Patchwork Workflow improvement

Patchwork Workflow improvement

[Holger Hans Peter Freyther]

[-- Attachment #1: Type: Text/Plain, Size: 589 bytes --]

Hi All,

right now I'm writing a Qt 4.6 based application that might help us a bit with 
the patchwork queue.

The idea of the script is to have two things at the end:
	1.) a branch with all the patches that applied and that can be published
	2.) a list of patches that don't apply.


What people are supposed to:

1.) Cherry pick from the branch
2.) Send updated patches, close the reports on patchwork...



The current version is able to download the information of all pending 
patches, the next thing is to do is to apply them and keep track of the 
result...



[-- Attachment #2: main.cc --]
[-- Type: text/x-c++src, Size: 3874 bytes --]

/*
 * Copyright (C) 2010 Holger Hans Peter Freyther
 *
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY APPLE COMPUTER, INC. ``AS IS'' AND ANY
 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
 * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL APPLE COMPUTER, INC. OR
 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
 * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

#include <QApplication>
#include <QWebFrame>
#include <QWebElement>
#include <QWebPage>

class Worker : public QObject {
    Q_OBJECT
public:
    Worker()
        : m_currentPage(1)
    {
        m_page = new QWebPage(this);
        connect(m_page, SIGNAL(loadFinished(bool)), SLOT(loadFinished(bool)));
    }

    void start()
    {
        qWarning("Trying to download page %d from patchwork", m_currentPage);
        QUrl url("http://patchwork.openembedded.org/project/openembedded/list/?order=date&page=" + QByteArray::number(m_currentPage));
        m_page->mainFrame()->load(url);
    }

    void startNext()
    {
        ++m_currentPage;
        start();
    }

    void handleCurrentPatch(QWebElement patch)
    {
        QWebElement link = patch.findFirst("a");
        QList<QWebElement> table = patch.findAll("td").toList();
        QString date, author, state;

        if (table.size() == 5) {
            date = table[1].toPlainText();
            author = table[2].toPlainText();
            state = table[4].toPlainText();
        }

        qWarning("\tDate: %s Author: '%s' State: %s '%s'", qPrintable(date), qPrintable(author), qPrintable(state), qPrintable(link.toPlainText()));
    }

    // query for all patches..
    void handleCurrentPage()
    {
        QWebElementCollection patches = m_page->mainFrame()->findAllElements("tr[id*=patch_row]");
        foreach(QWebElement element, patches)
            handleCurrentPatch(element);
    }

public Q_SLOTS:
    void loadFinished(bool result)
    {
        // Verify that we got the page we wanted
        QWebElement currentPage = m_page->mainFrame()->findFirstElement("span[title='Current Page']");
        if (currentPage.isNull()) {
            qWarning("Failed to find the current page.");
            return QApplication::exit(-1);
        }
        int currentPageNo = currentPage.toPlainText().toInt();
        if (currentPageNo != m_currentPage) {
            qWarning("The end was reached.");
            return QApplication::exit(0);
        }

        qWarning("Finished the download of page %d with status: %d", m_currentPage, result);
        handleCurrentPage();

        startNext();
    }

private:
    QWebPage* m_page;
    int m_currentPage;
};

int main(int argc, char **argv)
{
    QApplication app(argc, argv);

    Worker worker;
    worker.start();
    app.exec();
}

#include "main.moc"

Re: Patchwork Workflow improvement

[Holger Hans Peter Freyther]

On Saturday 23 January 2010 09:43:24 Holger Hans Peter Freyther wrote:

> The current version is able to download the information of all pending
> patches, the next thing is to do is to apply them and keep track of the
> result...

Ofcourse there is pwclient (which I couldn't find right now when I started 
this) but it will not be able to help us on the main purpose..

Re: Patchwork Workflow improvement

[Holger Hans Peter Freyther]

On Saturday 23 January 2010 09:43:24 Holger Hans Peter Freyther wrote:
> Hi All,

these urls fail to apply. It would be cool if people would help with the 
following randomized algorithm:

Pick a given number between 0 to 138 and then work on a block of 10 times.

1.) Open the URL in the browser
2.) Check if the patch was applied
2.1) If it was applied, close the report
2.2) If it was not applied, and the patch makes sense, send a follow
       up mail to the author and defer the current patch
2.3) If it was not applied, and the patch does not make sense,
       just close the bug as rejected..

