You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Jeff Trawick <tr...@gmail.com> on 2011/09/08 18:44:31 UTC
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
On Thu, Sep 8, 2011 at 11:16 AM, Jeff Trawick <tr...@gmail.com> wrote:
> Here's what I have at present:
> http://people.apache.org/~trawick/2.2.20-byterange-fixes.txt
>
> (compiled-in max ranges, uses same AP_ symbol as 2.2.21 even though
> the compiled-in version isn't the same type of "DEFAULT")
>
See also http://people.apache.org/~trawick/2.2.19-byterange-fixes.txt for 2.2.19
I've tested this latter patch with current framework and a separate
test for numranges>200 with 2.2.14, 2.2.15, and 2.2.19. (2.2.14 needs
r916627, or at least the subset of it for byterange_filter.c, before
applying the patch.)
--
Born in Roswell... married an alien...
T&R of 2.0/2.2 next week?
Posted by Jim Jagielski <ji...@jaguNET.com>.
We are v. close to having 2.0 and 2.2 in a tag-able state…
I propose a T&R next week and offer to RM.
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Jeff Trawick <tr...@gmail.com>.
Can anyone formally +1 the 2.2.14 and 2.2.19 patches on this thread
for promotion to apply_to?
I put the 2.0.x patch in 2.0.x/STATUS, as the code needs to be
committed first before being made available as a patch.
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Jim Jagielski <ji...@apache.org>.
Patch applied and committed…
On Sep 9, 2011, at 9:53 AM, Jim Jagielski wrote:
> Same here…
>
> On Sep 9, 2011, at 6:54 AM, Jeff Trawick wrote:
>
>> On Fri, Sep 9, 2011 at 4:29 AM, Joe Orton <jo...@redhat.com> wrote:
>>> On Thu, Sep 08, 2011 at 05:36:06PM -0400, Jeff Trawick wrote:
>>>> static apr_status_t base10_strtoff(apr_off_t *offset, const char *buf,
>>>> char **endptr)
>>>> {
>>>> const char *last;
>>>>
>>>> *offset = apr_atoi64(buf);
>>>
>>> I think this needs to DTRT with a 32-bit off_t.
>>>
>>> How about this?
>>>
>>> static apr_status_t strtoff(apr_off_t *offset, const char *nptr,
>>> char **endptr, int base)
>>> {
>>> errno = 0;
>>> if (sizeof(apr_off_t) == 4) {
>>> *offset = strtol(nptr, endptr, base);
>>> }
>>> else {
>>> *offset = apr_strtoi64(ptr, endptr, base);
>>> }
>>> return APR_FROM_OS_ERROR(errno);
>>> }
>>
>> works for me
>>
>> http://people.apache.org/~trawick/draft3-2.0.64-byterange-fixes.txt
>>
>> Thanks!
>>
>
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Jim Jagielski <ji...@apache.org>.
Same here…
On Sep 9, 2011, at 6:54 AM, Jeff Trawick wrote:
> On Fri, Sep 9, 2011 at 4:29 AM, Joe Orton <jo...@redhat.com> wrote:
>> On Thu, Sep 08, 2011 at 05:36:06PM -0400, Jeff Trawick wrote:
>>> static apr_status_t base10_strtoff(apr_off_t *offset, const char *buf,
>>> char **endptr)
>>> {
>>> const char *last;
>>>
>>> *offset = apr_atoi64(buf);
>>
>> I think this needs to DTRT with a 32-bit off_t.
>>
>> How about this?
>>
>> static apr_status_t strtoff(apr_off_t *offset, const char *nptr,
>> char **endptr, int base)
>> {
>> errno = 0;
>> if (sizeof(apr_off_t) == 4) {
>> *offset = strtol(nptr, endptr, base);
>> }
>> else {
>> *offset = apr_strtoi64(ptr, endptr, base);
>> }
>> return APR_FROM_OS_ERROR(errno);
>> }
>
> works for me
>
> http://people.apache.org/~trawick/draft3-2.0.64-byterange-fixes.txt
>
> Thanks!
>
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Jeff Trawick <tr...@gmail.com>.
