotifyd PROG FILE1[:MASK]...",
> and PROG is run with
>
> PROG ACTUAL_EVENTS FILEn [SUBFILE]
>
> This severely limits the number of unmodified PROGs you can run
> to achieve some useful result. For example, even cat'ting newly
2015-09-21 4:17 GMT-07:00 Denys Vlasenko :
> Applied both.
>
> However, I am not entirely sure such changes are useful.
> If people with old toolchains will start asking
> "why the F it stopped building?"
> I'll give them your home address ;) ;)
No worries, this API has existed for years now. And
2015-08-25 15:39 GMT+02:00 Bartosz Golaszewski :
> Add -x option which allows to specify the exit status of PROG for which
> inotifyd should exit.
>
> An example use case for this change is writing parallel system startup
> scripts with busybox' runit: inotifyd can be used to wait for a specific
>
Hi Denys,
after giving it a thought I would like to ask you to reconsider
merging of the readahead daemon mode.
The reason for implementing this applet is a project we did at
BayLibre - an open-hardware power consumption measurement device:
http://baylibre.com/acme/
It's a cape sitting on top of
2015-09-04 3:40 GMT+02:00 Denys Vlasenko :
> Let's keep unused defines.
> They don't hurt, and if we'd need one of them later,
> it's already there :)
Actually I started thinking: most applets that make use of the kernel
UAPI simply include appropriate linux headers. Can you take a look at
the fol
2015-09-01 12:14 GMT+02:00 Bartosz Golaszewski :
> The implementation is more or less feature complete. Remove the parts of
> kernel UAPI copied from include/uapi/linux/i2c.h that aren't used.
>
> Signed-off-by: Bartosz Golaszewski
> ---
> miscutils/i2c_tools.c | 13 -
> 1 file change
2015-08-27 0:45 GMT+02:00 Tito :
>
> On 08/25/2015 09:49 PM, Denys Vlasenko wrote:
>> Fixed in git
>>
>> On Tue, Aug 25, 2015 at 8:16 PM, Bartosz Gołaszewski
>> wrote:
>>> 2015-08-25 19:29 GMT+02:00 Xabier Oneca -- xOneca :
>>>> Hello Ba
2015-08-25 19:29 GMT+02:00 Xabier Oneca -- xOneca :
> Hello Bartosz,
>
> 2015-08-25 13:10 GMT+02:00 Bartosz Golaszewski :
>> +/*
>> + * Return NULL if string is not suffixed with key. Return pointer to the
>> + * beginning of prefix key in string. If key is an empty string return
>> pointer
>> +
2015-08-25 16:25 GMT+02:00 Denys Vlasenko :
> No, it's my fault.
> Without optimization, this code
>
> sockopt = offsetof(struct icmp6_hdr, icmp6_cksum);
> if (sockopt != 2)
> BUG_bad_offsetof_icmp6_cksum();
>
> will not optimize out the comparison. So the generated
2015-08-25 16:38 GMT+02:00 Denys Vlasenko :
> Applied. However, it has no users...
>
Because it was meant for readahead which didn't get much love... ;)
--
Best regards,
Bartosz Golaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busy
I'm unable to compile current git with debugging symbols.
networking/lib.a(ping.o): In function `ping6':
/home/brgl/workspace/gitbook/busybox/networking/ping.c:773: undefined
reference to `BUG_bad_offsetof_icmp6_cksum'
collect2: error: ld returned 1 exit status
Bisect points at commit c52cbea2.
2015-08-25 14:42 GMT+02:00 Denys Vlasenko :
> On Tue, Aug 25, 2015 at 1:09 PM, Bartosz Golaszewski
> wrote:
>> While working on an embedded system running several big services, X-org,
>> fluxbox, Qt5 etc. where the boot-time was awful, it turned out that the
>> readahead implementation from system
2015-08-24 21:29 GMT+02:00 Tito :
>
>
> On 08/24/2015 12:51 AM, Jody Bruchon wrote:
>>
>> On 2015-08-21 17:26, Tito wrote:
>>>
>>> what sense would it make to check if str\0 ends_with \0 if all strings
>>> do end with \0 ?
>>
>>
>> In this case, I view this as a problem of handling all possible val
2015-08-21 16:23 GMT+02:00 Bartosz Golaszewski :
> Implement optional daemon mode for readahead.