please help


# 139 patches failed to apply.
http://patchwork.openembedded.org/patch/454
http://patchwork.openembedded.org/patch/650
http://patchwork.openembedded.org/patch/651
http://patchwork.openembedded.org/patch/691
http://patchwork.openembedded.org/patch/714
http://patchwork.openembedded.org/patch/727
http://patchwork.openembedded.org/patch/729
http://patchwork.openembedded.org/patch/740
http://patchwork.openembedded.org/patch/743
http://patchwork.openembedded.org/patch/746
http://patchwork.openembedded.org/patch/750
http://patchwork.openembedded.org/patch/753
http://patchwork.openembedded.org/patch/770
http://patchwork.openembedded.org/patch/773
http://patchwork.openembedded.org/patch/890
http://patchwork.openembedded.org/patch/893
http://patchwork.openembedded.org/patch/902
http://patchwork.openembedded.org/patch/940
http://patchwork.openembedded.org/patch/942
http://patchwork.openembedded.org/patch/949
http://patchwork.openembedded.org/patch/959
http://patchwork.openembedded.org/patch/960
http://patchwork.openembedded.org/patch/972
http://patchwork.openembedded.org/patch/986
http://patchwork.openembedded.org/patch/988
http://patchwork.openembedded.org/patch/989
http://patchwork.openembedded.org/patch/994
http://patchwork.openembedded.org/patch/1000
http://patchwork.openembedded.org/patch/1004
http://patchwork.openembedded.org/patch/1005
http://patchwork.openembedded.org/patch/1007
http://patchwork.openembedded.org/patch/1008
http://patchwork.openembedded.org/patch/1016
http://patchwork.openembedded.org/patch/1020
http://patchwork.openembedded.org/patch/1021
http://patchwork.openembedded.org/patch/1022
http://patchwork.openembedded.org/patch/1023
http://patchwork.openembedded.org/patch/1024
http://patchwork.openembedded.org/patch/1030
http://patchwork.openembedded.org/patch/1043
http://patchwork.openembedded.org/patch/1036
http://patchwork.openembedded.org/patch/1046
http://patchwork.openembedded.org/patch/1041
http://patchwork.openembedded.org/patch/1048
http://patchwork.openembedded.org/patch/1056
http://patchwork.openembedded.org/patch/1064
http://patchwork.openembedded.org/patch/1066
http://patchwork.openembedded.org/patch/1067
http://patchwork.openembedded.org/patch/1074
http://patchwork.openembedded.org/patch/1078
http://patchwork.openembedded.org/patch/1082
http://patchwork.openembedded.org/patch/1083
http://patchwork.openembedded.org/patch/1089

Re: Patchwork Workflow improvement

[Holger Hans Peter Freyther]

[-- Attachment #1: Type: Text/Plain, Size: 6984 bytes --]

On Saturday 23 January 2010 09:43:24 Holger Hans Peter Freyther wrote:
> Hi All,


I have pushed "holger/staging-branch" to the server. It does contain all 
patches from patchwork that are still applying. Please go through the branch 
and cherry-pick or follow up on the emails.. I also have a new version of the 
program I use attached... I will probably switch it to use pwclient in the 
future.