On Fri, Sep 9, 2011 at 4:29 AM, Joe Orton <jo...@redhat.com> wrote:
> On Thu, Sep 08, 2011 at 05:36:06PM -0400, Jeff Trawick wrote:
>> static apr_status_t base10_strtoff(apr_off_t *offset, const char *buf,
>> char **endptr)
>> {
>> const char *last;
>>
>> *offset = apr_atoi64(buf);
>
> I think this needs to DTRT with a 32-bit off_t.
>
> How about this?
>
> static apr_status_t strtoff(apr_off_t *offset, const char *nptr,
> char **endptr, int base)
> {
> errno = 0;
> if (sizeof(apr_off_t) == 4) {
> *offset = strtol(nptr, endptr, base);
> }
> else {
> *offset = apr_strtoi64(ptr, endptr, base);
> }
> return APR_FROM_OS_ERROR(errno);
> }
works for me
http://people.apache.org/~trawick/draft3-2.0.64-byterange-fixes.txt
Thanks!
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Joe Orton <jo...@redhat.com>.
On Thu, Sep 08, 2011 at 05:36:06PM -0400, Jeff Trawick wrote:
> static apr_status_t base10_strtoff(apr_off_t *offset, const char *buf,
> char **endptr)
> {
> const char *last;
>
> *offset = apr_atoi64(buf);
I think this needs to DTRT with a 32-bit off_t.
How about this?
static apr_status_t strtoff(apr_off_t *offset, const char *nptr,
char **endptr, int base)
{
errno = 0;
if (sizeof(apr_off_t) == 4) {
*offset = strtol(nptr, endptr, base);
}
else {
*offset = apr_strtoi64(ptr, endptr, base);
}
return APR_FROM_OS_ERROR(errno);
}
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Jeff Trawick <tr...@gmail.com>.
On Thu, Sep 8, 2011 at 5:36 PM, Jeff Trawick <tr...@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 3:16 PM, Jeff Trawick <tr...@gmail.com> wrote:
>> On Thu, Sep 8, 2011 at 3:08 PM, William A. Rowe Jr. <wr...@rowe-clan.net> wrote:
>>> On 9/8/2011 1:45 PM, Jeff Trawick wrote:
>>>>>
>>>>> On the 2.0 side, nothings changed since 2.0.55 that should break the patch.
>>>>
>>>> BTW, do any of us have an updated 2.0 patch to reflect the important
>>>> changes since last weekend? If not, I'll need to work on thatin the
>>>> short term.
>>>
>>> I'm happy to help late this afternoon once I've worked out a wholly
>>> unrelated fix, and 2.2.x looks ready to tag. If you wanted sooner,
>>> please feel free to jump ahead of me.
>>
>> I'm starting that now. Hopefully I can make some progress (or finish)
>> before I have to switch gears in an hour or so.
>
> This probably fails due to my context switching, but it is a start :)
>
> http://people.apache.org/~trawick/draft-2.0.64-byterange-fixes.txt
with a whitespace and comment fix that make it easier to compare
2.0.64+patch with 2.2.19+patch:
http://people.apache.org/~trawick/draft2-2.0.64-byterange-fixes.txt
> This needs some check for epic fail...
>
> /* simple version for consistency with 2.2 code which uses
> * apr_strtoff()
> *
> * Base is always 10, leading whitespace and signs not allowed,
> * per restrictions in the caller.
> */
> static apr_status_t base10_strtoff(apr_off_t *offset, const char *buf,
> char **endptr)
> {
> const char *last;
>
> *offset = apr_atoi64(buf);
> last = buf;
> while (isdigit(*last)) {
> ++last;
> }
> *endptr = (char *)last;
> if (last == buf) {
> return APR_EINVAL;
> }
> return APR_SUCCESS;
> }
>
>
>
> --
> Born in Roswell... married an alien...
>
--
Born in Roswell... married an alien...
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Jeff Trawick <tr...@gmail.com>.
