200, Maciej Fijalkowski wrote:
>> This should fix the issues presented.
>>
>> There is one problem which is that the hash in test-rebase-detach
>> changes. The way I see it is just an RNG phase-shift, but it might be
>> a real bug. Some actual input will be appreciated.
# HG changeset patch
# User Maciej Fijalkowski
# Date 1473680234 -7200
# Mon Sep 12 13:37:14 2016 +0200
# Node ID 2c852d298fbf87c5bc0ad7b65563212169915ab3
# Parent df05c43bd1e64f1620d0b2e502f4603c1e5a8341
lazymanifest: write a more efficient, pypy friendly version of lazymanifest
diff
cool, thanks guys!
I've sent a new version that should pass all the tests.
On Sun, Sep 25, 2016 at 1:03 PM, Jun Wu wrote:
> Excerpts from Maciej Fijalkowski's message of 2016-09-25 09:13:29 +0200:
>> a proper debugger which is surprisingly hard to use in the case of
>> mercurial tests
>
> If you
The one that comes with os x or homebrew I'll update to a newer one I
promise
On 28 Sep 2016 2:09 PM, "Pierre-Yves David"
wrote:
>
>
> On 09/17/2016 01:45 PM, Maciej Fijalkowski wrote:
>
>> ok, I think at the end it is that:
>>
>> received listke
# HG changeset patch
# User Maciej Fijalkowski
# Date 1473680234 -7200
# Mon Sep 12 13:37:14 2016 +0200
# Node ID c770219dc4c253d7cd82519ce3c74438bb2829d3
# Parent df05c43bd1e64f1620d0b2e502f4603c1e5a8341
lazymanifest: write a more efficient, pypy friendly version of lazymanifest
diff
Fixed, it's error reporting and making sure we truncate the 22 length hash
On Fri, Sep 30, 2016 at 12:06 AM, Augie Fackler wrote:
> On Wed, Sep 28, 2016 at 01:47:32PM +0200, Maciej Fijalkowski wrote:
>> # HG changeset patch
>> # User Maciej Fijalkowski
>> # Date 1473
Preview of my work on speeding up mercurial on pypy. The more changing
benchmarks (commit, rebase etc.) are not included since they need more
plumbing (working on it).
The benchmarks are run multiple times, to simulate the effects of chg.
This *also* populates caches etc, so I'm excluding initial
# HG changeset patch
# User Maciej Fijalkowski
# Date 1476428549 -7200
# Fri Oct 14 09:02:29 2016 +0200
# Node ID 4e80a66124279e235dec7ae8f58c0a14b0b137bf
# Parent c770219dc4c253d7cd82519ce3c74438bb2829d3
osutil: implement listdir for linux
diff --git a/mercurial/pure/osutil.py b/mercurial
FYI This is the violation of C API, explicit assignment would be
probably better (at least it's CLEAR that you're doing something
dodgy)
On Thu, Oct 13, 2016 at 9:43 PM, Gregory Szorc wrote:
> # HG changeset patch
> # User Gregory Szorc
> # Date 1475939263 -7200
> # Sat Oct 08 17:07:43 2016
ion for
quite a while, if you ask me
On Sun, Oct 16, 2016 at 6:55 PM, Gregory Szorc wrote:
> On Fri, Oct 14, 2016 at 12:02 AM, Maciej Fijalkowski
> wrote:
>>
>> # HG changeset patch
>> # User Maciej Fijalkowski
>> # Date 1476428549 -7200
>>
On Fri, Aug 5, 2016 at 3:54 PM, Yuya Nishihara wrote:
> On Mon, 25 Jul 2016 15:10:58 +0200, Maciej Fijalkowski wrote:
>> # HG changeset patch
>> # User Maciej Fijalkowski
>> # Date 1469201285 -7200
>> # Fri Jul 22 17:28:05 2016 +0200
>> # Node ID 8e2d1b2
# HG changeset patch
# User Maciej Fijalkowski
# Date 1469708228 -7200
# Thu Jul 28 14:17:08 2016 +0200
# Node ID 2f4cd91aaa9a2ad1e4de097a1ec7a59127d51e20
# Parent 54ce0308751b2772dbfbe94004a139d53f292a3f
bdiff: implement cffi version of blocks
diff --git a/mercurial/pure/bdiff.py b
# HG changeset patch
# User Maciej Fijalkowski
# Date 1469708281 -7200
# Thu Jul 28 14:18:01 2016 +0200
# Node ID 300f14ea21432face8d7e6cdcf92ba9d2f1f92dc
# Parent 2f4cd91aaa9a2ad1e4de097a1ec7a59127d51e20
performance: disable workaround for an old bug of Python
Since disabling the gc does
Well, seems the comment is out of date. I know the issue - the GC got
triggered every X objects in 2.6, which caused O(n^2) performance when
allocating lots of objects. These days (since 2.7) it adapts the
threshold, which means it always amortizes to O(n).