thanks

The current list of failures:
# 142 patches failed to apply.
http://patchwork.openembedded.org/patch/454
http://patchwork.openembedded.org/patch/650
http://patchwork.openembedded.org/patch/651
http://patchwork.openembedded.org/patch/691
http://patchwork.openembedded.org/patch/714
http://patchwork.openembedded.org/patch/727
http://patchwork.openembedded.org/patch/729
http://patchwork.openembedded.org/patch/740
http://patchwork.openembedded.org/patch/743
http://patchwork.openembedded.org/patch/746
http://patchwork.openembedded.org/patch/747
http://patchwork.openembedded.org/patch/750
http://patchwork.openembedded.org/patch/753
http://patchwork.openembedded.org/patch/770
http://patchwork.openembedded.org/patch/773
http://patchwork.openembedded.org/patch/881
http://patchwork.openembedded.org/patch/890
http://patchwork.openembedded.org/patch/893
http://patchwork.openembedded.org/patch/902
http://patchwork.openembedded.org/patch/940
http://patchwork.openembedded.org/patch/942
http://patchwork.openembedded.org/patch/949
http://patchwork.openembedded.org/patch/950
http://patchwork.openembedded.org/patch/959
http://patchwork.openembedded.org/patch/960
http://patchwork.openembedded.org/patch/972
http://patchwork.openembedded.org/patch/986
http://patchwork.openembedded.org/patch/988
http://patchwork.openembedded.org/patch/989
http://patchwork.openembedded.org/patch/994
http://patchwork.openembedded.org/patch/1000
http://patchwork.openembedded.org/patch/1004
http://patchwork.openembedded.org/patch/1005
http://patchwork.openembedded.org/patch/1007
http://patchwork.openembedded.org/patch/1008
http://patchwork.openembedded.org/patch/1016
http://patchwork.openembedded.org/patch/1020
http://patchwork.openembedded.org/patch/1021
http://patchwork.openembedded.org/patch/1022
http://patchwork.openembedded.org/patch/1023
http://patchwork.openembedded.org/patch/1024
http://patchwork.openembedded.org/patch/1030
http://patchwork.openembedded.org/patch/1043
http://patchwork.openembedded.org/patch/1036
http://patchwork.openembedded.org/patch/1046
http://patchwork.openembedded.org/patch/1041
http://patchwork.openembedded.org/patch/1048
http://patchwork.openembedded.org/patch/1056
http://patchwork.openembedded.org/patch/1064
http://patchwork.openembedded.org/patch/1066
http://patchwork.openembedded.org/patch/1067
http://patchwork.openembedded.org/patch/1074
http://patchwork.openembedded.org/patch/1078
http://patchwork.openembedded.org/patch/1082
http://patchwork.openembedded.org/patch/1083
http://patchwork.openembedded.org/patch/1089
http://patchwork.openembedded.org/patch/1090
http://patchwork.openembedded.org/patch/1092
http://patchwork.openembedded.org/patch/1093
http://patchwork.openembedded.org/patch/1095
http://patchwork.openembedded.org/patch/1098
http://patchwork.openembedded.org/patch/1115
http://patchwork.openembedded.org/patch/1119
http://patchwork.openembedded.org/patch/1122
http://patchwork.openembedded.org/patch/1135
http://patchwork.openembedded.org/patch/1144
http://patchwork.openembedded.org/patch/1149
http://patchwork.openembedded.org/patch/1150
http://patchwork.openembedded.org/patch/1152
http://patchwork.openembedded.org/patch/1153
http://patchwork.openembedded.org/patch/1154
http://patchwork.openembedded.org/patch/1157
http://patchwork.openembedded.org/patch/1161
http://patchwork.openembedded.org/patch/1162
http://patchwork.openembedded.org/patch/1163
http://patchwork.openembedded.org/patch/1164
http://patchwork.openembedded.org/patch/1172
http://patchwork.openembedded.org/patch/1176
http://patchwork.openembedded.org/patch/1175
http://patchwork.openembedded.org/patch/1173
http://patchwork.openembedded.org/patch/1178
http://patchwork.openembedded.org/patch/1186
http://patchwork.openembedded.org/patch/1187
http://patchwork.openembedded.org/patch/1191
http://patchwork.openembedded.org/patch/1192
http://patchwork.openembedded.org/patch/1193
http://patchwork.openembedded.org/patch/1194
http://patchwork.openembedded.org/patch/1202
http://patchwork.openembedded.org/patch/1204
http://patchwork.openembedded.org/patch/1209
http://patchwork.openembedded.org/patch/1210
http://patchwork.openembedded.org/patch/1211
http://patchwork.openembedded.org/patch/1214
http://patchwork.openembedded.org/patch/1221
http://patchwork.openembedded.org/patch/1223
http://patchwork.openembedded.org/patch/1225
http://patchwork.openembedded.org/patch/1226
http://patchwork.openembedded.org/patch/1229
http://patchwork.openembedded.org/patch/1231
http://patchwork.openembedded.org/patch/1247
http://patchwork.openembedded.org/patch/1250
http://patchwork.openembedded.org/patch/1261
http://patchwork.openembedded.org/patch/1266
http://patchwork.openembedded.org/patch/1267
http://patchwork.openembedded.org/patch/1269
http://patchwork.openembedded.org/patch/1283
http://patchwork.openembedded.org/patch/1284
http://patchwork.openembedded.org/patch/1286
http://patchwork.openembedded.org/patch/1289
http://patchwork.openembedded.org/patch/1303
http://patchwork.openembedded.org/patch/1314
http://patchwork.openembedded.org/patch/1317
http://patchwork.openembedded.org/patch/1316
http://patchwork.openembedded.org/patch/1320
http://patchwork.openembedded.org/patch/1322
http://patchwork.openembedded.org/patch/1326
http://patchwork.openembedded.org/patch/1327
http://patchwork.openembedded.org/patch/1334
http://patchwork.openembedded.org/patch/1352
http://patchwork.openembedded.org/patch/1354
http://patchwork.openembedded.org/patch/1351
http://patchwork.openembedded.org/patch/1366
http://patchwork.openembedded.org/patch/1379
http://patchwork.openembedded.org/patch/1382
http://patchwork.openembedded.org/patch/1393
http://patchwork.openembedded.org/patch/1402
http://patchwork.openembedded.org/patch/1418
http://patchwork.openembedded.org/patch/1431
http://patchwork.openembedded.org/patch/1438
http://patchwork.openembedded.org/patch/1439
http://patchwork.openembedded.org/patch/1440
http://patchwork.openembedded.org/patch/1453
http://patchwork.openembedded.org/patch/1454
http://patchwork.openembedded.org/patch/1455
http://patchwork.openembedded.org/patch/1456
http://patchwork.openembedded.org/patch/1457
http://patchwork.openembedded.org/patch/1452
http://patchwork.openembedded.org/patch/1467
http://patchwork.openembedded.org/patch/1468
http://patchwork.openembedded.org/patch/1489
http://patchwork.openembedded.org/patch/1504
http://patchwork.openembedded.org/patch/1505

[-- Attachment #2: main.cc --]
[-- Type: text/x-c++src, Size: 5619 bytes --]

