Ros-dev

ros-dev@reactos.org

1 participants
5793 discussions

RosTests/RosApps in GitHub CI
by Thomas Faber 03 Nov '17

03 Nov '17

IIUC we currently always build rostests, but never rosapps in CI. Would it be possible to analyze the changed paths instead and * enable rostests iff modules/rostests/ or sdk/ was changed * enable rosapps iff modules/rosapps/ or sdk/ was changed? That should speed up builds that don't change rostests, while also making sure that changes to rosapps aren't let into master if they break build. If this is nontrivial I may have a look at it myself but someone would need to point me to the right configuration/scripts please. Thanks! -Thomas

3 3

Re: [ros-dev] [ros-diffs] [reactos] 01/01: CID 1206831 Dereference after null check
by Jerome Gardou 31 Oct '17

31 Oct '17

HI, that doesn't look good, as shown by https://reactos.org/testman/compare.php?ids=56275,56276 Jérôme Le 29/10/2017 à 11:17, Samuel Serapion a écrit : > https://git.reactos.org/?p=reactos.git;a=commitdiff;h=b3b2a23f05e5188dc1475… > > commit b3b2a23f05e5188dc1475961fcd7f036f0046d25 > Author: Samuel Serapion <samcharly(a)hotmail.com> > AuthorDate: Fri Oct 20 14:00:32 2017 -0400 > > CID 1206831 Dereference after null check > > BytesRead is an optional out parameter and must be checked before being written to. > --- > sdk/lib/rtl/memstream.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/sdk/lib/rtl/memstream.c b/sdk/lib/rtl/memstream.c > index 0549424ca4..8fe4169fb1 100644 > --- a/sdk/lib/rtl/memstream.c > +++ b/sdk/lib/rtl/memstream.c > @@ -185,7 +185,8 @@ RtlReadMemoryStream( > > Stream->Current = (PUCHAR)Stream->Current + CopyLength; > > - *BytesRead = CopyLength; > + if (BytesRead) > + *BytesRead = CopyLength; > > return S_OK; > } >

3 2

Re: [ros-dev] [ros-diffs] [reactos] 01/01: CID 1206831 Dereference after null check
by hermes.belusca＠sfr.fr 31 Oct '17