Either way, I don't care if we disable i
On Thu, Aug 11, 2016 at 10:13 PM, Pierre-Yves David
wrote:
>
>
> On 08/11/2016 01:22 PM, Yuya Nishihara wrote:
>>
>> On Wed, 10 Aug 2016 16:45:35 +0200, Maciej Fijalkowski wrote:
>>>
>>> # HG changeset patch
>>> # User Maciej Fijalkowski
>>&g
# HG changeset patch
# User Maciej Fijalkowski
# Date 1471726818 -7200
# Sat Aug 20 23:00:18 2016 +0200
# Node ID 21b2401d468d6b24c1658468e4fc5ce8744f925b
# Parent 300f14ea21432face8d7e6cdcf92ba9d2f1f92dc
manifest: write a more efficient version of lazymanifest, in pure python
Questions
On Sun, Aug 14, 2016 at 4:14 AM, Yuya Nishihara wrote:
> On Fri, 12 Aug 2016 23:42:38 +0200, Maciej Fijalkowski wrote:
>> Well, seems the comment is out of date. I know the issue - the GC got
>> triggered every X objects in 2.6, which caused O(n^2) performance when
>> alloc
# HG changeset patch
# User Maciej Fijalkowski
# Date 1471727118 -7200
# Sat Aug 20 23:05:18 2016 +0200
# Node ID 17bb55a4276ea6d85e10c128fa2b83c7e170e9de
# Parent 21b2401d468d6b24c1658468e4fc5ce8744f925b
osutil: fix the bug on OS X when we return more in listdir
The pointer arithmetic
# HG changeset patch
# User Maciej Fijalkowski
# Date 1471727161 -7200
# Sat Aug 20 23:06:01 2016 +0200
# Node ID 02c99939ee6a6157de657bf8fffb7b998cd827b0
# Parent 17bb55a4276ea6d85e10c128fa2b83c7e170e9de
bdiff: implement cffi version of bdiff
diff --git a/mercurial/pure/bdiff.py b
# HG changeset patch
# User Maciej Fijalkowski
# Date 1469452252 -7200
# Mon Jul 25 15:10:52 2016 +0200
# Node ID 7fa693de90d23c034e36e0fe1b9552feb4eca839
# Parent 8e2d1b2d5248620ff4ccfcb61e1b383cf0e4e204
mpatch: write a cffi version of mpatch.patches
diff --git a/mercurial/pure/mpatch.py
On Sun, Aug 21, 2016 at 5:19 PM, Yuya Nishihara wrote:
> On Sat, 20 Aug 2016 23:04:06 +0200, Maciej Fijalkowski wrote:
>> On Sun, Aug 14, 2016 at 4:14 AM, Yuya Nishihara wrote:
>> > On Fri, 12 Aug 2016 23:42:38 +0200, Maciej Fijalkowski wrote:
>> >> Well, seems the
# HG changeset patch
# User fijal
# Date 1473321829 -7200
# Thu Sep 08 10:03:49 2016 +0200
# Node ID e812104013b1281b24dff50dd902897d24b0e29f
# Parent 64e432c2740087b7c042346f55af9512ead0f849
mpatch: add missing file
diff --git a/setup_mpatch_cffi.py b/setup_mpatch_cffi.py
new file mode 1006
ups, it seems this has been committed (?)