/*
 * Copyright (C) 2010 Holger Hans Peter Freyther
 *
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY APPLE COMPUTER, INC. ``AS IS'' AND ANY
 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
 * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL APPLE COMPUTER, INC. OR
 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
 * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

#include <QApplication>
#include <QWebFrame>
#include <QWebElement>
#include <QWebPage>

class Worker : public QObject {
    Q_OBJECT
public:
    Worker()
        : m_currentPage(1)
    {
        m_page = new QWebPage(this);
        connect(m_page, SIGNAL(loadFinished(bool)), SLOT(loadFinished(bool)));
    }

    void start()
    {
        qWarning("Trying to download page %d from patchwork", m_currentPage);
        QUrl url("http://patchwork.openembedded.org/project/openembedded/list/?order=date&page=" + QByteArray::number(m_currentPage));
        m_page->mainFrame()->load(url);
    }

    void startNext()
    {
        ++m_currentPage;
        start();
    }

    // Use git apply first and then use git am...
    void applyPatch(int patch_no)
    {
        QByteArray apply_cmd = "cd /home/ich/source/embedded/openembedded/; ";
        apply_cmd +=  "wget -o /dev/null -O /dev/stdout http://patchwork.openembedded.org/patch/" + QByteArray::number(patch_no) + "/mbox/";
        apply_cmd += " | git apply --check > /dev/null 2>&1";
        int result = system(apply_cmd.constData());
        if (result == 0) {
            QByteArray am_cmd = "cd /home/ich/source/embedded/openembedded/; ";
            am_cmd +=  "wget -o /dev/null -O /dev/stdout http://patchwork.openembedded.org/patch/" + QByteArray::number(patch_no) + "/mbox/";
            am_cmd += " | git am > /dev/null 2>&1";
            result = system(am_cmd.constData());
        }

        if (result != 0)
            m_failingPatches.append("http://patchwork.openembedded.org/patch/" + QByteArray::number(patch_no));
    }

    void handleCurrentPatch(QWebElement patch)
    {
        QWebElement link = patch.findFirst("a");
        QList<QWebElement> table = patch.findAll("td").toList();
        QString date, author, state;

        if (table.size() == 5) {
            date = table[1].toPlainText();
            author = table[2].toPlainText();
            state = table[4].toPlainText();
        }

        QString href = link.attribute("href");
        if (href.isEmpty() || !href.startsWith("/patch/")) {
            qWarning("Can not parse href from the link..");
            return;
        }

        // strip the '/patch/' and the last '/' from the url
        int patch_no = href.mid(sizeof "/patch/" - 1, href.length() - sizeof "/patch/").toInt();
        if (patch_no <= 0) {
            qWarning("Failed to find the patch number.");
            return;
        }

        qWarning("\tDate: %s Author: '%s' State: %s '%s'", qPrintable(date), qPrintable(author), qPrintable(state), qPrintable(link.toPlainText()));
        applyPatch(patch_no);

    }

    // query for all patches..
    void handleCurrentPage()
    {
        QWebElementCollection patches = m_page->mainFrame()->findAllElements("tr[id*=patch_row]");
        foreach(QWebElement element, patches)
            handleCurrentPatch(element);
    }

public Q_SLOTS:
    void loadFinished(bool result)
    {
        // Verify that we got the page we wanted
        QWebElement currentPage = m_page->mainFrame()->findFirstElement("span[title='Current Page']");
        if (currentPage.isNull()) {
            qWarning("Failed to find the current page.");
            return QApplication::exit(-1);
        }
        int currentPageNo = currentPage.toPlainText().toInt();
        if (currentPageNo != m_currentPage) {
            qWarning("The end was reached.");
            return QApplication::exit(0);
        }

        qWarning("Finished the download of page %d with status: %d", m_currentPage, result);
        handleCurrentPage();

        startNext();
    }

public:
    QWebPage* m_page;
    int m_currentPage;
    QList<QString> m_failingPatches;
};

int main(int argc, char **argv)
{
    QApplication app(argc, argv);

    Worker worker;
    worker.start();
    app.exec();

    if (worker.m_failingPatches.isEmpty())
        return 0;

    qWarning("# %d patches failed to apply.", worker.m_failingPatches.size());
    foreach(QString url, worker.m_failingPatches)
        qWarning("%s", qPrintable(url));
}

#include "main.moc"

Re: Patchwork Workflow improvement

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 23-01-10 09:43, Holger Hans Peter Freyther wrote:
> Hi All,
> 
> right now I'm writing a Qt 4.6 based application that might help us a bit with 
> the patchwork queue.
> 
> The idea of the script is to have two things at the end:
> 	1.) a branch with all the patches that applied and that can be published
> 	2.) a list of patches that don't apply.
> 
> 
> What people are supposed to:
> 
> 1.) Cherry pick from the branch
> 2.) Send updated patches, close the reports on patchwork...