31 Oct '17

<span style="font-family:arial,helvetica,sans-serif; font-size:12px">Certainly not a "feature", but just that (certainly because it is only for user-mode AND the out pointer is not optional) the MS dev who introduced these functions didn't want to (or just more simply forgot to) not check for such NULL pointer.</span><br> <span style="font-family:arial,helvetica,sans-serif; font-size:12px">And thus, if you pass NULL, it's just your fault if your app crashes.</span><br> <span style="font-family:arial,helvetica,sans-serif; font-size:12px">And of course, since ReactOS also want to behave similarly... we don't check for NULL either!</span><br> <br> <span style="font-family:arial,helvetica,sans-serif; font-size:12px">H.</span> <div class="gl_quote" style="margin-top: 20px; padding-top: 5px;"> <div class="gl_quote" style="margin-top: 20px; padding-top: 5px;">De : xxxx<br> A : ros-dev(a)reactos.org<br> Envoyé: mardi 31 octobre 2017 16:10<br> Objet : Re: [ros-dev] [ros-diffs] [reactos] 01/01: CID 1206831 Dereference after null check<br>   <div class="gl_quoted"> <div dir="ltr">Seems like this API has a 'feature' where by it throws exceptions if <span style="font-size:12.8px">BytesRead is null?</span></div> <div class="gmail_extra">  <div class="gmail_quote">On Sun, Oct 29, 2017 at 8:02 AM, Jerome Gardou <span dir="ltr"><<a href="mailto:jerome.gardou@reactos.org" target="_blank">jerome.gardou(a)reactos.org</a>></span> wrote: <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">HI,<br> <br> that doesn't look good, as shown by <a href="https://reactos.org/testman/compare.php?ids=56275,56276" rel="noreferrer" target="_blank">https://reactos.org/testman/co<wbr>mpare.php?ids=56275,56276</a><br> <br> Jérôme<br> <br> <br> Le 29/10/2017 à 11:17, Samuel Serapion a écrit : <blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><a href="https://git.reactos.org/?p=reactos.git;a=commitdiff;h=b3b2a23f05e5188dc1475…" rel="noreferrer" target="_blank">https://git.reactos.org/?p=rea<wbr>ctos.git;a=commitdiff;h=b3b2a2<wbr>3f05e5188dc1475961fcd7f036f004<wbr>6d25</a><br> <br> commit b3b2a23f05e5188dc1475961fcd7f0<wbr>36f0046d25<br> Author: Samuel Serapion <<a href="mailto:samcharly@hotmail.com" target="_blank">samcharly(a)hotmail.com</a>><br> AuthorDate: Fri Oct 20 14:00:32 2017 -0400<br> <br>      CID 1206831 Dereference after null check<br>           BytesRead is an optional out parameter and must be checked before being written to.<br> ---<br>   sdk/lib/rtl/memstream.c | 3 ++-<br>   1 file changed, 2 insertions(+), 1 deletion(-)<br> <br> diff --git a/sdk/lib/rtl/memstream.c b/sdk/lib/rtl/memstream.c<br> index 0549424ca4..8fe4169fb1 100644<br> --- a/sdk/lib/rtl/memstream.c<br> +++ b/sdk/lib/rtl/memstream.c<br> @@ -185,7 +185,8 @@ RtlReadMemoryStream(<br>         Stream->Current = (PUCHAR)Stream->Current + CopyLength;<br>   -    *BytesRead = CopyLength;<br> +    if (BytesRead)<br> +        *BytesRead = CopyLength;<br>         return S_OK;<br>   }<br>  </blockquote> <br> <br> ______________________________<wbr>_________________<br> Ros-dev mailing list<br> <a href="mailto:Ros-dev@reactos.org" target="_blank">Ros-dev(a)reactos.org</a><br> <a href="http://www.reactos.org/mailman/listinfo/ros-dev" rel="noreferrer" target="_blank">http://www.reactos.org/mailman<wbr>/listinfo/ros-dev</a></blockquote> </div> </div> <br> <br> <br> _______________________________________________<br> Ros-dev mailing list<br> Ros-dev(a)reactos.org<br> http://www.reactos.org/mailman/listinfo/ros-dev</div> <div class="gl_quoted"> </div> </div> </div>

1 0

Re: [ros-dev] [ros-diffs] [reactos] 01/01: [CRT] fix bug in _rotr implementation
by Jerome Gardou 28 Oct '17

28 Oct '17

Hmpf, somehow the change on my github profile was messed up. Sorry about that, the next commit should be under my real name. Jérôme Le 28/10/2017 à 10:28, zefklop a écrit : > https://git.reactos.org/?p=reactos.git;a=commitdiff;h=07c72955457267ee26f60… > > commit 07c72955457267ee26f6086e75c5595fbce76e06 > Author: zefklop <zefklop(a)users.noreply.github.com> > AuthorDate: Sat Oct 28 10:28:27 2017 +0200 > > [CRT] fix bug in _rotr implementation > --- > sdk/lib/crt/stdlib/rot.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/sdk/lib/crt/stdlib/rot.c b/sdk/lib/crt/stdlib/rot.c > index 825a76d3ed..7637a5e9b6 100644 > --- a/sdk/lib/crt/stdlib/rot.c > +++ b/sdk/lib/crt/stdlib/rot.c > @@ -38,7 +38,7 @@ unsigned int _rotr( unsigned int value, int shift ) > if ( shift < 0 ) > return _rotl(value,-shift); > > - if ( shift > max_bits<<3 ) > + if ( shift > max_bits ) > shift = shift % max_bits; > return (value >> shift) | (value << (max_bits-shift)); > } >

1 0

Status Meeting (October 2017)
by Giannis Adamopoulos 26 Oct '17

26 Oct '17

Hi all! Let me invite you to the monthly status meeting taking place today, October 26, 19:00 UTC. As of now we know that both Colin and Pierre will not be able to attend, or be able to control our IRC server so unless we get any further notice from either of them, the meeting will take place in #reactos-meeting in freenode. No agenda proposals have been submitted to me so far. You are welcome to suggest some points to discuss. One that I would like to propose is about the upcoming release that should be branched soon. Of course we will have our status reports. So please have them ready, so we get it done quickly! Giannis