>
> Bloatcheck without daemon mode:
> function old new delta
> readahead_main 127 123 -4
> --
2015-08-17 13:07 GMT+02:00 Waldemar Brodkorb :
> Hi,
> Bartosz Gołaszewski wrote,
>
>> 2015-08-15 12:27 GMT+02:00 Bartosz Gołaszewski :
>> > You're right of course from the engineering point of view, but at the
>> > same time I have always regarded busybox
2015-08-17 16:56 GMT+02:00 Denys Vlasenko :
> On Tue, Aug 4, 2015 at 12:04 PM, Bartosz Golaszewski
> wrote:
>> Add -x option which allows to specify the exit status of PROG for which
>> inotifyd should exit.
>>
>> An example use case for this change is writing parallel system startup
>> scripts wi
2015-08-15 12:27 GMT+02:00 Bartosz Gołaszewski :
> You're right of course from the engineering point of view, but at the
> same time I have always regarded busybox as a piece of software that
> get's things done. Uclibc is unfortunately still the default libc in
> buildroot
2015-08-13 16:27 GMT+02:00 Laurent Bercot :
> On 13/08/2015 15:57, Bartosz Golaszewski wrote:
>>
>> Fanotify syscalls are not supported in uClibc. Add a new include
>> file - missing_syscalls.h - implementing the syscalls directly in
>> busybox.
>
>
> It's not a good idea to start fixing libc prob
2015-07-27 11:45 GMT+02:00 purushi1 . :
> Hi Bartosz Golaszewski,
>
> I gave it a try and its the same result :(.
> I disabled the following in defaults/config.busybox and build the kernel.
>
> CONFIG_INIT=n
> CONFIG_GETTY=n
> CONFIG_LOGIN=n
> CONFIG_PASSWD=n
> CONFIG_SU=n
> CONFIG_SULOGIN=n
> CON
2015-07-27 7:45 GMT+02:00 purushi1 . :
> Hi,
>
> As per kernel hardening requirement, i am trying to disable
> "CONFIG_FEATURE_SYSLOG" in defaults/config.busybox and build the kernel.
>
> However, after doing a make, i still see "CONFIG_FEATURE_SYSLOG" enabled
>
>
>
> grep -r "CONFIG_F
2015-06-07 18:08 GMT+02:00 Denys Vlasenko :
> Applied all five patches, thanks!
>
> On Fri, Jun 5, 2015 at 10:47 AM, Bartosz Golaszewski
> wrote:
>> Since there are no more statements in the if block after the while loop
>> in list_i2c_busses_and_exit(), there's no need for a goto.
>>
>> Signed-of
2015-06-03 6:54 GMT+02:00 Wei, Catherine :
> Hi:
> Lately I'm trying to debug the top command in an embedded linux
> box. I've added some printf in top.c, for example:
>
> fprintf(stderr,"catherine %lu %lu %lu %lu %lu
> %lu\n",used,mfree,shared,buffers,cached,total);
> printf("catherine---
2015-04-21 14:29 GMT+02:00 Bartosz Golaszewski :
> I've stumbled upon this bug, where calling i2cdetect -l prints usage and
> exits. The culprit is the opt_complementary string which causes getopt32()
> to bail-out if no non-option arguments are encountered.
>
> I think the check for (argc < 1) in
Hi Bernhard,
2015-04-15 14:26 GMT+02:00 Bernhard Reutner-Fischer :
> On 15 April 2015 at 13:43, Bartosz Golaszewski wrote:
>> In some scenarios busybox performance may be prefered over its
>> small size.
>>
>> Allow users to select the compiler optimization level by adding
>> appropriate config o
2015-04-15 12:45 GMT+02:00 Bernhard Reutner-Fischer :
> On 15 April 2015 at 12:36, Jody Bruchon wrote:
>> On 4/15/2015 6:15 AM, Bartosz Golaszewski wrote:
>>>
>>> Allow users to select the compiler optimization level by adding
>>> appropriate config options.
>>
>> I would suggest dropping -O0 and
2015-03-04 17:04 GMT+01:00 Ari Sundholm :
>> bb_perror_msg_and_die()?
>>
>
> Would we not get an unrelated error string appended?
>
> By changing the above to bb_perror_msg_and_die("no files specified"):
>
> $ ./busybox truncate -c -s 1M
> truncate: no files specified: No such file or directory
>
>
2015-03-04 15:11 GMT+01:00 Ari Sundholm :
> From: Ari Sundholm
>
> bloat-o-meter:
>
> function old new delta
> .rodata 154411 154491 +80
> --
2015-01-18 10:31 GMT+01:00 Matthias Andree :
>
> That was the minor part because I can disable it by build configuration.