I think people that apply those patches to oe.dev should fix up the
commit message as well. The stack Khem pushed this week is just awfull.

I guess we should flag patches with bad commit messages as 'rejected' in
pw ASAP. We can weed out most of them by reject anything that lacks a
':' in the subject line.

regards,

Koen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFLWtKwMkyGM64RGpERAgSAAJ4y0iproyWSH/ojheSFN40fdj16NgCfRfj3
rD4D/FAM3X5NeYdEUTRsbDQ=
=4N5e
-----END PGP SIGNATURE-----

Re: New state for patches not complying with commit policy (was: Patchwork Workflow improvement)

[Paul Menzel]

[-- Attachment #1: Type: text/plain, Size: 1932 bytes --]

Am Samstag, den 23.01.2010, 11:42 +0100 schrieb Koen Kooi:
> On 23-01-10 09:43, Holger Hans Peter Freyther wrote:
> > right now I'm writing a Qt 4.6 based application that might help us
> a bit with 
> > the patchwork queue.
> > 
> > The idea of the script is to have two things at the end:
> >       1.) a branch with all the patches that applied and that can be published
> >       2.) a list of patches that don't apply.
> > 
> > 
> > What people are supposed to:
> > 
> > 1.) Cherry pick from the branch
> > 2.) Send updated patches, close the reports on patchwork...
> 
> I think people that apply those patches to oe.dev should fix up the
> commit message as well. The stack Khem pushed this week is just
> awfull.
> 
> I guess we should flag patches with bad commit messages as 'rejected' in
> pw ASAP. We can weed out most of them by reject anything that lacks a
> ':' in the subject line.

I am just getting into Patchwork [1][2], so I am sorry if I miss the
obvious.

If it is possible, I suggest to add a new state (»Rejected – Commit
Policy«) for patches not complying with the commit policy [3]?

As a bonus an automatically answer is sent to the message with the patch
to the list with the following content, which is just a draft/suggestion
of course.


        Dear patch author,
        
        
        thank you for your patch. Unfortunately we had to reject it
        because it does not comply with our commit policy [3].
        
        It would be great if you could resend your patch with this issue
        fixed so that we can review and commit it as soon as possible.
        
        
        Thank you,
        
        OpenEmbedded people


Is there a way to automate this?


Thanks,

Paul


[1] http://patchwork.openembedded.org/
[2] http://ozlabs.org/~jk/projects/patchwork/
[3] http://wiki.openembedded.net/index.php/Commit_Policy

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 205 bytes --]

Re: Patchwork Workflow improvement

[Holger Hans Peter Freyther]

On Saturday 23 January 2010 11:42:56 Koen Kooi wrote:

> > 1.) Cherry pick from the branch
> > 2.) Send updated patches, close the reports on patchwork...
> 
> I think people that apply those patches to oe.dev should fix up the
> commit message as well. The stack Khem pushed this week is just awfull.

Right... cherry-pick and use git commit --amend afterwards



> 
> I guess we should flag patches with bad commit messages as 'rejected' in
> pw ASAP. We can weed out most of them by reject anything that lacks a
> ':' in the subject line.

that is a difficult decision. E.g.  of course we should handle patches that have 
a good commit message faster than patches that are of lower quality. But then 
again just rejecting a patch is also bad, e.g. maybe the contributor just 
didn't know better and needs a helpful hand?

I was going through them older stuff, e.g. someone sending a patch against a 
package but not as part of a fix to the recipe... So the contributor did his 
homework, he looked for similiar build errors and found a patch, it is low 
hanging fruit to fix that in OE.


The bottom line is. People don't send us "bad" patches to piss us off, they try 
to contribute and we should help them to be useful contributors.

z.

Re: New state for patches not complying with commit policy

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 23-01-10 13:19, Paul Menzel wrote:
> Am Samstag, den 23.01.2010, 11:42 +0100 schrieb Koen Kooi:
>> On 23-01-10 09:43, Holger Hans Peter Freyther wrote:
>>> right now I'm writing a Qt 4.6 based application that might help us
>> a bit with 
>>> the patchwork queue.
>>>
>>> The idea of the script is to have two things at the end:
>>>       1.) a branch with all the patches that applied and that can be published
>>>       2.) a list of patches that don't apply.
>>>
>>>
>>> What people are supposed to:
>>>
>>> 1.) Cherry pick from the branch
>>> 2.) Send updated patches, close the reports on patchwork...
>>
>> I think people that apply those patches to oe.dev should fix up the
>> commit message as well. The stack Khem pushed this week is just
>> awfull.
>>
>> I guess we should flag patches with bad commit messages as 'rejected' in
>> pw ASAP. We can weed out most of them by reject anything that lacks a
>> ':' in the subject line.
> 
> I am just getting into Patchwork [1][2], so I am sorry if I miss the
> obvious.
> 
> If it is possible, I suggest to add a new state (»Rejected – Commit
> Policy«) for patches not complying with the commit policy [3]?