On Thu, Sep 8, 2016 at 10:04 AM, Maciej Fijalkowski wrote:
> # HG changeset patch
> # User fijal
> # Date 1473321829 -7200
> # Thu Sep 08 10:03:49 2016 +0200
> # Node ID e812104013b1281b24dff50dd902897d24b0
# HG changeset patch
# User fijal
# Date 1473323878 -7200
# Thu Sep 08 10:37:58 2016 +0200
# Node ID 8db87163c95d79da03644e3850551f3bfa802262
# Parent e20945831e4f9362d2e113bafe527e2a0554f4b4
osutil: fix the declaration to work on different sizes of off_t
diff --git a/setup_osutil_cffi.py b/
# HG changeset patch
# User Maciej Fijalkowski
# Date 1473680234 -7200
# Mon Sep 12 13:37:14 2016 +0200
# Node ID a43b7edeb6d66afc5c4eab9e4c8bb35b94cbfa97
# Parent df05c43bd1e64f1620d0b2e502f4603c1e5a8341
lazymanifest: write a more efficient, pypy friendly version of lazymanifest
diff
t;>
>>
>> On 09/13/2016 12:47 PM, Pierre-Yves David wrote:
>>>
>>>
>>>
>>> On 09/12/2016 03:13 PM, Augie Fackler wrote:
>>>>
>>>> On Mon, Sep 12, 2016 at 02:36:19PM +0200, Maciej Fijalkowski wrote:
>>>>>
>&g
https://paste.pound-python.org/show/JB0RSPFycVzWS8X6C7hc/
On Sat, Sep 17, 2016 at 12:25 PM, Pierre-Yves David
wrote:
>
>
> On 09/17/2016 11:17 AM, Maciej Fijalkowski wrote:
>>
>> Hi
>>
>> When I'm trying to do that, I get that:
>>
>> https:
ok, I think at the end it is that:
received listkey for "obsolete": 16746270 bytes
and it takes a while, but there is no progress nothing which seems
like it's hanging
On Sat, Sep 17, 2016 at 1:44 PM, Maciej Fijalkowski wrote:
> https://paste.pound-python.org/show/JB0RSPFycV
managed to pull that. How do I incorporate my fixes for the problems now?
On Sat, Sep 17, 2016 at 1:45 PM, Maciej Fijalkowski wrote:
> ok, I think at the end it is that:
>
> received listkey for "obsolete": 16746270 bytes
>
> and it takes a while, but there is no pro
I have fixed the problem with the test that you listed. I'm still
completely unable to finish the test run - with -j it randomly hangs
and does nothing with no extra output
On Sat, Sep 17, 2016 at 1:47 PM, Maciej Fijalkowski wrote:
> managed to pull that. How do I incorporate my fixes
# HG changeset patch
# User Maciej Fijalkowski
# Date 1473680234 -7200
# Mon Sep 12 13:37:14 2016 +0200
# Node ID 7551f1e60b2155462d89a9571eec065e9f67debc
# Parent df05c43bd1e64f1620d0b2e502f4603c1e5a8341
lazymanifest: write a more efficient, pypy friendly version of lazymanifest
diff
I'll resend it with v2
On Sat, Sep 17, 2016 at 8:21 PM, Maciej Fijalkowski wrote:
> # HG changeset patch
> # User Maciej Fijalkowski
> # Date 1473680234 -7200
> # Mon Sep 12 13:37:14 2016 +0200
> # Node ID 7551f1e60b2155462d89a9571eec065
# HG changeset patch
# User Maciej Fijalkowski
# Date 1473680234 -7200
# Mon Sep 12 13:37:14 2016 +0200
# Node ID 7551f1e60b2155462d89a9571eec065e9f67debc
# Parent df05c43bd1e64f1620d0b2e502f4603c1e5a8341
lazymanifest: write a more efficient, pypy friendly version of lazymanifest
diff
This should fix the issues presented.
There is one problem which is that the hash in test-rebase-detach
changes. The way I see it is just an RNG phase-shift, but it might be
a real bug. Some actual input will be appreciated.
On Sat, Sep 17, 2016 at 8:22 PM, Maciej Fijalkowski wrote:
> #
34 matches
Mail list logo