>
> What about mempcpy()? That is a showstopper.
Please try the attached patch. I didn't try to compile it on a FreeBSD
system, but if it works for you we can ask Denys to mer
2015-01-06 21:55 GMT+01:00 Matthias Andree :
> Happy new year!
>
> I just tried to upgrade the busybox-unstable port on FreeBSD to 1.23.0,
> but one issue of the two given below is a showstopper.
>
>
> The first one pertains to an incompatibility around sendfile. Easily
> sidestepped by deconfiguri
2014-12-17 20:26 GMT+01:00 Denys Vlasenko :
> Applied, thanks!
Are you sure you pushed this patch? I can't see it in current git.
Bart
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2014-12-16 14:21 GMT+01:00 tito :
> Hi,
> seems to me that i2cget and i2cset share a lot of code,
> couldn't they be condensed in one applet like it is
> done for example in deluser.c:
>
> int do_deluser = (ENABLE_DELUSER && (!ENABLE_DELGROUP || applet_name[3] ==
> 'u'));
>
Actually it's rather
2014-12-12 16:32 GMT+01:00 Xabier Oneca -- xOneca :
> Does this help?
>
> commit:
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/i2c/i2c-dev.c?id=0dc9a32d9a975ccd25b9f531451165c93e1c7313
>
> v2.6.14 vs. v2.6.15:
> https://git.kernel.org/cgit/linux/kernel/git/tor
2014-12-11 23:33 GMT+01:00 Xabier Oneca -- xOneca :
> Well, I think it's better with conditional compiling... This way would
> make the code easier to understand. Indeed, it doesn't make sense to
> try to open a device in a path when you known it wont be there.
> Moreover, with the current code i
2014-12-11 18:08 GMT+01:00 Rich Felker :
> How is this an improvement? It increases the code size and performs
> unnecessary and potentially harmful filesystem operations. And it's
> just covering up the "dangerous" issue rather than fixing it -- using
> mkstemp then deleting the file and reusing t
2014-12-10 13:48 GMT+01:00 walter harms :
> instead of checking two possible path let the user/system decide.
> This can be done automagicly by checking the LINUX_VERSION like:
>
> #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,0) /* used in loop.c */
>
> or you can let the user decide by providing
2014-12-10 14:20 GMT+01:00 Ralf Friedl :
> I suggest ddoing it the other way, enable sendfile only for Linux.
> As it is, sendfile is enabled by default, which would automatically select
> Linux, which might select other things that don't work on non-Linux systems.
>
There are many config options
2014-12-10 11:54 GMT+01:00 walter harms :
> ok,
> but given that busybox is not used by pple it seems reasonable to check only
> for one target.
> either automatic
> #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,0) /* used in loop.c */
> or manual: let the user decide via busybox.conf
> re,
> wh
2014-12-09 18:09 GMT+01:00 walter harms :
>
> could you please explain shortly why you have a external buffer here ?
>
> you could easly do
> filename=xasprintf("/dev/i2c-%d", i2cbus);
> or
> char filename[32];
> or
This is done to pass the filename to the calling function. It isn't
used by i2cget
2014-12-10 8:47 GMT+01:00 AGATE Stephen :
> It's not difficult to use the original if you need it:
This holds true for many other tools implemented in busybox and yet
they're there. To use i2c-tools you need the whole dynamic loading
infrastructure (libi2c.so) and four or more executables.
Thanks
2014-11-27 13:22 GMT+01:00 Denys Vlasenko :
> Thanks for catching it. So you were using 4k copy buffer.
> This explains why copies were slow for you...
Yeah, I just ran make defconfig. I forgot it's configurable - regular
coreutils copy tools use 4k buffers as well.
> I pushed a tweaked patch to
2014-11-26 17:43 GMT+01:00 Denys Vlasenko :
> On Wed, Nov 26, 2014 at 4:06 PM, Denys Vlasenko
> wrote:
>> Looks like it will always allocate the copy buffer,
>> and always attempt the final read, even if sendfile worked perfectly.
>> Can this be avoided?
>
> Please review attached patch.
I can't
2014-11-26 14:05 GMT+01:00 Denys Vlasenko :
> This feels like feature creep.
This isn't really a piece of code that will be compiled by default and
affect the overall executable size.