We can add that easily, we did the same for the 'applied' state.

> As a bonus an automatically answer is sent to the message with the patch
> to the list with the following content, which is just a draft/suggestion
> of course.

I'm unsure if we can do that. Patchwork does support sending emails, but
I left it disabled since I suck at email stuff.
If someone knows how to make patchwork do that we can consider turning
such a feature on.

regards,

Koen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFLWwNXMkyGM64RGpERAp59AJ476WIGYubuaxiqdmQ4yl4r+z2EWgCeOGqM
6D0qHrY1hvikflj9fpVLXAk=
=VKyT
-----END PGP SIGNATURE-----

Re: Patchwork Workflow improvement

[Paul Menzel]

[-- Attachment #1: Type: text/plain, Size: 1517 bytes --]

Dear Holger,


Am Samstag, den 23.01.2010, 10:56 +0100 schrieb Holger Hans Peter Freyther:
> On Saturday 23 January 2010 09:43:24 Holger Hans Peter Freyther wrote:

[…]

> I have pushed "holger/staging-branch" to the server. It does contain all 
> patches from patchwork that are still applying. Please go through the branch 
> and cherry-pick or follow up on the emails..

if I understood correctly you mean with »patches from patchwork that are
still applying« that the mbox file can be applied with `git am` as
`pw-am.sh` does [1].

That requires that all patches are submitted using `git send-email`,
which has been suggested lately on this list by several people, or that
everyone uses a MUA which provides easy handling of mbox files created
by `git format-patch`. Is that correct?

I will start a separate thread regarding `git send-email`.

> I also have a new version of the 
> program I use attached... I will probably switch it to use pwclient in the 
> future.

Thank you for your effort and bringing up `pwclient`. It is so sad, that
there exists a lot of tools and stuff one does not know about which
could safe work and time for oneself. I wonder why others projects did
not face the same problems yet. Probably every project or developers
wrote her/his own scripts (as OE did with `pw-am.sh`) and everyone
reinvents the wheel.


Thanks,

Paul


[1] http://wiki.openembedded.net/index.php/Patchwork
[2] http://wiki.openembedded.net/index.php/Commit_Policy

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 205 bytes --]

Re: Patchwork Workflow improvement

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 23-01-10 15:53, Paul Menzel wrote:
> Dear Holger,
> 
> 
> Am Samstag, den 23.01.2010, 10:56 +0100 schrieb Holger Hans Peter Freyther:
>> On Saturday 23 January 2010 09:43:24 Holger Hans Peter Freyther wrote:
> 
> […]
> 
>> I have pushed "holger/staging-branch" to the server. It does contain all 
>> patches from patchwork that are still applying. Please go through the branch 
>> and cherry-pick or follow up on the emails..
> 
> if I understood correctly you mean with »patches from patchwork that are
> still applying« that the mbox file can be applied with `git am` as
> `pw-am.sh` does [1].
> 
> That requires that all patches are submitted using `git send-email`,

That is not a hard requirement, but seems to be the most fool-proof one

>  Probably every project or developers
> wrote her/his own scripts (as OE did with `pw-am.sh`) and everyone
> reinvents the wheel.

In this case it was purposely reinvented since getting OE developers to
install extra software is like pulling teeth, so a simple shell script
was added.

regards,

Koen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFLWxH0MkyGM64RGpERAgAPAJ0VTGsXK2zviMV5G8Prq/Had2jzEgCfWwxu
u1NK7hM+MNyDm2tDhAFIIug=
=ONPl
-----END PGP SIGNATURE-----

Re: Patchwork Workflow improvement

[Petr Štetiar]

Holger Hans Peter Freyther <holger+oe@freyther.de> [2010-01-23 09:43:24]:

> Hi All,

Hi,

>  * THIS SOFTWARE IS PROVIDED BY APPLE COMPUTER, INC. ``AS IS'' AND ANY

Apple? :)

-- ynezz

Re: New state for patches not complying with commit policy (was: Patchwork Workflow improvement)

[Phil Blundell]

On Sat, 2010-01-23 at 13:19 +0100, Paul Menzel wrote:
> If it is possible, I suggest to add a new state (»Rejected – Commit
> Policy«) for patches not complying with the commit policy [3]?

If the patch is basically correct then it seems a little bit officious
to reject it purely because it fails to conform to the commit policy in
some way.  Most of the issues listed in the policy are easy enough to
fix up when applying the patch, after all.

By all means send a note to the patch submitter explaining how they
could write a better commit message (or whatever) in the future, but I
don't think there is any need to reject the patch just for that.

p.

Re: Patchwork Workflow improvement

[Rolf Leggewie]

Holger Hans Peter Freyther wrote:
> right now I'm writing a Qt 4.6 based application that might help us a bit with 
> the patchwork queue.