On Thu, Sep 8, 2011 at 3:16 PM, Jeff Trawick <tr...@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 3:08 PM, William A. Rowe Jr. <wr...@rowe-clan.net> wrote:
>> On 9/8/2011 1:45 PM, Jeff Trawick wrote:
>>>>
>>>> On the 2.0 side, nothings changed since 2.0.55 that should break the patch.
>>>
>>> BTW, do any of us have an updated 2.0 patch to reflect the important
>>> changes since last weekend? If not, I'll need to work on thatin the
>>> short term.
>>
>> I'm happy to help late this afternoon once I've worked out a wholly
>> unrelated fix, and 2.2.x looks ready to tag. If you wanted sooner,
>> please feel free to jump ahead of me.
>
> I'm starting that now. Hopefully I can make some progress (or finish)
> before I have to switch gears in an hour or so.
This probably fails due to my context switching, but it is a start :)
http://people.apache.org/~trawick/draft-2.0.64-byterange-fixes.txt
This needs some check for epic fail...
/* simple version for consistency with 2.2 code which uses
* apr_strtoff()
*
* Base is always 10, leading whitespace and signs not allowed,
* per restrictions in the caller.
*/
static apr_status_t base10_strtoff(apr_off_t *offset, const char *buf,
char **endptr)
{
const char *last;
*offset = apr_atoi64(buf);
last = buf;
while (isdigit(*last)) {
++last;
}
*endptr = (char *)last;
if (last == buf) {
return APR_EINVAL;
}
return APR_SUCCESS;
}
--
Born in Roswell... married an alien...
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Jeff Trawick <tr...@gmail.com>.
On Thu, Sep 8, 2011 at 3:08 PM, William A. Rowe Jr. <wr...@rowe-clan.net> wrote:
> On 9/8/2011 1:45 PM, Jeff Trawick wrote:
>>>
>>> On the 2.0 side, nothings changed since 2.0.55 that should break the patch.
>>
>> BTW, do any of us have an updated 2.0 patch to reflect the important
>> changes since last weekend? If not, I'll need to work on thatin the
>> short term.
>
> I'm happy to help late this afternoon once I've worked out a wholly
> unrelated fix, and 2.2.x looks ready to tag. If you wanted sooner,
> please feel free to jump ahead of me.
I'm starting that now. Hopefully I can make some progress (or finish)
before I have to switch gears in an hour or so.
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by "William A. Rowe Jr." <wr...@rowe-clan.net>.
On 9/8/2011 1:45 PM, Jeff Trawick wrote:
>>
>> On the 2.0 side, nothings changed since 2.0.55 that should break the patch.
>
> BTW, do any of us have an updated 2.0 patch to reflect the important
> changes since last weekend? If not, I'll need to work on thatin the
> short term.
I'm happy to help late this afternoon once I've worked out a wholly
unrelated fix, and 2.2.x looks ready to tag. If you wanted sooner,
please feel free to jump ahead of me.
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Jeff Trawick <tr...@gmail.com>.
On Thu, Sep 8, 2011 at 2:45 PM, Jeff Trawick <tr...@gmail.com> wrote:
> On Thu, Sep 8, 2011 at 2:17 PM, William A. Rowe Jr. <wr...@rowe-clan.net> wrote:
>> On 9/8/2011 11:44 AM, Jeff Trawick wrote:
>>> On Thu, Sep 8, 2011 at 11:16 AM, Jeff Trawick <tr...@gmail.com> wrote:
>>>> Here's what I have at present:
>>>> http://people.apache.org/~trawick/2.2.20-byterange-fixes.txt
>>>>
>>>> (compiled-in max ranges, uses same AP_ symbol as 2.2.21 even though
>>>> the compiled-in version isn't the same type of "DEFAULT")
>>>>
>>>
>>> See also http://people.apache.org/~trawick/2.2.19-byterange-fixes.txt for 2.2.19
>>
>> Lovely! I think this file is far more useful, since anyone who already
>> adopted .20 should be in a good position to adopt .21, while someone
>> sitting back on .15 (or .9, gasp) might have a much bigger headache.
>>
>>> I've tested this latter patch with current framework and a separate
>>> test for numranges>200 with 2.2.14, 2.2.15, and 2.2.19. (2.2.14 needs
>>> r916627, or at least the subset of it for byterange_filter.c, before
>>> applying the patch.)