> In what way time(1) is not sufficient?
Internal timing measures time spent on testing only, instead of
time(1)
2014-11-26 15:21 GMT+01:00 Denys Vlasenko :
> On Thu, Jul 3, 2014 at 2:27 PM, Bartosz Gołaszewski
> wrote:
>> have you read my message from June 10th in this thread? Could you
>> review the patches I attached back then? Busybox still uses ARG_MAX as
>> default, which
red to just sendfile()).
So this helper
1301 *can splice directly through a process-private pipe.
1302 *
1303 */
So apparently sendfile() is ok in this case.
Best regards,
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2014-10-20 14:10 GMT+02:00 Baruch Siach :
>
> Why depend on HTTPD, then?
>
My bad, I'll send another one.
Best regards,
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2014-07-28 11:40 GMT+02:00 Bartosz Gołaszewski :
> 2014-07-03 14:27 GMT+02:00 Bartosz Gołaszewski :
>> Hi Denys,
>>
>> have you read my message from June 10th in this thread? Could you
>> review the patches I attached back then? Busybox still uses ARG_MAX as
>&g
to look what was type of meminfo - looking at the
> memset line alone is sufficient.
>
You're right, my bad. I misread it as meminfo[].
Best regards,
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
7;:');
>
> Would be needed to fix it.
This should be memset(meminfo, 0, MI_MAX*sizeof(unsigned long))
otherwise you're taking the size of a pointer.
Best regards,
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2014-07-03 14:27 GMT+02:00 Bartosz Gołaszewski :
> Hi Denys,
>
> have you read my message from June 10th in this thread? Could you
> review the patches I attached back then? Busybox still uses ARG_MAX as
> default, which is smaller than sysconf(_SC_ARG_MAX) on every system I
>
sconf() as default.
Best regards,
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
need for code for this kind of thing.
>
> Rich
I wasn't sure of that, thanks.
If we only use the INIT_LAST macro, I propose to rename it to
INIT_FUNC and not use the priority argument for constructor attribute.
Please see the patch in attachment.
Bartosz Gołaszewski
0001-Rena
2014-06-22 16:31 GMT+02:00 Denys Vlasenko :
> Applied all patches with some editing.
> Thanks a lot!
Hi Denys,
I see, that you removed the INIT_FIRST macro from my implementation.
Are you sure llist_t *tests will always be initialized before calling
the test registering functions?
B
a to define some functions that can be used by both
> commands.
This is what I want to do next, but I'm not sure when exactly I am
going to have time for this. The problem is - xargs prepares the
arguments in the form of one long string for the read_args(
RG_MAX
737 unsigned bb_arg_max(void) FAST_FUNC;
738 #else
739 # define bb_arg_max() ((unsigned)(32 * 1024))
740 #endif
741 unsigned bb_clk_tck(void) FAST_FUNC;
Line 735 will never be compiled unless we include .
Also: I would make sysconf(_SC_ARG_MAX) the default expansion of
bb_arg_max() as it
ined - probably should add it to platform.h. I'll
resubmit the patch in the evening.
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2014-06-17 17:02 GMT+02:00 Denys Vlasenko :
>
> Please try attached patch.
Thanks!
Please take a look at two patches I've submitted. The first one
proposes to use the system defined limit for command-line length to
further reduce the number of execve() calls. If you'll accept it, then
I'll also t
27;ll resubmit the patch later.
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2014-06-12 10:05 GMT+02:00 walter harms :
> hi,
> could you give a few word why you need that feature ?
It offers a 'find | xargs ...'-like functionality. It's faster, as it
limits the number of execve() calls to one per '-exec +' parameter.
> btw:
> Why you choose a linked list ? It seems more s
2014-04-16 1:06 GMT+02:00 Bartosz Golaszewski :
> This set of patches proposes to add a simple unit-testing framework to
> Busybox.
>
> First four patches implement the actual framework and document it, while
> last two add some example test cases.
>
> Bartosz Golaszewski (6):
> unit-tests: add
can do when
I'll have some time.
Best regards,
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2014-05-29 18:56 GMT+02:00 Laurent Bercot :
>> Does this provided a noticeable performance increase? Do you have
>> benchmarks?
I've checked the times just by looking up all the applets in a loop
and measuring the time using gettimeofday() - the results are: ~220
microseconds for bsearch and ~150
> unit-tests: add two example test cases
> unit-tests: modify the existing strrstr test code to use the unit-test
> framework
>
Any comments/reviews?