Holger,

thank you for working on better integration of patchwork into the
workflow.  I was wondering if it was possible for the script to check
for patches attached to open tickets that reverse-apply (IOW, the
committer forgot to close the ticket).  If your script would then close
these tickets as APPLIED, that would be splendid.

Regards

Rolf

Re: Patchwork Workflow improvement

[Holger Hans Peter Freyther]

On Saturday 23 January 2010 16:19:32 Petr Štetiar wrote:
> Holger Hans Peter Freyther <holger+oe@freyther.de> [2010-01-23 09:43:24]:
> > Hi All,
> 
> Hi,
> 
> >  * THIS SOFTWARE IS PROVIDED BY APPLE COMPUTER, INC. ``AS IS'' AND ANY
> 
> Apple? :)

Haha.. I copied that from my WebKit sources :)

Re: Patchwork Workflow improvement

[Holger Hans Peter Freyther]

On Saturday 23 January 2010 15:53:05 Paul Menzel wrote:
> Dear Holger,

> > I also have a new version of the
> > program I use attached... I will probably switch it to use pwclient in
> > the future.
> 
> Thank you for your effort and bringing up `pwclient`. It is so sad, that
> there exists a lot of tools and stuff one does not know about which
> could safe work and time for oneself. I wonder why others projects did
> not face the same problems yet. Probably every project or developers
> wrote her/his own scripts (as OE did with `pw-am.sh`) and everyone
> reinvents the wheel.

Well, I remembered that pwclient exists, so I went to the patchwork website at 
ozlabs.org but I could not find any reference to the tools (or the XML RPC 
API). It was just a matter that writing my version was faster than the time I 
would have to spent searching... :(.

I will probably give up on my version and now look into pw-am.sh.

pwclient (was: Patchwork Workflow improvement)

[Paul Menzel]

[-- Attachment #1: Type: text/plain, Size: 1330 bytes --]

Am Samstag, den 23.01.2010, 17:46 +0100 schrieb Holger Hans Peter
Freyther:
> On Saturday 23 January 2010 15:53:05 Paul Menzel wrote:
> > Dear Holger,
> 
> > > I also have a new version of the
> > > program I use attached... I will probably switch it to use pwclient in
> > > the future.
> > 
> > Thank you for your effort and bringing up `pwclient`. It is so sad, that
> > there exists a lot of tools and stuff one does not know about which
> > could safe work and time for oneself. I wonder why others projects did
> > not face the same problems yet. Probably every project or developers
> > wrote her/his own scripts (as OE did with `pw-am.sh`) and everyone
> > reinvents the wheel.
> 
> Well, I remembered that pwclient exists, so I went to the patchwork website at 
> ozlabs.org but I could not find any reference to the tools (or the XML RPC 
> API). It was just a matter that writing my version was faster than the time I 
> would have to spent searching... :(.

It seems to be available at [1] and in the Patchwork Git repository
under `apps/patchwork/bin`.

> I will probably give up on my version and now look into pw-am.sh.

If the only problem was that you could not find `pwclient` I hope we can
improve `pwclient`.


Thanks,

Paul


[1] http://patchwork.ozlabs.org/help/pwclient/

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 205 bytes --]

Re: New state for patches not complying with commit policy

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 23-01-10 17:11, Phil Blundell wrote:
> On Sat, 2010-01-23 at 13:19 +0100, Paul Menzel wrote:
>> If it is possible, I suggest to add a new state (»Rejected – Commit
>> Policy«) for patches not complying with the commit policy [3]?
> 
> If the patch is basically correct then it seems a little bit officious
> to reject it purely because it fails to conform to the commit policy in
> some way.  Most of the issues listed in the policy are easy enough to
> fix up when applying the patch, after all.

And yet OE developers aren't fixing those up (latest example: the patch
stack Khem pushed last week), however easy it may be.

Maybe it's a case of sampling bias since I only inspect bad commits, but
in any case, the commit message of the commits that end up in .dev needs
to get improved.

regards,

Koen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFLWzlJMkyGM64RGpERAvyDAKCE7pBD3krfl2wBN4GvYi59Wi+g8wCeLNmA
Wg9qotVtOlkyOtLgSkhTrPQ=
=ar+b
-----END PGP SIGNATURE-----

Re: pwclient

[Paul Menzel]

[-- Attachment #1: Type: text/plain, Size: 1884 bytes --]

Am Samstag, den 23.01.2010, 17:53 +0100 schrieb Paul Menzel:
> Am Samstag, den 23.01.2010, 17:46 +0100 schrieb Holger Hans Peter Freyther:
> > On Saturday 23 January 2010 15:53:05 Paul Menzel wrote:
> > > Dear Holger,
> > 
> > > > I also have a new version of the
> > > > program I use attached... I will probably switch it to use pwclient in
> > > > the future.
> > > 
> > > Thank you for your effort and bringing up `pwclient`. It is so sad, that
> > > there exists a lot of tools and stuff one does not know about which
> > > could safe work and time for oneself. I wonder why others projects did
> > > not face the same problems yet. Probably every project or developers
> > > wrote her/his own scripts (as OE did with `pw-am.sh`) and everyone
> > > reinvents the wheel.
> > 
> > Well, I remembered that pwclient exists, so I went to the patchwork website at 
> > ozlabs.org but I could not find any reference to the tools (or the XML RPC 
> > API). It was just a matter that writing my version was faster than the time I 
> > would have to spent searching... :(.
> 
> It seems to be available at [1] and in the Patchwork Git repository
> under `apps/patchwork/bin`.