3 2

JIRA up again - Do we still need FishEye?
by Colin Finck 16 Oct '17

16 Oct '17

Hi all! Our two Atlassian tools JIRA and FishEye have been unavailable for some hours again. I have brought JIRA back up, but FishEye has been eating RAM and CPU like crazy again. In my opinion, this is unacceptable for a tool that is just a simple commit browser to us. Also I don't have any time to constantly monitor it this weekend, so I'm leaving it shut down for now. With us having https://github.com/reactos/reactos/commits/master and https://git.reactos.org/?p=reactos.git;a=summary now, do we still need FishEye at all? If so, please tell me what features of it you would miss if it stays down. Maybe there are other alternatives. My last FishEye "performance" bug report took 3 months to resolve, and I'm reluctant to go the same route again for another such bug.. Cheers, Colin

3 2

Question about commits with no commit log, or commits with no files but just the log, on Git.
by Hermès BÉLUSCA - MAÏTO 16 Oct '17

16 Oct '17

‌Hello all,<br> <br> After having noticed https://github.com/reactos/reactos/commit/50ae5e7c5268222718174221366169e2b…<br> I am wondering whether it is possible to add mechanisms (like git pre-commit hooks) that check for, and<br> prevent potential problematic commits like, commits without commit log, or commits with no files (but with a log).<br> <br> Any ideas?<br> <br> Cheers,<br> Hermes

4 3

Testing Pull Requests
by Colin Finck 14 Oct '17

14 Oct '17

Hi all! Apart from basic build testing through AppVeyor and Travis-CI, you can now also run all regression tests for the code belonging to a Pull Request. Just log in to our BuildBot at https://build.reactos.org with your development account credentials (formerly "SVN credentials"), select a builder in the waterfall view and fill the field "Branch" in the "Force a Build" section. You have two options here: * "refs/pull/<ID>/head" Builds your branch, unmodified. That means, if you haven't kept your branch up to date with the main ReactOS repo, latest changes won't be included. * "refs/pull/<ID>/merge" Builds the changes of your branch merged to the main ReactOS repo. That means, latest changes in the main ReactOS repo will be included. Replace <ID> by the number of your Pull Request. As the resulting hash may be hard to predict, the "prepare source" step will now give detailed information about the latest commit included in your build: https://build.reactos.org/builders/Build%20GCCLin_x86/builds/19201/steps/pr… Have fun! Colin

3 3

GitHub Newbie Question
by Hermès BÉLUSCA - MAÏTO 10 Oct '17

10 Oct '17

<span style="font-family:arial,helvetica,sans-serif; font-size:12px">‌Hi all,<br> <br> I've a newbie question related to GitHub pull requests: How can one see whether commits pertaining to a pull request have both a valid full committer name and email set? So far I can only see the nicknames.<br> <br> Cheers,<br> Hermes</span>

3 2

Re: [ros-dev] GitHub Newbie Question
by hermes.belusca＠sfr.fr 10 Oct '17

10 Oct '17

<span style="font-family:arial,helvetica,sans-serif; font-size:12px">Oops yeah sorry, it was about "author name", not committer name. But you've understood my point.<br> Thanks!</span><br>   <div class="gl_quote" style="margin-top: 20px; padding-top: 5px;">De : gigaherz(a)gmail.com<br> A : hermes.belusca@sfr.fr,ros-dev@reactos.org<br> Envoyé: mardi 10 octobre 2017 12:06<br> Objet : Re: [ros-dev] GitHub Newbie Question<br>   <div class="gl_quoted">Open the commit URL, such as https://github.com/reactos/reactos/commit/44060c284194546703805453f3985ed2b…, but with .patch appended to the end. This shows the author name as part of the "From:" line. The commiter doens't matter since it will be set to whoever presses the merge button. On 10 October 2017 at 11:56, Hermès BÉLUSCA - MAÏTO wrote: > ‌Hi all, > > I've a newbie question related to GitHub pull requests: How can one see > whether commits pertaining to a pull request have both a valid full > committer name and email set? So far I can only see the nicknames. > > Cheers, > Hermes > _______________________________________________ > Ros-dev mailing list > Ros-dev(a)reactos.org > http://www.reactos.org/mailman/listinfo/ros-dev _______________________________________________ Ros-dev mailing list Ros-dev(a)reactos.org http://www.reactos.org/mailman/listinfo/ros-dev</div> <div class="gl_quoted"> </div> </div>

1 0

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Ros-dev