Best regards,
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2014-04-02 13:52 GMT+02:00 Mason :
>
> Would it be possible to define MIN and MAX as inline functions?
>
> static inline int MIN(int a, int b) { return a < b ? a : b; }
Macros accept any comparable type, so you would have to have specific
functions for int, unsigned, double etc.
2014-04-02 12:07
> I have such a directory on my systems (/mnt/tmpfs, 0755, with /tmp
> actually being a symlink to /mnt/tmpfs/tmp, 1777). Some distributions
> use an initramfs and create /lib/rw (ugh). Unfortunately, nothing is
> standardized... Maybe we could use /dev ? Ha ha, only serious.
Is it necessary to us
2014-02-07 Denys Vlasenko :
> On Thu, Feb 6, 2014 at 9:34 PM, Bartosz Golaszewski
> wrote:
>> Pass the REG_NOTBOL flag to subsequent regexec calls.
>
> Which use case it fixes?
> I would like to add it to testsuite.
It's already there:
testing "grep -w ^str doesn't match str not at the beginning
2014-02-07 walter harms :
> Hi Bartosz Golaszewski
>
> I do not like those !ENABLE_EXTRA_COMPAT
> since the human brain makes a lot of problems with reverse logic.
> would you mind to change that here ?
>
>
> re,
> wh
Hi Walter,
not all of them can be changed e.g.:
#if !ENABLE_EXTRA_COMPAT
2014/1/18 Karl Godt :
> Bartosz Golaszewski (6):
> grep: don't bail out on first mismatch if '-w' option is set
>
> seems not to work as it should, it seems to grep '1' , '11' , '111' , ...
> if
> word is '1' , probably because it is a number .
I didn't check, but it might have been fixed by:
o be the culprit after initial look - but have not tested.
>
> Would you Bartosz care to look at it?
>
> Thanks,
> Timo
Hi Timo,
I have submitted a patch fixing this issue.
--
Pozdrawiam,
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
way - unit-tests suite can be just in the form of a new applet with
tests distributed among the relevant files and compiled out if this
applet is not selected.
--
Best regards,
Bartosz Gołaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2013/11/5 Bartosz Gołaszewski :
> Anyway, what do you think about the unit tests then?
Any comments?
Best regards,
Bartosz Golaszewski
___
busybox mailing list
busybox@busybox.net
http://lists.busybox.net/mailman/listinfo/busybox
2013/11/5 Bernd Petrovitsch :
> []
>
> "dbus" probably started the same way. SCNR
>
> Bernd
Maybe, but my goal is to have messages, methods and signals, simple
credentials verification, an easy-to-use client API in C (and even
easier in C++ and Python), simple monitoring and contr
2013/11/4 John Spencer :
> Bartosz Gołaszewski wrote:
>>
>> Hi!
>>
>> While working on my small free-time project - a compact message bus
>> for embedded systems (think small dbus)
>
>
> a good idea. it already exists though: http://unixbus.org/ubus/
>
Hi!
While working on my small free-time project - a compact message bus
for embedded systems (think small dbus) - I wrote a simple unit
testing framework.
The project is hosted on github: https://github.com/Zajcev/busybus
(BTW - is it ok to use the name 'busybus'?)
The test framework and some in
2013/10/25 Alexander Kriegisch :
> Actually it was not empty. Whoever owns the list might want to check why my
> message was mutilated. What I had written was:
>
/etc/profile
$HOME/.profile
See http://linux.die.net/man/1/ash
>
> And Raja who got the message CC sent me a private
2013/7/30 Denys Vlasenko :
> bloat-o-meter looks wrong. It should show something like
>
> function old new delta
> visible- 70 +70
> do_display 431
2013/5/16 Bartosz Gołaszewski :
> 2013/5/15 Denys Vlasenko :
>> On Sun, May 12, 2013 at 11:48 PM, Bartosz Gołaszewski
>> wrote:
>>> I have added support for '-v' argument for gzip, gunzip, bzip2 and
>>> bunzip2. The change is only available if ENABLE_D
2013/5/15 Denys Vlasenko :
> On Sun, May 12, 2013 at 11:48 PM, Bartosz Gołaszewski
> wrote:
>> I have added support for '-v' argument for gzip, gunzip, bzip2 and
>> bunzip2. The change is only available if ENABLE_DESKTOP is set.
>>
>> Currently there
1 - 100 of 103 matches
Mail list logo