>>
>> On the 2.0 side, nothings changed since 2.0.55 that should break the patch.
>
> BTW, do any of us have an updated 2.0 patch to reflect the important
> changes since last weekend? If not, I'll need to work on thatin the
> short term.
>
>> On the 2.2 side, we might want to publish an apply_to_2.2.14 and 2.2.19,
>> just given that 2.2.9+ (which the .14 applies to) has only now reached 3
>> years old. I see no real value in working out a <2.2.8 patch /shrug
>
> works for me
>
> I'll get a 2.2.14-no-caveats patch together (i.e., no prepatch of
> r916627 required).
http://people.apache.org/~trawick/2.2.14-byterange-fixes.txt
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by Jeff Trawick <tr...@gmail.com>.
On Thu, Sep 8, 2011 at 2:17 PM, William A. Rowe Jr. <wr...@rowe-clan.net> wrote:
> On 9/8/2011 11:44 AM, Jeff Trawick wrote:
>> On Thu, Sep 8, 2011 at 11:16 AM, Jeff Trawick <tr...@gmail.com> wrote:
>>> Here's what I have at present:
>>> http://people.apache.org/~trawick/2.2.20-byterange-fixes.txt
>>>
>>> (compiled-in max ranges, uses same AP_ symbol as 2.2.21 even though
>>> the compiled-in version isn't the same type of "DEFAULT")
>>>
>>
>> See also http://people.apache.org/~trawick/2.2.19-byterange-fixes.txt for 2.2.19
>
> Lovely! I think this file is far more useful, since anyone who already
> adopted .20 should be in a good position to adopt .21, while someone
> sitting back on .15 (or .9, gasp) might have a much bigger headache.
>
>> I've tested this latter patch with current framework and a separate
>> test for numranges>200 with 2.2.14, 2.2.15, and 2.2.19. (2.2.14 needs
>> r916627, or at least the subset of it for byterange_filter.c, before
>> applying the patch.)
>
> On the 2.0 side, nothings changed since 2.0.55 that should break the patch.
BTW, do any of us have an updated 2.0 patch to reflect the important
changes since last weekend? If not, I'll need to work on thatin the
short term.
> On the 2.2 side, we might want to publish an apply_to_2.2.14 and 2.2.19,
> just given that 2.2.9+ (which the .14 applies to) has only now reached 3
> years old. I see no real value in working out a <2.2.8 patch /shrug
works for me
I'll get a 2.2.14-no-caveats patch together (i.e., no prepatch of
r916627 required).
Re: [PATCH] byterange patches for 2.2.19 and 2.2.20
Posted by "William A. Rowe Jr." <wr...@rowe-clan.net>.
On 9/8/2011 11:44 AM, Jeff Trawick wrote:
> On Thu, Sep 8, 2011 at 11:16 AM, Jeff Trawick <tr...@gmail.com> wrote:
>> Here's what I have at present:
>> http://people.apache.org/~trawick/2.2.20-byterange-fixes.txt
>>
>> (compiled-in max ranges, uses same AP_ symbol as 2.2.21 even though
>> the compiled-in version isn't the same type of "DEFAULT")
>>
>
> See also http://people.apache.org/~trawick/2.2.19-byterange-fixes.txt for 2.2.19
Lovely! I think this file is far more useful, since anyone who already
adopted .20 should be in a good position to adopt .21, while someone
sitting back on .15 (or .9, gasp) might have a much bigger headache.
> I've tested this latter patch with current framework and a separate
> test for numranges>200 with 2.2.14, 2.2.15, and 2.2.19. (2.2.14 needs
> r916627, or at least the subset of it for byterange_filter.c, before
> applying the patch.)
On the 2.0 side, nothings changed since 2.0.55 that should break the patch.
On the 2.2 side, we might want to publish an apply_to_2.2.14 and 2.2.19,
just given that 2.2.9+ (which the .14 applies to) has only now reached 3
years old. I see no real value in working out a <2.2.8 patch /shrug