I failed the first time trying to compose the help by myself, but
reading and trying [1] you find `pwclient` under the »project info« page
[2] at the bottom [3] directly in our installation. There you can
download the configuration file [4] too.

> > I will probably give up on my version and now look into pw-am.sh.
> 
> If the only problem was that you could not find `pwclient` I hope we can
> improve `pwclient`.


Thanks,

Paul

> [1] http://patchwork.ozlabs.org/help/pwclient/
[2] http://patchwork.openembedded.org/project/openembedded/
[3] http://patchwork.openembedded.org/help/pwclient/
[4] http://patchwork.openembedded.org/project/openembedded/pwclientrc/

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 205 bytes --]

Re: Patchwork Workflow improvement

[Khem Raj]

On Sat, Jan 23, 2010 at 4:56 AM, Holger Hans Peter Freyther
<holger+oe@freyther.de> wrote:
> On Saturday 23 January 2010 11:42:56 Koen Kooi wrote:
>
>> > 1.) Cherry pick from the branch
>> > 2.) Send updated patches, close the reports on patchwork...
>>
>> I think people that apply those patches to oe.dev should fix up the
>> commit message as well. The stack Khem pushed this week is just awfull.
>
yeah rebasing bit me in foot. my ammended commit messages did not make
it to repo because I did'nt  finish rebasing (git rebase --continue)
and pushed it. It would have been nice if git warned me about being in
the middle of rebase before pushing.

Re: Patchwork Workflow improvement

[Khem Raj]

On Sat, Jan 23, 2010 at 3:40 PM, Khem Raj <raj.khem@gmail.com> wrote:
> On Sat, Jan 23, 2010 at 4:56 AM, Holger Hans Peter Freyther
> <holger+oe@freyther.de> wrote:
>> On Saturday 23 January 2010 11:42:56 Koen Kooi wrote:
>>
>>> > 1.) Cherry pick from the branch
>>> > 2.) Send updated patches, close the reports on patchwork...
>>>
>>> I think people that apply those patches to oe.dev should fix up the
>>> commit message as well. The stack Khem pushed this week is just awfull.
>>
> yeah rebasing bit me in foot. my ammended commit messages did not make
> it to repo because I did'nt  finish rebasing (git rebase --continue)
> and pushed it. It would have been nice if git warned me about being in
> the middle of rebase before pushing.
>


I guess we could install commit hooks similar to uclibc to rejects
pushes which do
not have proper commit message format and also bogus branch merges that is seen
so common

-Khem

Re: Patchwork Workflow improvement

[Rolf Leggewie]

Khem Raj wrote:
> I guess we could install commit hooks similar 

[x] show code

;-)

Useful Git hooks for ensuring code quality. (was: Patchwork Workflow improvement)

[Paul Menzel]

[-- Attachment #1: Type: text/plain, Size: 1441 bytes --]

Am Sonntag, den 24.01.2010, 01:28 +0100 schrieb Rolf Leggewie: 
> Khem Raj wrote:
> > I guess we could install commit hooks similar 
> 
> [x] show code

I mentioned this too in my other (rather long) mail »[oe] [RFC]
Perfect(?) workflow for patches.« [1].

I hope I will find some time to code something up. But here are some
links.

I copy the links from [1] into this message. Unfortunately I could not
find the hooks uclibc is using.

Besides in `git help hooks`, Git hooks are covered very well in the Pro
Git book section 7.3 [2]. There is a pretty good blog post regarding
this issue [3] too.

1. Examples in the Pro Git book from section 7.4.
http://progit.org/book/ch7-4.html

2. Git hooks used by arora project (also referred to in [3]):
http://github.com/Arora/arora/tree/master/git_hooks/

3. Script which can be hooked up to automatically change state in
Patchwork (for Subversion/SVN though):
http://lists.ozlabs.org/pipermail/patchwork/attachments/20091006/13e2340a/attachment.asc

4. Examples in the Git project
http://git.kernel.org/?p=git/git.git;a=tree;f=contrib/hooks


I suggest to also create a directory `contrib/hooks` where we should
save tested hooks to share them.


Thanks,

Paul


[1] http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-January/016715.html
[2] http://progit.org/book/ch7-3.html
[3] http://benjamin-meyer.blogspot.com/2008/10/git-hooks.html

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 205 bytes --]